It’s often recommended to install from backports. These are newer versions of apps that have been explicitly taken out of testing and packaged for the stable release. i.e. if you’re running ‘buster’ you would install from buster-backports.

It’s possible that one would use the identifiers ‘stable’ and ’testing’ which pegs you to whatever is current, rather than to a specific release.

echo deb http://deb.debian.org/debian buster-backports main | sudo tee /etc/apt/sources.list.d/buster-backports.list
sudo apt update

# seach for the most recent amd64 image.
sudo apt search -t buster-backports linux-image-5
sudo apt install -t buster-backports linux-image-5.2.0-0.bpo.3-amd64-unsigned
sudo apt install -t buster-backports btrfs-progs

Install from Release Candidate

If there’s no backport, you can install from the release candidate. These are simply upcoming versions of debian that haven’t been released yet

To install the kernel from the experimental version of debian, add the repo to your sources and explicitly add the kernel (this is safe to add to your repos because experimental packages aren’t installed by default)

sudo su -c "echo deb http://deb.debian.org/debian experimental main > /etc/apt/sources.list.d/experimental.list"
sudo apt update
sudo apt -t experimental search linux-image
sudo apt -t experimental install linux-image-XXX
sudo apt -t experimental install btrfs-progs

Install from Generic

You can also download the packages and manually install.

Navigate to

http://kernel.ubuntu.com/~kernel-ppa/mainline/

And download, similar to this (from a very long time ago :-)

wget https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.0/linux-headers-5.0.0-050000_5.0.0-050000.201903032031_all.deb
wget https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.0/linux-headers-5.0.0-050000-generic_5.0.0-050000.201903032031_amd64.deb
wget https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.0/linux-image-unsigned-5.0.0-050000-generic_5.0.0-050000.201903032031_amd64.deb
wget https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.0/linux-modules-5.0.0-050000-generic_5.0.0-050000.201903032031_amd64.deb

Troubleshooting

The following signatures couldn’t be verified because the public key is not available:

sudo apt-key adv –recv-key –keyserver keyserver.ubuntu.com 648ACFD622F3D138

Sources

https://unix.stackexchange.com/questions/432406/install-the-latest-rc-kernel-on-debian https://wiki.debian.org/HowToUpgradeKernel https://www.tecmint.com/upgrade-kernel-in-ubuntu/ https://raspberrypi.stackexchange.com/questions/12258/where-is-the-archive-key-for-backports-debian-org/60051#60051

2 - Ceph

Ceph is a distributed object storage system that supports both replicated and erasure coding of data. So you can design it to be fast or efficient. Or even both.

It’s complex compared to other solutions, but it’s also the main focus of Redhat and other development. So it may eclipse other technologies just through adaption.

It also comes ‘baked-in’ with Proxmox. If you’re already using PVE it’s worth deploying over others, as long as you’re willing to devote the time to learning it.

2.1 - Ceph Proxmox

Overview

There are two main use-cases;

Virtual Machines
Bulk Storage

They both provide High Availability. But VMs need speed whereas Bulk should be economical. What makes Ceph awesome is you can do both - all with the same set of disks.

Preparation

Computers

Put 3 or 4 PCs on a LAN. Have at least a couple HDDs in addition to a boot drive. 1G of RAM per TB of disk is recommended¹ and it will use it. You can have less RAM, it will just be slower.

Network

The speed of Ceph is essentially a third or half your network speed. With a 1 Gig NIC you’ll average around 60 MB/Sec for file operations (multiple copies are being saved behind the scenes). This sounds terrible, but in reality its fine for a cluster that serves up small files and/or media streams. But it will take a long time to get data onto the system.

If you need more, install a secondary NIC for Ceph traffic. Do that before you configure the system, if you can. Doing it after is hard. You can use a mesh config via the PVE docs² or purchase a switch. Reasonable 2.5 Gb switches and NICs can now be had.

Installation

Proxmox

Install PVE and cluster the servers.

Ceph

Double check the Ceph repo is current by comparing what you have enabled:

grep ceph  /etc/apt/sources.list.d/*

Against what PVE has available.

curl -s https://enterprise.proxmox.com/debian/ | grep ceph

If you don’t have the latest, take a look at Install PVE to update the Ceph repo.

After that, log into the PVE web GUI, click on each server and in that server’s submenu, click on Ceph. It will prompt for permission to install. When the setup windows appears, select the newest version and the No-Subscription repository. You can also refer to the official notes.

The wizard will ask some additional configuration options for which you can take the defaults and finish. If you have an additional ceph-specific network hardware, set it up with a separate IP range and choose interface for both the public network and cluster network.

Configuration

Ceph uses multiple daemons on each node.

Monitors, to keep track of what’s up and down.
Managers, to gather performance data.
OSDs, a service per disk to store and retrieve data.
Metadata Servers, to handle file permissions and such.

To configure them, you will:

Add a monitor and manager to each node
Add each node’s disks as OSD, or (Object Storage Devices)
Add metadata servers to each node
Create Pools, where you group OSDs and choose the level of resiliency
Create a Filesystem

Monitors and Managers

Ceph recommends at least three monitors³ and manager processes⁴. To install, click on a server’s Ceph menu and select the Monitor menu entry. Add a monitor and manager to the first three nodes.

OSDs

The next step is to add disks - aka Object Storage Devices. Select a server from the left-hand menu, and from that server’s Ceph menu, select OSD. Then click Create: OSD, select the disks and click create. If they are missing, enter the server’s shell and issue the command wipefs -a /dev/sdX on the disks in question.

If you have a lot of disks, you can also do this at the shell

# Assuming you have 5 disks, starting at 'b'
for X in {b..f}; do echo pveceph osd create /dev/sd$X; done

Metadata Servers

To add MDSs, click on the CephFS submenu for each server and click Create under the Metadata Servers section. Don’t create a CephFS yet, however.

Pools and Resiliency

We are going to create two pools; a replicated pool for VMs and an erasure coded pool for bulk storage.

If you want to mix SSDs and HDDs see the storage tiering page before creating the pools. You’ll want to set up classes of storage and create pools based on that page.

Replicated Pool

On this pool we’ll use the default replication level that gives us three copies of everything. These are guaranteed at a per-host level. Loose any one or two hosts, no problem. But loose individual disks from all three hoots at the same time and you’re out of luck.

This is the GUI default so creation is easy. Navigate to a host and select Ceph –> Pools –> Create. The defaults are fine and all you need do is give it a name, like “VMs”. You may notice there is already a .mgr pool. That’s created by the manager service and safe to ignore.

If you only need storage for VMs and containers, you can actually stop here. You can create VMs directly on your new pool, and containers on local storage then migrate (Server –> Container –> Resources –> Root Disk -> Volume Action –> Target Storage)

Erasure Coded Pool

Erasure coding requires that you determine how many data and parity bits to use, and issue the create command in the terminal. For the first question, it’s pretty simple - if you have three servers you’ll need 2 data and 1 parity. The more systems you have the more efficient you’ll be, though when you get to 6 you should probably increase your parity. Unlike the replicated pool, you can only loose one host with this level of protection.

Here’s the command⁵ for a 3 node system that can withstand one node loss (2,1). For a 4 node system you’d use (3,1) and so on. Increase the first number as your node count goes up, and the second as you desire more resilience. Ceph doesn’t require a ‘balanced’ cluster in terms of drives, but you’ll loose some capacity if you don’t have roughly the same amount of space on each node.

# k is for data and m is for parity
pveceph pool create POOLNAME --erasure-coding k=2,m=1 --pg_autoscale_mode on --application cephfs

Note that we specified application in the command. If you don’t, you won’t be able to use it for a filesystem later on. We also specified PGs (placement groups) as auto-scaling. This is how Ceph chunks data as it gets sent to storage. If you know how much data you have to load, you can specify the starting number of PGs with the --pg_num parameter. This will make things a little faster for an initial copy. Redhat suggests⁶ the OSD*100 / K+M. You’ll get a warning⁷ from Ceph if it’s not a power of 2 (2, 4, 8, 16, 32, 64, 128, 256, 512) so use the closest number, such as --pg_num 512.

If you don’t know how much data you’re going to bing in, but expect it to be a lot, you can turn on the bulk flag, rather than specifying pg_num.

# Make sure to add '-data' at the end
ceph osd pool set POOLNAME-data bulk true

When you look at the pools in the GUI you’ll see it also created two pools, one for data and one for metadata, which isn’t compatible with EC pools yet. You’ll also notice that you can put VMs and containers on this pool just like the replicated pool. It will just be slower.

Filesystem

The GUI won’t allow you to choose the erasure coded pool you just created so you’ll use the command line again. The name you pick for your Filesystem will be how it’s mounted.

ceph osd pool set POOLNAME-data bulk true
ceph fs new FILE-SYSTEM-NAME POOLNAME-metadata POOLNAME-data --force

To mount it cluster-wide, go back to the Web GUI and select Datacenter at the top left, then Storage. Click Add and select CephFS as the type. In the subsequent dialog box, put the name you’d like it mounted as in the ID field, such as “bulk” and leave the rest at their defaults.

You can now find it mounted at /mnt/pve/IDNAME and you can bind mount it to your Containers or setup NFS for your VMs.

Operation

Failure is Always an Option

Ceph defaults to a failure domain of ‘host’. That means you can loose a whole host with all it’s disks and continue operating. You can also lose individual disks from different hosts, but operations are NOT guaranteed. If you have two copies, as well and continue operating. After a short time, Ceph will re-establish parity as disks fail or hosts remain off-line. Should they come back, it will re-adjust. Though in both cases this can take some time.

Rebooting a host

Ceph immediately panics and starts reestablishing resilience. When the host comes back up, it starts redoing it back. This is OK, but Redhat suggests to avoid it with a few steps.

On the node you want to reboot:

sudo ceph osd set noout
sudo ceph osd set norebalance
sudo reboot

# Log back in and check that the pgmap reports all pgs as normal (active+clean). 
sudo ceph -s

# Continue on to the next node
sudo reboot
sudo ceph -s

# When done
sudo ceph osd unset noout
sudo ceph osd unset norebalance

# Perform a final status check to make sure the cluster reports HEALTH_OK:
sudo ceph status

Troubleshooting

Pool Creation Error

If you created a pool but left off the –application flag it will be set to RDP by default. You’d have to change it from RDP to CephFS like so, for both the data and metadata

ceph osd pool application enable srv-data cephfs --yes-i-really-mean-it
ceph osd pool application disable srv-data rdb --yes-i-really-mean-it
ceph osd pool application enable srv-metadata cephfs --yes-i-really-mean-it
ceph osd pool application disable srv-metadata rdb --yes-i-really-mean-it

Cluster IP Address Change

If you want to change your IP addresses, you may be able to just change the public network in the /etc/pve/ceph.conf and then destroy and recreate what it says. This worked. I don’t know if it’s good. I think the OSD cluster network needed changed also.

Based on https://www.reddit.com/r/Proxmox/comments/p7s8ne/change_ceph_network/

2.2 - Ceph Tiering

Overview

If you have a mix of workloads you should create a mix of pools. Cache tiering is out¹. So use a mix of NVMEs, SSDs, and HDDs with rules to control what pool uses what class of device.

In this example, we’ll create a replicated SSD pool for our VMs, and a erasure coded HDD pool for our content and media files.

Initial Rules

When an OSD is added, its device class is automatically assigned. The typical ones are ssd or hdd. But either way, the default config will use them all as soon as you add them. Let’s change that by creating some additional rules

Replicated Data

For replicated data, it’s as easy as creating a couple new rules and then migrating data, if you have any.

New System

If you haven’t yet created any pools, great! we can create the rules so they are available when creating pools. Add all your disks as OSDs (visit the PVE ceph menu for each server in the cluster). Then add these rules at the command line of any PVE server to update the global config.

# Add rules for both types
#
# The format is
#    ceph osd crush rule create-replicated RULENAME default host CLASSTYPE
#
ceph osd crush rule create-replicated replicated_hdd_rule default host hdd
ceph osd crush rule create-replicated replicated_ssd_rule default host ssd

And you’re done! When you create a pool in the PVE GUI, click the advanced button and choose the appropriate CRUSH rule from the drop-down. Or you can create one now while you’re at the command line.

# Create a pool for your VMs on replicated SSD. Default replication is used (so 3 copies)
#  pveceph pool create POOLNAME --crush_rule RULENAME --pg_autoscale_mode on
pveceph pool create VMs --crush_rule  replicated_ssd_rule --pg_autoscale_mode on

Existing System

With an existing system you must migrate your data. If you’ve haven’t added your SSDs yet, do so now. It will start moving data using the default rule, but we’ll apply a new rule that will take over.

# Add rules for both types
ceph osd crush rule create-replicated replicated_hdd_rule default host hdd
ceph osd crush rule create-replicated replicated_ssd_rule default host ssd

# If you've just added SSDs, apply the new rule right away to minimize the time spent waiting for data moves.
# Use the SSD or HDD rule as you prefer. In this example we're moving POOLNAME to SSDs
ceph osd pool set VMs crush_rule replicated_sdd_rule

Erasure Coded Data

On A New System

EC data is a little different. You need a profile to describe the resilience and class, and Ceph manages the CRUSH rule directly. But you can have the pveceph to do this for you.

# Create pool name 'Content' with 2 data and 1 parity. Add --application cephfs as we're using this for file storage. The --crush_rule affects the metadata pool so its on fast storage.
pveceph pool create Content --erasure-coding k=2,m=1,device-class=hdd --crush_rule  replicated_ssd_rule --pg_autoscale_mode on --application cephfs

You’ll notice separate pools for data and metadata were automatically created as the latter doesn’t support EC pools yet.

On An Existing System

Normally, you set device class as part of creating a profile and you cannot change the profile after creating the pool². However, you can change the CRUSH rule and that’s all we need for changing the class.

# Create a new profile that to base a CRUSH rule on. This one uses HDD
#  ceph osd erasure-code-profile set PROFILENAME crush-device-class=CLASS k=2 m=1
ceph osd erasure-code-profile set ec_hdd_2_1_profile crush-device-class=hdd k=2 m=1

# ceph osd crush rule create-erasure RULENAME PROFILENAME (from above)
ceph osd crush rule create-erasure erasure_hdd_rule ec_hdd_2_1_profile

# ceph osd pool set POOLNAME crush_rule RULENAME
ceph osd pool set Content-data crush_rule erasure_hdd_rule

Don’t forget about the metadata pool and it’s a good time to turn on bulk setting if you’re going to store a lot of data.

# Put the metadata pool on SSD for speed
ceph osd pool set Content-metadata crush_rule replicated_ssd_rule
ceph osd pool set Content-data bulk true

Other Notes

NVME

There’s some reports that NVMe aren’t separated from SSDs. You may need to create that class and turn off auto detection, though this is quite old information.

Investigation

When investigating a system, you may want to drill down with thees commands.

ceph osd lspools
ceph osd pool get VMs crush_rule
ceph osd crush rule dumpreplicated_ssd_rule
# or view rules with
ceph osd getcrushmap | crushtool -d -

Data Loading

The fastest way is to use a Ceph Client at the source of the data, or at least separate the interfaces.

With 1Gb NIC, one of the Ceph storage servers also connected to an external NFS and coping data to a CephFS.

12 MB/sec

Same, but reversed with NFS server itself running the Ceph client pushing the data.

103 MB/sec

Creating and Destroying

https://dannyda.com/2021/04/10/how-to-completely-remove-delete-or-reinstall-ceph-and-its-configuration-from-proxmox-ve-pve/

# Adjust letters as needed
for X in {b..h}; do pveceph osd create /dev/sd${X};done
mkdir -p /var/lib/ceph/mon/
mkdir  /var/lib/ceph/osd

# Adjust numbers as needed
for X in {16..23};do systemctl stop ceph-osd@${X}.service;done
for X in {0..7}; do umount /var/lib/ceph/osd/ceph-$X;done
for X in {a..h}; do ceph-volume lvm zap /dev/sd$X --destroy;done

2.3 - Ceph Client

This assumes you already have a working cluster and a ceph file system.

Install

You need the ceph software. You use the cephadm tool, or add the repos and packages manually. You also need to pick what version by it’s release name; ‘Octopus, Nautilus, etc’

sudo apt install software-properties-common gnupg2
wget -q -O- 'https://download.ceph.com/keys/release.asc' | sudo apt-key add -

# Discover the current release. PVE is a good place to check when at the command line
curl -s https://enterprise.proxmox.com/debian/ | grep ceph

# Note the release name after debian, 'debian-squid' in this example.
sudo apt-add-repository 'deb https://download.ceph.com/debian-squid/ bullseye main'

sudo apt update; sudo apt install ceph-common -y

#
# Alternatively 
#

curl --silent --remote-name --location https://github.com/ceph/ceph/raw/squid/src/cephadm/cephadm
chmod +x cephadm
./cephadm add-repo --release squid
./cephadm install ceph-common

Configure

On a cluster member, generate a basic conf and keyring for the client

# for a client named 'minecraft'

ceph config generate-minimal-conf > /etc/ceph/minimal.ceph.conf

ceph-authtool --create-keyring /etc/ceph/ceph.client.minecraft.keyring --gen-key -n client.minecraft

You must add file system permissions by adding lines to the bottom of the keyring, then import it to the cluster.

nano  /etc/ceph/ceph.client.minecraft.keyring

# Allowing the client to read the root and write to the subdirectory '/srv/minecraft'
caps mds = "allow rwps path=/srv/minecraft"
caps mon = "allow r"
caps osd = "allow *"

Import the keyring to the cluster and copy it to the client

ceph auth import -i /etc/ceph/ceph.client.minecraft.keyring
scp minimal.ceph.conf ceph.client.minecraft.keyring [email protected]:

On the client, copy the keyring and rename and move the basic config file.

ssh [email protected]

sudo cp ceph.client.minecraft.keyring /etc/ceph
sudo cp minimal.ceph.conf /etc/ceph/ceph.conf

Now, you may mount the filesystem

# the format is "User ID" @ "Cluster ID" . "Filesystem Name" = "/some/folder/on/the/server" "/some/place/on/the/client"
# You can get the cluster ID from your server's ceph.conf file and the filesystem name 
# with a ceph fs ls, if you don't already know it. It will be the part after name, as in "name: XXXX, ..."

sudo mount.ceph [email protected]=/srv/minecraft /mnt

You can and entry to your fstab like so

[email protected]=/srv/minecraft /mnt ceph noatime,_netdev    0       2

Troubleshooting

source mount path was not specified
unable to parse mount source: -22

You might have accidentally installed the distro’s older version of ceph. The mount notation above is based on “quincy” aka ver 17

ceph --version

  ceph version 17.2.3 (dff484dfc9e19a9819f375586300b3b79d80034d) quincy (stable)

Try an apt remove --purge ceph-common and then apt update before trying a apt install ceph-common again.

**unable to get monitor info from DNS SRV with service name: ceph-mon**

Check your client’s ceph.conf. You may not have the right file in place

**mount error: no mds server is up or the cluster is laggy**

This is likely a problem with your client file.

Sources

https://docs.ceph.com/en/quincy/install/get-packages/ https://knowledgebase.45drives.com/kb/creating-client-keyrings-for-cephfs/ https://docs.ceph.com/en/nautilus/cephfs/fstab/

3 - Gluster

Gluster is a distributed file system that supports both replicated and dispersed data.

Supporting dispersed data is a differentiating feature. Only a few can distribute the data in a erasure-coded or RAID-like fashion, making efficient use of space while providing redundancy. Have 5 cluster members? Just add one ‘parity bit’ for just a 20% overhead and you can loose a host. Add more parity if you like with the incremental cost. Other systems require you to duplicate your data for a 50% hit.

It’s also generally perceived as less complex than competitors like Ceph, as it has fewer moving parts and is focused on block storage. And since it uses native filesystems, you can always access your data directly. Redhat has ceased it’s corporate sponsorship, but the project is still quite active.

So you just need file storage and you have a lot of data, use gluster.

3.1 - Gluster on XCP-NG

Let’s set up a distributed and dispersed example cluster. We’ll XCP-NG for this. This is similar to an erasure-coded ceph pool.

Preparation

We use three hosts, each connected to a common network. With three we can disperse data enough to take one host at a time out of service. We use 4 disks on each host in this example but any number will work as long as they are all the same.

Network

Hostname Resolution

Gluster requires¹ the hosts be resolvable by hostname. Verify all the hosts can ping each other by name. You may want to create a hosts file and copy to all three to help.

If you have free ports on each server, consider using the second interface for storage, or a mesh network for better performance.

# Normal management and or guest network
192.168.1.1 xcp-ng-01.lan
192.168.1.2 xcp-ng-02.lan
192.168.1.3 xcp-ng-03.lan

# Storage network in a different subnet (if you have a second interface)
192.168.10.1 xcp-ng-01.storage.lan
192.168.10.2 xcp-ng-02.storage.lan
192.168.10.3 xcp-ng-03.storage.lan

Firewall Rules

Gluster requires a few rules; one for the daemon itself and one per ‘brick’ (drive) on the server. You can also just allow the cluster members cart-blanc access. We’ll do both examples here. Add these to all cluster members.

vi /etc/sysconfig/iptables

# Note that the last line in the existing file is a REJECT. Make sure to insert these new rules BEFORE that line.
-A RH-Firewall-1-INPUT -p tcp -s xcp-ng-01.storage.lan -j ACCEPT 
-A RH-Firewall-1-INPUT -p tcp -s xcp-ng-02.storage.lan -j ACCEPT 
-A RH-Firewall-1-INPUT -p tcp -s xcp-ng-03.storage.lan -j ACCEPT 

# Possibly for clients
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 24007:24008 -s client-01.storage.lan -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 49152:49156 -s client-01.storage.lan -j ACCEPT

service iptables restart

vi /etc/sysconfig/iptables

# The gluster daemon needs ports 24007 and 24008
# Individual bricks need ports starting at 49152. Add an additional port per brick.
# Here we have 49152-49155 open for 4 brickes. 

# TODO - test this command
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp -s xcp-ng-01.storage.lan --dport 24007:24008 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp -s xcp-ng-01.storage.lan --dport 49152:49155 -j ACCEPT

-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp -s xcp-ng-02.storage.lan --dport 24007:24008 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp -s xcp-ng-02.storage.lan --dport 49152:49155 -j ACCEPT

-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp -s xcp-ng-03.storage.lan --dport 24007:24008 -j ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp -s xcp-ng-03.storage.lan --dport 49152:49155 -j ACCEPT

Disk

Gluster works with filesystems. This is convenient because if all else fails, you still have files on disks you can access. XFS is well regarded with gluster admins, so we’ll use that.

# Install the xfs programs
yum install -y xfsprogs

# Wipe the disks before using, then format the whole disk. Repeat for each disk
wipefs -a /dev/sda
mkfs.xfs /dev/sda

Let’s mount those disks. The convention² is to put them in /data organized by volume. We’ll use ‘volume01’ later in the config, so lets use that here as well.

On each server

# For 4 disks - Note, gluster likes to call them 'bricks'
mkdir -p /data/glusterfs/volume01/brick0{1,2,3,4}
mount /dev/sda /data/glusterfs/volume01/brick01
mount /dev/sdb /data/glusterfs/volume01/brick02
mount /dev/sdc /data/glusterfs/volume01/brick03
mount /dev/sdd /data/glusterfs/volume01/brick04

Add the appropriate config to your /etc/fstab so they mount at boot

Installation

A Note About Versions

XCP-NG is CentOS 7 based and provides GlusterFS v8 in their Repo. This version went EOL in 2021. You can add the CentOS Storage Special Interest group repo to get to v9, but no current version can be installed.

# Not recommended
yum install centos-release-gluster  --enablerepo=epel,base,updates,extras

# On each host

yum install -y glusterfs-server
systemctl enable --now glusterd


# On the first host

gluster peer probe xcp-ng-02.storage.lan
gluster peer probe xcp-ng-03.storage.lan

gluster pool list

UUID                                  Hostname              State
a103d6a5-367b-4807-be93-497b06cf1614  xcp-ng-02.storage.lan Connected 
10bc7918-364d-4e4d-aa16-85c1c879963a  xcp-ng-03.storage.lan Connected 
d00ea7e3-ed94-49ed-b56d-e9ca4327cb82  localhost             Connected

# Note - localhost will always show up for the host you're running the command on

Configuration

Gluster talks about data as being distributed and dispersed.

Distributed

# Distribute data amongst 3 servers, each with a single brick
gluster volume create MyVolume server1:/brick1 server2:brick1 server3:brick1

Any time you have more that one drive, it’s distributed. That can be across different disks on the same host, or across different hosts. There is no redundancy, however, and any loss of disk is loss of data.

Disperse

# Disperse data amongst 3 bricks, each on a different server
gluster volume create MyVolume disperse server1:/brick1 server2:/brick1 server3:/brick1

Dispersed is how you build redundancy across servers. Any one of these servers or bricks can fail and the data is safe.

# Disperse data amongst 6 six bricks, but some on the same server. Problem!
gluster volume create MyVolume disperse \
  server1:/brick1 server2:/brick1 server3:/brick1
  server1:/brick2 server2:/brick2 server3:/brick2

If you try and disperse your data across multiple bricks on the same server, you’ll run into the problem of sub-optimal parity. You’ll see the error message:

Multiple bricks of a disperse volume are present on the same server. This setup is not >optimal. Bricks should be on different nodes to have best fault tolerant configuration

Distributed-Disperse

# Disperse data into 3 brick subvolumes before distributing
gluster volume create MyVolume disperse 3 \
  server1:/brick1 server2:/brick1 server3:/brick1
  server1:/brick2 server2:/brick2 server3:/brick2

By specifying disperse COUNT you tell gluster that you want to create a subvolumes every COUNT bricks. In the above example, it’s every three bricks, so two subvolumes get created from the six bricks. This ensures the parity is optimally handled as it’s distributed.

You can also take advantage of bash shell expansion like below. Each subvolume is one line, repeated for each of the 4 bricks it will be distributed across.

gluster volume create volume01 disperse 3 \
  xcp-ng-0{1..3}.storage.lan:/data/glusterfs/volume01/brick01/brick \
  xcp-ng-0{1..3}.storage.lan:/data/glusterfs/volume01/brick02/brick \
  xcp-ng-0{1..3}.storage.lan:/data/glusterfs/volume01/brick03/brick \
  xcp-ng-0{1..3}.storage.lan:/data/glusterfs/volume01/brick04/brick

Operation

Mounting and Optimizing Volumes

mount -t glusterfs xcp-ng-01.storage.lan:/volume01 /mnt

gluster volume set volume01 group metadata-cache

gluster volume set volume01 performance.readdir-ahead on 
gluster volume set volume01 performance.parallel-readdir on

gluster volume set volume01 group nl-cache 
gluster volume set volume01 nl-cache-positive-entry on

Adding to XCP-NG

mount -t glusterfs xcp-ng-01.lan:/volume01/media.2 /root/mnt2/
mkdir mnt2/xcp-ng

xe sr-create content-type=user type=glusterfs name-label=GlusterSharedStorage shared=true \
  device-config:server=xcp-ng-01.lan:/volume01/xcp-ng \
  device-config:backupservers=xcp-ng-02.lan:xcp-ng-03.lan

Scrub and Bitrot

Scrub is off by default. You can enable scrub at which point the scrub daemon will begin “signing” files³ (by calculating checksum). The file-system parity isn’t used. So if you enable and immediately begin a scrub you will see many “Skipped files” as their checksum hasn’t been calculated yet.

Client Installation

The FUSE client is recommended⁴. The docs cover a .deb based install, but you can also install from the repo. On Debian:

sudo apt install lsb-release gnupg

OS=$(lsb_release --codename --short)

# Assuming the current version of gluster is 11 
wget -O - https://download.gluster.org/pub/gluster/glusterfs/11/rsa.pub | sudo pt-key add -
echo deb [arch=amd64] https://download.gluster.org/pub/gluster/glusterfs/11/LATEST/Debian/${OS}/amd64/apt ${OS} main | sudo tee /etc/apt/sources.list.d/gluster.list
sudo apt update; sudo apt install glusterfs-client

You need quite a few options to use this successfully at boot in the fstab

192.168.3.235:/volume01 /mnt glusterfs nofail,x-systemd.automount,x-systemd.requires=network-online.target,x-systemd.device-timeout=10 0 0

How to reboot a node

You may find that your filesystem has paused during a reboot. Take a look at your network timeout and see if setting it lower helps.

https://unix.stackexchange.com/questions/452939/glusterfs-graceful-reboot-of-brick

gluster volume set volume01 network.ping-timeout 5

Using notes from https://www.youtube.com/watch?v=TByeZBT4hfQ

3.2 - Tuning

From https://docs.gluster.org/en/main/Administrator-Guide/Performance-Tuning/

sudo gluster volume set media group metadata-cache
sudo gluster volume set media performance.readdir-ahead on 
sudo gluster volume set media  performance.parallel-readdir on
sudo gluster volume set media  group nl-cache
sudo gluster volume set media  nl-cache-positive-entry on
sudo gluster volume bitrot  media enable

Also turned off atime at the brick and mount level in the etc/fstab

Some other notes in relation to syncthing https://forum.syncthing.net/t/glusterfs-vs-syncthing/24423/21?

File System Table example

cat /etc/fstab

LABEL=BRICK1 /data/brick1   xfs defaults,noatime    0   0
LABEL=BRICK1 /data/brick2   xfs defaults,noatime    0   0
shire1:/media   /mnt/gluster/media  glusterfs   defaults,noatime,_netdev    0   0

4 - Hashdeep

Hashdeep is a reasonable way to generate checksums and test integrity on file trees.

On the source side, navigate to the parent directory of the one you want to hash and use the relative paths flag so as to avoid any issue with absolute path differences between the source and destination folders.

cd /srv/some/Parent/Directory
hashdeep -j 1 -c md5 -r -l > ~/hashes DirectoryYouWantToHash

The flags we are using are:

-j 1     # Threads to use. Adjust down if you have slow IO
-c md5   # Algorythem. md5 is the fastet in testing
-r       # Recursive
-l       # Use relative paths so it's more portable

Copy the hashes text file to the destination side and repeat.

cd /srv/some/Possibly/Different/Parent/Directory
hashdeep -r -a -v -l -k ~/hashes DirectoryToCheck

The flags we are using are:

-r # Recursive, as before
-a # Archive checking mode
-v # Verbose summary
-l # Relative as before
-k # File that has the sums to use for checking

5 - LINSTOR DRBD

LINSTOR is a management layer, frequently deployed with DRBD which does the actual work of replicating a block device. It’s mostly known for being a very fast solution for virtualization clustering. It focuses on keeping a replica of a VMs disk up to date for quick migration and redundancy, with primary IO operating locally.

Install

DKMS Build Tools

https://linuxcontainers.org/incus/docs/main/howto/storage_linstor_setup/

This involves building the kernel modules from source and pulls in all the g++ and libs adding up to 575G of needed space.

Your instructed to use add-apt-repository but that’s an Ubuntu command and you can’t install it via software-properties-common anymore. So you’d next want to install from a repo provided by Linbit, but they only make part of their stack available for Trixie. To get the drbd module, You can bodge it by putting in a link to the proxmox repo that does have it all.

sudo apt update
sudo apt install curl gnupg

curl -fsSL https://packages.linbit.com/package-signing-pubkey.asc | gpg --dearmor | sudo tee /usr/share/keyrings/linbit-keyring.gpg > /dev/null

echo "deb [signed-by=/usr/share/keyrings/linbit-keyring.gpg] http://packages.linbit.com/public/ proxmox-9 drbd-9" | sudo tee /etc/apt/sources.list.d/linbit.list

You can then proceed.

# Note: If you're running the zabbly+ kernel, you'll need to add a .config link
sudo ln -s /boot/config-$(uname -r) /usr/src/linux-headers-$(uname -r)/.config

# In most cases the headers are already installed, but you can check with:
sudo apt install linux-headers-$(uname -r)

# Install the parts that are already build
sudo apt install lvm2 drbd-utils linstor-satellite

# Allow the dynamic kernel module system to do it's thing
sudo apt install drbd-dkms

Start the Satellite service on all nodes

sudo systemctl enable --now linstor-satellite

LINSTOR Controller

LINSTOR uses a single controller. On that system, run:

# Install the controller components
sudo apt install linstor-controller linstor-client python3-setuptools

# Enable the controller service
sudo systemctl enable --now linstor-controller

Configure

On the controller

TODO: I was able to add the node simply by simply omitting the IP address and entering the command “linstor node interface create shire2” and “linstor node interface create shire3”

# You will see a message about "Unsupported storage providers". This is expected as we have only installed LVM 
linstor node create server01 <server_1> --node-type combined
linstor node create server02 <server_2> --node-type satellite
linstor node create server03 <server_3> --node-type satellite

# These command should now show all nodes and relevant features
linstor node list

linstor node info

Add the drives to linstore. You may want to run a wipefs -a /dev/sdX on the nodes to preempt any error messages.

# Change the server names such as "server01" and device names as appropriate.
linstor physical-storage create-device-pool --storage-pool nvme_pool --pool-name nvme_pool lvmthin server01 /dev/nvme1n1
linstor physical-storage create-device-pool --storage-pool nvme_pool --pool-name nvme_pool lvmthin server02 /dev/nvme1n1
linstor physical-storage create-device-pool --storage-pool nvme_pool --pool-name nvme_pool lvmthin server03 /dev/nvme1n1

# Double check that they are all up.
linstor storage-pool list

Tell Incus that you have a linstor volume

# Tell Incus where the linstor controller is
incus config set storage.linstor.controller_connection=http://<server_1>:3370

# Create a storage pool named 'remote' in incus on the individual nodes
incus storage create remote linstor --target server01
incus storage create remote linstor --target server02
incus storage create remote linstor --target server03

# link the incus storeage pool to the linstore storage pool named 'nvme_pool`
incus storage create remote linstor linstor.resource_group.storage_pool=nvme_pool

Launch an instance to test

sudo incus launch images:debian/13 c2 --storage remote
Launching c2

If this command hangs, you must cancel out (the process will keep trying for a while in the background) and check to make sure all the nodes see each other with the command sudo drbdadm status at each node.

Operations

Moving Instances to the CLuster

The normal thing to do is migrate existing containers. The only gotcha there is size. Incus creates a 10G volume on the remote disk. This is LVM thin provisioned, so it won’t use it unless it needs it, but it does set an upper limit on how big the instance can get. If you’re brining in an instance larger than 10G you’ll need to change the default volume size temporarily or you’ll get a cryptic error message about rsync and other issues.

# Check your size then move with an optional parameter to set the volume size if needed.
sudo du -sh /var/lib/incus/storage-pools/local/containers/

4.5G    /var/lib/incus/storage-pools/local/containers/desktop
68G     /var/lib/incus/storage-pools/local/containers/database
...
...
# This one is under the default 10GB so nothing extra is needed
incus move desktop --storage remote

# Since this on is 68G, set the size up, move it, then unset it
incus storage set remote volume.size 100Gib
incus move database --storage remote 
incus storage unset remote volume.size 

# Some resouces suggesting setting this option, but it fails
#incus move database --storage remote -d root,size=100GiB

Verifying its running locally.

When you move your instance it lands on two nodes. The nodes are chosen by linstore’s auto-place logic and incus doesn’t get to pick. So you’ll want to check instance is running on one of those two nodes for the best speed.

Let’s first look at a case where it lined up.

Take a look in incus to see where your instance is running. Here we can see it’s running on the server “shire2”

 sudo incus list name=something
+-----------+---------+-----------------------+---------------------------------------------+-----------+-----------+----------+
|   NAME    |  STATE  |         IPV4          |                    IPV6                     |   TYPE    | SNAPSHOTS | LOCATION |
+-----------+---------+-----------------------+---------------------------------------------+-----------+-----------+----------+
| something | RUNNING | 192.168.250.16 (eth0) | fd96:bd12:59eb:1:1266:6aff:fea3:b529 (eth0) | CONTAINER | 0         | shire2   |
+-----------+---------+-----------------------+---------------------------------------------+-----------+-----------+----------+

Now let’s ask linstore where it’s disk is at.

# Get the volume UUID
sudo linstor resource-definition list --show-props Aux/Incus/name | grep something

| incus-volume-38a5a7036bdf4001a073bc559ec06212 | remote        | DRBD,STORAGE | ok    | incus-volume-something

# See where its running at
linstor resource list
╭───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
┊ ResourceName                                  ┊ Node   ┊ Layers       ┊ Usage  ┊ Conns ┊      State ┊ CreatedOn           ┊
╞═══════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════╡
┊ incus-volume-434bcfb9288d460fbfee6482dbf9670c ┊ shire1 ┊ DRBD,STORAGE ┊ Unused ┊ Ok    ┊   UpToDate ┊ 2026-03-21 20:47:02 ┊
┊ incus-volume-434bcfb9288d460fbfee6482dbf9670c ┊ shire2 ┊ DRBD,STORAGE ┊ InUse  ┊ Ok    ┊   UpToDate ┊ 2026-03-21 20:47:02 ┊
┊ incus-volume-434bcfb9288d460fbfee6482dbf9670c ┊ shire3 ┊ DRBD,STORAGE ┊ Unused ┊ Ok    ┊ TieBreaker ┊ 2026-03-21 20:47:02 ┊
╰───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯

You can see that it’s InUse and UpToDate on shire2, the same server the instance is running on.

$ linstor resource list
╭───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
┊ ResourceName                                  ┊ Node   ┊ Layers       ┊ Usage  ┊ Conns ┊      State ┊ CreatedOn           ┊
╞═══════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════╡
┊ incus-volume-38a5a7036bdf4001a073bc559ec06212 ┊ shire1 ┊ DRBD,STORAGE ┊ Unused ┊ Ok    ┊   UpToDate ┊ 2026-03-22 08:53:11 ┊
┊ incus-volume-38a5a7036bdf4001a073bc559ec06212 ┊ shire2 ┊ DRBD,STORAGE ┊ InUse  ┊ Ok    ┊   Diskless ┊ 2026-03-22 08:53:10 ┊
┊ incus-volume-38a5a7036bdf4001a073bc559ec06212 ┊ shire3 ┊ DRBD,STORAGE ┊ Unused ┊ Ok    ┊   UpToDate ┊ 2026-03-22 08:53:11 ┊
╰───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯

You’ll notice the state Diskless on shire2…where it’s useage is InUse. That’s not ideal. Let’s fix that.

# Create a replica on the same node it's runing on. It will automatically start using it when it's done
linstor resource create shire2 incus-volume-38a5a7036bdf4001a073bc559ec06212 --storage-pool ssd_pool

linstor resource list
╭──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
┊ ResourceName                                  ┊ Node   ┊ Layers       ┊ Usage  ┊ Conns ┊             State ┊ CreatedOn           ┊
╞══════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════╡
┊ incus-volume-38a5a7036bdf4001a073bc559ec06212 ┊ shire1 ┊ DRBD,STORAGE ┊ Unused ┊ Ok    ┊          UpToDate ┊ 2026-03-22 08:53:11 ┊
┊ incus-volume-38a5a7036bdf4001a073bc559ec06212 ┊ shire2 ┊ DRBD,STORAGE ┊ InUse  ┊ Ok    ┊ SyncTarget(0.81%) ┊ 2026-03-22 08:53:10 ┊
┊ incus-volume-38a5a7036bdf4001a073bc559ec06212 ┊ shire3 ┊ DRBD,STORAGE ┊ Unused ┊ Ok    ┊          UpToDate ┊ 2026-03-22 08:53:11 ┊
╰──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯

# After a while that will change to UpToDate and you can delete one of the other replicas
linstor resource delete shire3 incus-volume-38a5a7036bdf4001a073bc559ec06212

Troubleshooting

Building initial module drbd/9.3.0-1 for 6.18.8-zabbly+ grep: /lib/modules/6.18.8-zabbly+/build/.config: No such file or directory

Assuming you’ve already addressed the note about the missing .config file above, make sure you’re running the same version. If you’ve updated but not rebooted, you’ll see this error.

uname -a

If it’s trying to build a version that’s different than your running kernel, like an older version, you see a message something like

Building for 6.18.8-zabbly+ and 6.19.6-zabbly+

It can fail because you don’t have the old headers anymore. Simply purge the old kernel so it doesn’t try and build for it.

sudo apt purge *6.18.8-zabbly+*

DRBD not available.

If you check your note status you may see something like this

╭─────────────────────────────────────────────────────────────────────╮ ┊ Node ┊ DRBD ┊ LUKS ┊ NVMe ┊ Cache ┊ BCache ┊ WriteCache ┊ Storage ┊ ╞═════════════════════════════════════════════════════════════════════╡ ┊ shire1 ┊ + ┊ - ┊ - ┊ + ┊ - ┊ + ┊ + ┊ ┊ shire2 ┊ - ┊ - ┊ - ┊ + ┊ - ┊ + ┊ + ┊ ┊ shire3 ┊ - ┊ - ┊ - ┊ + ┊ - ┊ + ┊ + ┊ ╰─────────────────────────────────────────────────────────────────────╯

Only one of the nodes looks up. If you check the node that’s up it will show this:

cat /proc/drbd
version: 9.3.1 (api:2/proto:118-124)

If you’ve built the new DKMS module, but the other nodes show an older version such as version 8, try a reboot. It’s probably running default module. You may also be able to try a `modprobe -r drbd && modprobe drbd’ to load the new one

6 - MergerFS

This is a FUSE utility that merges multiple individual filesystem disks into one large filesystem. It shines when you have a random collection of disks and can’t use RAID. It also tolerates failure as even if you loose a disk, the data on the remaining are still available normally.

When combined with SnapRAID, you can even recover.

Installation

The developer recommends users install the most recent version available from the releases page as the Debian repos can be very out of date.

Navigate to https://github.com/trapexit/mergerfs/releases, download, and then:

sudo dpkg -i mergerfs*

Configuration

# First, format your disks and give each of them their own mount point
wipefs -a /dev/sda
mkfs.ext4 /dev/sda
blkid /dev/sda
... repeat for each disk and note the UUID

# Mount them all permanently
mkdir -v /mnt/disk-{1..4}
echo "UUID=c3659080-ce25-4c13-bc96-016959f1e097 /mnt/disk-1 ext4 defaults,noatime 0 0" >> /etc/fstab
echo "UUID=3cf40be6-647f-4dbe-a617-fb0a2b7b6b9e /mnt/disk-2 ext4 defaults,noatime 0 0" >> /etc/fstab
... and so on
mount --all

# Install MergerFS, create a mount point for it and add a final line to the fstab
sudo apt install mergerfs
mkdir /mnt/pool
echo "/mnt/disk* /mnt/pool fuse.mergerfs cache.files=off,category.create=mfs,func.getattr=newest,    0 0" >> /etc/fstab

According to the developer¹, the basic FUSE options above work for most cases if you’re on a recent kernel uname -r > 6.6.


cache.files	off	Let the OS handle caching, so as to not double-cache things
category.create	mfs	Space allocation - see below
func.getattr	newest	Make directory tree mod times accurate so changes are detected²

Note:

The default space allocation used to keep folder contents together. This is good for some types of content, but deep directory trees could become unbalanced and run out of space early. So we changed to mfs which is simply ‘most free space’.

The default category.create may work better now that it’s changed and the [other policies] are interesting to read. There are also [other options] for fuse.

The docs also recommend starting with dropcacheonclose=false but that defaults to false anyway, so i’ve left it out.

[[other policies]:https://trapexit.github.io/mergerfs/latest/config/functions_categories_policies/#policy-descriptions [other options]:https://trapexit.github.io/mergerfs/latest/config/options/#mount-options

7 - MooseFS

MooseFS is a great distributed filesystem that supports erasure coding for maximum efficiency. It’s also free to use, but there’s one catch. No redundancy. You’re only allowed to run one master node for free. The license is cheaper than enterprise solutions, but far from free.

But if you don’t mind doing a full shutdown when rebooting the master node, it’s a great system.

# Make sure all hosts can find the mfs master

sudo -i
echo "10.96.146.1     mfsmaster" >> /etc/hosts

# Install the repo (for Deb 13)

apt install -y gpg curl
mkdir -p /etc/apt/keyrings
curl https://repository.moosefs.com/moosefs.key | gpg -o /etc/apt/keyrings/moosefs.gpg --dearmor 
echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/moosefs.gpg] http://repository.moosefs.com/moosefs-4/apt/debian/trixie trixie main" > /etc/apt/sources.list.d/moosefs.list
apt update

# Install the Master Server on the first instance
apt install moosefs-master moosefs-cgi moosefs-cgiserv moosefs-cli

# Install the chunk master on it as well as secondary servers

## Prepare the disks


umount /data/rustfs*
mkfs.xfs  -i size=512 -n ftype=1 -L RUSTFS0 /dev/sda -f
mkfs.xfs  -i size=512 -n ftype=1 -L RUSTFS1 /dev/sdc -f

# if needed
sudo mkdir -p /data/rustfs{0..1} 
echo "LABEL=RUSTFS0 /data/rustfs0   xfs   defaults,noatime,nodiratime   0   0" >> /etc/fstab
echo "LABEL=RUSTFS1 /data/rustfs1   xfs   defaults,noatime,nodiratime   0   0" >> /etc/fstab

mount -a

apt install moosefs-chunkserver

# Configure the chunkserver

cd /etc/mfs
cp mfschunkserver.cfg.sample mfschunkserver.cfg
cp mfshdd.cfg.sample mfshdd.cfg


# Note - If there is an error above because these files are missing, you may not have the correct Repo. 
#        make sure to run `apt update` and that the moosefs repos are listed.
#        Of course, they are only full of comments by default, so worst case you can just create them yourself.
touch mfschunkserver.cfg
touch mfshdd.cfg

echo /data/rustfs0/ >> /etc/mfs/mfshdd.cfg
echo /data/rustfs1/ >> /etc/mfs/mfshdd.cfg

chown -R mfs:mfs /data/*
chmod -R 770 /data/*

mfschunkserver start

# Assuming it started, enable it at boot.
systemctl enable moosefs-chunkserver.service

# Install the client and mount
apt install moosefs-client

mkdir /mnt/mfs
mount -t moosefs mfsmaster: /mnt/mfs

sudo chown -R rustfs-user:rustfs-user /data/*
sudo systemctl restart rustfs.service

8 - NFS

NFS is the fastest way move files around a small network. It beats both samba and afp in throughput (Circa 2014) in my testing and with a little extra config works well between apple and linux.

8.1 - General Use

The NFS server supports multiple protocol versions, but we’ll focus on the current 4.X version of the protocol. It’s been out since 2010 and simplifies security.

Installation

Linux Server

This will install the server and a few requisites.

sudo  apt-get install nfs-kernel-server

Configuration

Set NFSv4 only

In order to streamline the ports needed (in case one uses firewalls) and reduce required services, we will limit the server to v4¹ only.

Edit nfs-common

sudo vi /etc/default/nfs-common

NEED_STATD=“no” NEED_IDMAPD=“yes”

And the defaults

sudo vi /etc/default/nfs-kernel-server

RPCNFSDOPTS="-N 2 -N 3" RPCMOUNTDOPTS="–manage-gids -N 2 -N 3"

Disable rpcbind

sudo systemctl mask rpcbind.service sudo systemctl mask rpcbind.socket

Create Exports

In NFS parlance, you ’export’ a folder when you share it. We’ll use the same location for our exports as suggested in the Debian example.

sudo vim /etc/exports

/srv/nfs4 192.168.1.0/24(rw,async,fsid=0,crossmnt,no_subtree_check,all_squash,anonuid=1000,anongid=1000,insecure)

         /srv/nfs4 # This is the actual folder on the server's file system you're sharing
    192.168.1.0/24 # This is the network you're sharing with
                rw # Read-Write mode
             async # Allow cached writes
            fsid=0 # This signifies this is the 'root' of the exported file system and that
                   # clients will mount this share as '/'
          crossmnt # Allow subfolders that are seperate filesystem to be accessed also
  no_subtree_check # Disable checking for access rights outside the exported file system
        all_squash # all user IDs will translated to anonymous
      anonuid=1000 # all anonymous connections will be mapped to this user account in /etc/passwd
      anongid=1000 # all anonymous connections will be mapped to this group account in /etc/passwd
          insecure # Allows macs to mount using source ports from non-root source ports

If you can’t put all your content under this folder, it’s recommended you create pseudo file system for security reasons. See the notes for more info on that, but keep things simple if you can.

Configure Host-Based Firewall

If you have a system with ufw you can get this working fairly easily. NFS is already defined as a well-known service.

ufw allow from 192.168.1.0/24 to any port nfs

Restart the Service

You don’t actually need to restart. You put your changed into effect by issuing the exportfs command. This is best practice so you don’t to disrupt currently connected clients.

exportfs -rav

Client Configuration

Apple OS X

Modern Macs support NFSv4 with a couple tweaks

# In a terminal, issue the command
sudo mount -t nfs -o nolocks,resvport,locallocks 192.168.1.2:/srv ./mnt

You can also mount in finder with a version 4 flag. That may help but is somewhat awkward

nfs://vers=4,192.168.1.5/srv/nfs4

You can also edit the mac’s config file. This will allow you to use the finder to mount NFS 4 exports.

sudo vim /etc/nfs.conf

#
# nfs.conf: the NFS configuration file
#
#nfs.client.mount.options = nolock
#
nfs.client.mount.options = vers=4.1,nolocks,resvport,locallocks

You can now hit command-k and enter the string below to connect

nfs://my.server.or.ip/

Some sources suggest editing the autofs.conf file to add ’nolocks,locallocks to the automount options. This may or may not have an effect.

sudo vim  /etc/autofs.conf
AUTOMOUNTD_MNTOPTS=nosuid,nodev,nolocks,locallocks

Troubleshooting

Must use v3

If you must use v3, you can set static ports. Use the internet for this.

lockd: cannot monitor

You may want to check your mac’s nfs options and set ’nolock’ or possibly ‘vers=4’ as above. Don’t set them both on at once as in the next issue.

mount_nfs: can’t mount / from home onto /Volumes/mnt: Invalid argument

You can’t combine -o vers=4 with options like ’nolocks’, presumably because it’s not implemented fully. This may have changed by now.

https://developer.apple.com/library/mac/documentation/Darwin/Reference/Manpages/man8/mount_nfs.8.html

No Such File or Directory mount.nfs: mounting some.ip:/srv failed, reason given by server: No such file or directory

Version 4 maps directories and starts with ‘/’. Try mounting just the root path as opposed to /srv/nfs4.

mount  -o nfsvers=4.1 some.ip:/ /srv

<There was a problem ….

Check that you have ‘insecure’ in your nfs export options on the server

/srv  192.168.1.0/24(rw,async,fsid=0,insecure,crossmnt,no_subtree_check)

Can’t create or see files

Don’t forget that file permissions apply as the user you specified above. Set chown and chmod accordingly

Can Create Files But Not Modify or Delete

Check the parent directory permissions

NFS doesn’t mount at boot

Try adding some mount [options].

some.ip:/ /srv  nfs nofail,x-systemd.automount,x-systemd.requires=network-online.target,x-systemd.device-timeout=10,vers=4.1 0 0

mount.nfs: requested NFS version or transport protocol is not supported

Try specifying the nfs version

mount  -o nfsvers=4.1 some.ip:/ /srv

https://wiki.debian.org/NFSServerSetup [options]:https://unix.stackexchange.com/questions/459731/systemd-failed-to-mount-nfs-share-mount-nfs-network-is-unreachable-until-late ↩︎

8.2 - Armbian NFS Server

This is usually a question of overhead. NFS has less CPU overhead and faster speeds circa 2023, and anecdotal testing showed fewer issues with common clients like VLC, Infuse and Kodi. However, there’s no advertisement¹ like SMB has, so you have to pre-configure all clients.

This is the basic config for an anonymous, read-only share.

apt install nfs-kernel-server

echo "/mnt/pool *(fsid=0,ro,all_squash,no_subtree_check)" >> /etc/exports

exportfs -rav

mDNS SRV records has some quasi supports, but not with common clients ↩︎

8.3 - NFS Container

This is problematic. NFS requires kernel privileges so the usual answer is “don’t”. Client’s also. So from a security and config standpoint, it’s better have PVE act as the NFS client and use bind mounts for the containers. But this can blur the line between services and infrastructure.

Either way, here’s my notes from setting up an Alpine NFS server.

Create privileged container and enable nesting

https://forum.proxmox.com/threads/is-it-possible-to-run-a-nfs-server-within-a-lxc.24403/page-2

Create a privileged container by unchecking “Unprivileged” during creation. May be possible to convert an existing container from unprivileged to privileged by backing-up and restoring. In the container Options -> Features, enable Nesting. (The NFS feature doesn’t seem necessary for running an NFS server. May be required for an NFS client - I haven’t checked

For Alpine, CAP_SETPCAP is also needed

vi /etc/pve/lxc/100.conf

# clear cap.drop
lxc.cap.drop:

# copy drop list from /usr/share/lxc/config/common.conf
lxc.cap.drop = mac_admin mac_override sys_time sys_module sys_rawio

# copy drop list from /usr/share/lxc/config/alpine.common.conf with setpcap commented

lxc.cap.drop = audit_write
lxc.cap.drop = ipc_owner
lxc.cap.drop = mknod
# lxc.cap.drop = setpcap
lxc.cap.drop = sys_nice
lxc.cap.drop = sys_pacct
lxc.cap.drop = sys_ptrace
lxc.cap.drop = sys_rawio
lxc.cap.drop = sys_resource
lxc.cap.drop = sys_tty_config
lxc.cap.drop = syslog
lxc.cap.drop = wake_alarm

Then proceed with https://wiki.alpinelinux.org/wiki/Setting_up_a_nfs-server.

9 - SnapRAID

SnapRAID protects against disk failure, but but not in the normal way. Rather than calculating parity with every write, it calculates parity on a schedule. Think of it as off-line RAID that’s like a backup. If you delete a file, you can recover it. If you loose a disk, you can rebuild it. Just not in real-time. All you need is a spare disk to save the parity to.

The natural question is: If you have an extra disk, why wouldn’t you just run RAID to begin with? The answers are usually:

You need something like backup but don’t have enough space for a full copy.
You already have a collection of individual drives without raid, and want to protect them.
You’d like to setup RAID, but don’t have the space to copy it out and back in.
Your case is full so you plug in a USB drive occasionally for a snap.
You want to send the snap to another site, like the cloud.
You have a mix of drive sizes and can’t do normal RAID.
You want to protect your existing data from bit-rot.

The developer mentions it’s targeted at large media collections, so the first few points are most common.

I say disk a lot, but I should really say filesystem, as it works at that level. If you already have a large mdadm raid0 filesystem, your parity storage will have to be as large as the whole mdadm filesystem. So mostly too late in that case.

One common use is paired with MergerFS. It has individual file systems so you can snap it easily. And MergerFS lets the (remaining) content to continue being served while the missing filesystem is recovered.

Installation

You can install from apt but on most systems it’s a major version behind. The developer recommends downloading the current version and building.

https://www.snapraid.it/download

# If you're OK with v12
apt install snapraid

# If you prefer v13 and don't mind all the space from the build tools
sudo apt-get update && sudo apt-get install gcc git make -y

git clone https://github.com/amadvance/snapraid.git
cd snapraid
./autogen.sh
./configure
make
make check
sudo make install

Configuration

SnapRAID needs a configuration file and a content file. The first is where you tell it what to do, and the second is for it to keep track of what it’s done.

It also needs a place to store parity information. This can grow as large as the largest disk. So if your largest disk is 10TB you’ll need a 10TB parity disk.

For an example, let’s assume we have 4 disks that range from 4 to 8 TB. To that we’ll add an 8TB disk to store parity. The developer’s suggests¹ this command to maximize space (though if you go past 16TB use xfs).

mkfs.ext4 -m 0 -T largefile4 DEVICE

That give us:

lsblk

NAME     MAJ:MIN  RM   SIZE RO TYPE MOUNTPOINTS
sdb        8:16    0     4T  0 disk
└─sdb5     8:21    0   3.9T  0 part /mnt/pics
sdc        8:32    0     4T  0 disk /mnt/movies
sdd        8:48    0     6T  0 disk /mnt/movies2
sde        8:64    0     8T  0 disk /mnt/tv.shows
sdf        8:64    0     8T  0 disk /mnt/snapraid   # <-- Just added this

So you’d create the config file with the following options:

sudo vi /etc/snapraid.conf

/var/snapraid/snapraid.content

data d1 /mnt/pics/
data d2 /mnt/movies/
data d3 /mnt/movies2/
data d4 /mnt/tv.shows/
parity /mnt/snapraid/snapraid.parity
# This is the 'working' file for snapraid
content /var/snapraid/snapraid.content
# A couple more copies is suggested as backup
content /mnt/pics/.snapraid.content
content /mnt/movies/.snapraid.content

Operations

Backup

The command to sync up the parity disk is just sync. This will take a long time for it’s initial run.

snapraid sync

Undelete

You can restore a deleted file or directory with:

snapraid fix -f SOMETHING

Rebuilding

If you loose a disk, just add a new one, format and mount it at the same place, and type:

# Assuming 'd2' was the replaced disk
snapraid -d d2 fix

Scrub

If your disks are suspect, you can run periodic scrubs to check for bit-rot.

snapraid scrub
snapraid status

There are may other options and the manual is worth checking out.

https://www.snapraid.it/faq#fs ↩︎

10 - sshfs

You can mount a remote system via sshfs. It’s slow, but better than nothing.

# Mount a host dir you have ssh access to
sshfs [email protected]:/var/www/html /var/www/html

# Mount a remote system over a proxy jump host
sshfs [email protected]:/var/www/html /var/www/html -o ssh_command="ssh -J [email protected]",allow_other,default_permissions

It’s often handy to add a shortcut to your ssh config so you don’t have to type as much.

cat .ssh/config 
Host some.host
    ProxyJump some.in.between.host

11 - ZFS

Overview

ZFS: the last word in file systems - at least according to Sun Microsystems back in 2004. But it’s pretty much true for traditional file servers. You add disks to a pool where you decide how much redundancy you want, then create file systems that sit on top. ZFS directly manages it all. No labeling, partitioning or formatting required.

There is error detection and correction, zero-space clones and snapshots, compression and deduplication, and delta-based replication for backup. It’s also highly resistent to corruption from power loss and crashes because it uses Copy-On-Write.

This last feature means that as files are changed, only the changed-bits are written out, and then the metadata updated at the end as a separate and final step to include these changed bits. The original file stays the same until the very end. An interruption in the middle of a write (such as from a crash) leaves the file undamaged.

Lastly, everything is checksummed and automatically repaired should you ever suffer from silent corruption.

11.1 - Basics

The Basics

Let’s create a pool and mount a file system.

Create a Pool

A ‘pool’ is group of disks and that’s where RAID levels are established. You can choose to mirror drives, or use distributed parity. A common choice is RAIDZ1 so that you can sustain the loss of one drive. You can increase the redundancy to RAIDZ2 and RAIDZ3 as desired.

zpool create pool01 raidz1 /dev/sdc /dev/sdd /dev/sde /dev/sdf

zpool list
NAME               SIZE  ALLOC   FREE  
pool01              40T     0T    40T

Create a Filesystem

A default root file system is created and mounted automatically when the pool is created. Using this is fine and you can easily change where it’s mounted.

zfs list

NAME     USED  AVAIL     REFER  MOUNTPOINT
pool01     0T    40T        0T  /pool01

zfs set mountpoint=/mnt pool01

But often, you need more than one file system - say one for holding working files and another for long term storage. This allows you to easily back up things different things, differently.

# Get rid of the initial fs and pool
zpool destroy pool01

# Create it again but leave the root filesystem unmounted with the `-m` option. You can also use drive short-names
zpool create -m none pool01 raidz1 sdc sdd sde sdf

# Add a couple filesystems and mount under the /srv directory
zfs create pool01/working -o  mountpoint=/srv/working
zfs create pool01/archive -o  mountpoint=/srv/archive             
              ^     ^
             /       \
pool name --          filesystem name

Now you can do things like snapshot the archive folder regularly while skipping the working folder. The only downside is that they are separate filessytems. Moving things doesn’t happen instantly.

Compression

Compression is on by default and this will save space for things that can benefit from it. It also makes things faster as moving compressed data takes less time. CPU use for the default algorithm lz4, is negligible and it quickly detects files that aren’t compressible and gives up so that CPU time isn’t wasted.

zfs get compression pool01

NAME    PROPERTY     VALUE           SOURCE
pool01  compression  lz4             local

Next Step

Now you’ve got a filesystem, take a look at creating and work with snapshots.

11.2 - Snapshot

Create a Snapshot

About to do something dangerous? Let’s create a ‘save’ point so you can reload your game, so to speak. They don’t take any space (to start with) and are nearly instant.

# Create a snapshot named `save-1`
zfs snapshot pool01/archive@save-1

The snapshot is a read-only copy of the filesystem at that time. It’s mounted by default in a hidden directory and you can examine and even copy things out of it, if you desire.

ls /srv/archive/.zfs/save-1

Delete a Snapshot

While a snapshot doesn’t take up any space to start with, it begins to as you make changes. Anything you delete stays around on the snapshot. Things you edit consume space as well for the changed bits. So when you’re done with a snapshot, it’s easy to remove.

zfs snapshot destroy pool01/archive@save-1

Rollback to a Snapshot

Mess things up in your archive folder after all? Not a problem, just roll back to the same state as your snapshot.

zfs rollback pool01/archive@save-1

Importantly, this is a one-way trip back. You can’t branch and jump around like it was a filesystem multiverse of alternate possibilities. ZFS will warn you about this and if you have more than one snapshot in between you and where you’re going, it will let you know they are about to be deleted.

Auto Snapshot

One of the most useful tools is the zfs-auto-snapshot utility. This will create periodic snapshots of your filesystem and keep them pruned for efficiency. By default, it creates a snapshot every 15 min, and then prunes them down so you have one:

Every 15 min for an hour
Every hour for a day
Every day for a week
Every week for month
Every month

Install with the command:

sudo apt install zfs-auto-snapshot

That’s it. You’ll see new folders based on time created in the hidden .zfs folder at the root of your filesystems. Each filesystem will get it’s own. Anytime you need to look for a file you’ve deleted, you’ll find it there.

# Look at snapshots
ls /srv/archive/.zfs/

Excluding datasets from auto snapshot

# Disable
zfs set com.sun:auto-snapshot=false rpool/export

Excluding frequent or other auto-snapshot

There are sub-properties you can set under the basic auto-snapshot value

zfs set com.sun:auto-snapshot=true pool02/someDataSet
zfs set com.sun:auto-snapshot:frequent=false  pool02/someDataSet

zfs get com.sun:auto-snapshot pool02/someDataSet
zfs get com.sun:auto-snapshot:frequent pool02/someDataSet

# Possibly also the number to keep if other than the default is desired
zfs set com.sun:auto-snapshot:weekly=true,keep=52

# Take only weekly
zfs set com.sun:auto-snapshot:weekly=true rpool/export

Deleting Lots of auto-snapshot files

You can’t use globbing or similar to mass-delete snapshots, but you can string together a couple commands.

# Disable auto-snap as needed
zfs set com.sun:auto-snapshot=false pool04
zfs list -H -o name -t snapshot | grep auto | xargs -n1 zfs destroy

Missing Auto Snapshots

On some centOS-based systems, like XCP-NG, you will only see frequent snapshots. This is because only the frequent cron job uses the correct path. You must add a PATH statement to the other cron jobs

https://forum.level1techs.com/t/setting-up-zfs-auto-snapshot-on-centos-7/129574/12

Next Step

Now that you have snapshots, let’s send them somewhere for backup with replication.

References

https://www.reddit.com/r/zfs/comments/829v5a/zfs_ubuntu_1604_delete_snapshots_with_wildcard/

11.3 - Replication

Replication is how you backup and copy ZFS. It turns a snapshot into a bit-stream that you can pipe to something else. Usually, you pipe it over the network to another system where you connect it to zfs receive.

It is also the only way. The snapshot is what allows point-in-time handling and the receive ensures consistency. And a snapshot is a filesystem, but it’s more than just the files. Two identically named filesystems with the same files you put in place by rsync are not the same filesystem and you can’t jump-start a sync this way.

Basic Examples

# On the receiving side, create a pool to hold the filesystem
zpool create -m none pool02 raidz1 sdc sdd sde sdf

# On the sending side, pipe over SSH. The -F forces the filesystem on the receiving side to be replaced
zfs snapshot pool01@snap1
zfs send pool01@snap1 | ssh some.other.server zfs receive -F pool02

Importantly, this replaces the root filesystem on the receiving side. The filesystem you just copied over is accessible when the replication is finished - assuming it’s mounted and your only using the default root. If you’re using multiple filesystems, you’ll want to recursively send things so you can pick up children like the archive filesystem.

# The -r and -R trigger recursive operations
zfs snapshot -r pool01@snap1
zfs send -R pool01@snap1 | ssh some.other.server zfs receive -F pool02

You can also pick a specific filesystem to send. You can name it whatever you like on the other side, or replace something already named.

# Sending just the archive filesystem
zfs snapshot pool01/archive@snap1
zfs send pool01/archive@snap1 | ssh some.other.server zfs receive -F pool02/archive

And of course, you may have two pools on the same system. One line in a terminal is all you need.

zfs send -R pool01@snap1 | zfs receive -F pool02

Using Mbuffer or Netcat

These are much faster than ssh if you don’t care about someone capturing the traffic. But it does require you to start both ends separately.

# On the receiving side
ssh some.other.system
mbuffer -4 -s 128k -m 1G -I 8990 | zfs receive -F pool02/archive

# On the sending side
zfs send pool01/archive@snap1 | mbuffer -s 128k -m 1G -O some.other.system:8990

You can also use good-ole netcat. It’s a little slower but still faster than SSH. Combine it with pv for some visuals.

# On the receiving end
nc -l 8989 | pv -trab -B 500M | zfs recv -F pool02/archive

# On the sending side
zfs send pool01/archive@snap1 | nc some.other.system 8989

Estimating The Size

You may want to know how big the transmit is to estimate time or capacity. You can do this with a dry-run.

zfs send -nv pool01/archive@snap1

Use a Resumable Token

Any interruptions and you have to start all over again - or do you? If you’re sending a long-running transfer, add a token on the receiving side and you can restart from where it broke, turning a tragedy into just an annoyance.

# On the receiving side, add -s
ssh some.other.system
mbuffer -4 -s 128k -m 1G -I 8990 | zfs receive -s -F pool01/archive

# Send the stream normally
zfs send pool01/archive@snap1 | mbuffer -s 128k -m 1G -O some.other.system:8990

# If you get interrupted, on the receiving side, look up the token
zfs get -H -o value receive_resume_token pool01

# Then use that on the sending side to resume where you left off
zfs send -t 'some_long_key' | mbuffer -s 128k -m 1G -O some.other.system:8990

If you decide you don’t want to resume, clean up with the -A command to release the space consumed by the pending transfer.

# On the receiving side
zfs recv -A pool01/archive

Sending an Incremental Snapshot

After you’ve sent the initial snapshot, subsequent ones are much smaller. Even very large backups can be kept current up if you ‘pre-seed’ before taking the other pool remote.

# at some point in past
zfs snapshot pool01/archive@snap1
zfs send pool01/archive@snap1 ....

# now we'll snap again and send just the changes between the two using  -i 
zfs snapshot pool01/archive@snap2
zfs send -i pool01/archive/@snap1 pool01/archive@snap2 ...

Sending Intervening Snapshots

If you are jumping more than one snapshot ahead, the intervening ones are skipped. If you want to include them for some reason, use the -I option.

# This will send all the snaps between 1 and 9
zfs send -I pool01/archive/@snap1 pool01/archive@snap9 ...

Changes Are Always Detected

You’ll often need to use -F to force changes as even though you haven’t used the remote system, it may think you have if it’s mounted and atimes are on.

You must have a snapshot in common

You need at least 1 snapshot in common at both locations. This must have been sent from one to the other, not just named the same. Say you create snap1 and send it. Later, you create snap2 and, thinking you don’t need it anymore, delete snap1. Even though you have snap1 on the destination you cannot send snap2 as a delta. You need snap1 in both locations to generate the delta. You are out of luck and must send snap2 as a full snapshot.

You can use a zfs feature called a bookmark as an alternative, but that is something you set up in advance and won’t save you from the above scenario.

A Full Example of an Incremental Send

Take a snapshot, estimate the size with a dry-run, then use a resumable token and force changes

zfs snapshot pool01/archive@12th
zfs send -nv -i pool01/archive@11th pool01/archive@12th
zfs send -i pool01/archive@11th pool01/archive@12th | pv -trab -B 500M | ssh some.other.server zfs recv -F -s pool01/archive

Here’s an example of a recursive snapshot to a file. The snapshot takes a -r for recursive, and the send a -R.

# Take the snapshot
zfs snapshot -r pool01/archive@21

# Estimate the size
zfs send -nv -R -i pool01/archive@20 pool01/archive@21

# Mount a portable drive to /mnt - a traditional ext4 or whatever
zfs send -Ri pool01/archive@20 pool01/archive@21 > /mnt/backupfile

# When you get to the other location, receive from the file
zfs receive -s -F pool02/archive < /mnt/backupfile

Sending The Whole Pool

This is a common thought when you’re starting, but it ends up being deceptive because pools aren’t things that can be sent. Only filesets. So what you’re actually doing is a recursive send of the root fileset with a implicit snapshot created on the fly. This is find, but you won’t be able to refer to it later for updates, so you’re better off not.

# Unmount -a (all available filesystems) on the given pool
zfs unmount -a pool01

# Send the unmounted filesystem with an implicit snapshot
zfs send -R pool01 ...

Auto Replication

You don’t want to do this by hand all the time. One way is with a simple script. If you’ve already installed zfs-auto-snapshot you may have something that looks like this:

# use '-o name' to get just the snapshot name without all the details 
# use '-s creation' to sort by creation time
zfs list -t snapshot -o name -s creation pool01/archive

pool01/archive@auto-2024-10-13_00-00
pool01/archive@auto-2024-10-20_00-00
pool01/archive@auto-2024-10-27_00-00
pool01/archive@auto-2024-11-01_00-00
pool01/archive@auto-2024-11-02_00-00

You can get the last two like this, then use send and receive. Adjust the grep to get just the daily as needed.

CURRENT=$(zfs list -t snapshot -o name -s creation pool01/replication | grep auto | tail -1 )
LAST=$(zfs list -t snapshot -o name -s creation pool01/replication | grep auto | tail -2 | head -1)

zfs send -i $LAST $CURRENT | pv -trab -B 500M | ssh some.other.server zfs recv -F -s pool01/archive

This is pretty basic and can fall out of sync. You can bring it up a notch by asking the other side to list it’s snapshots with a zfs list over ssh and comparing against yours to find the most recent match. And add a resume token. But by that point you may consider just using a tool like:

Zrepl

This looks to replace your auto snapshots as well, and that’s probably fine. I have’t used it myself as I scripted back in the day. I will probably start, though.

Next Step

Sometimes, good disks go bad. Learn how to catch them before they do, and replace them when needed.

Scrub and Replacement

11.4 - Disk Replacement

Your disks will fail, but you’ll usually get some warnings because ZFS proactively checks every occupied bit to guard against silent corruption. This is normally done every month, but you can launch one manually if you’re suspicious. They take a long time, but operate at low priority.

# Start a scrub
zpool scrub pool01

# Check the status
zpool scrub -s pool01

You can check the status of your pool at anytime with the command zpool status. When there’s a problem, you’ll see this:

zpool status

NAME            STATE     READ WRITE CKSUM
pool01          DEGRADED     0     0     0
  raidz1        DEGRADED     0     0     0
    /dev/sda    ONLINE       0     0     0
    /dev/sdb    ONLINE       0     0     0
    /dev/sdc    ONLINE       0     0     0
    /dev/sdd    FAULTED     53     0     0  too many errors

Time to replace that last drive before it goes all the way bad

# You don't need to manually offline the drive if it's faulted, but it's good practice to as there's other states it can be in
zpool offline pool01 /dev/sdd

# Physically replace that drive. If you're shutting down to do this, the replacement usually has the same device path
zpool replace pool01 /dev/sdd

There’s a lot of strange things that can happen with drives and depending on your version of ZFS it might be using UUIDs or other drive identification strings. Check the link below for some of those conditions.

zpool

11.5 - Large Pools

You’ve guarded against disk failure this by adding redundancy, but was it enough? There’s a very mathy calculator at https://jro.io/r2c2/ that will allow you chart different parity configs. But a reasonable rule-of-thumb is to devote 20%, or 1 in 5 drives, to parity.

RAIDZ1 - up to 5 Drives
RAIDZ2 - up to 10 Drives
RAIDZ3 - up to 15 Drives

Oracle however, recommends Virtual Devices when you go past 9 disks.

Pools and Virtual Devices

When you get past 15 drives, you can’t increase parity. You can however, create virtual devices. Best practice from Oracle says to do this even earlier as a VDev should be less than 9 disks ¹. So given 24 disks, you should have 3 VDevs of 8 each. Here’s an example with 2 parity. Slightly better than 1 in 5 and suitable for older disks.

### Build a 3-Wide RAIDZ2  across 24 disks
zpool create \
  pool01 \
  -m none \
  -f \
  raidz2 sdb sdc sdd sde sdf sdg sdh sdi \
  raidz2 sdj sdk sdl sdm sdn sdo sdp sdq \
  raidz2 sdr sds sdt sdu sdv sdw sdx sdy

Using Disk IDs

Drive letters can be hard to trace back to a physical drive. A a better² way is the /dev/disk/by-id identifiers.

ls /dev/disk/by-id | grep ata | grep -v part

zpool create -m none -o ashift=12 -O compression=lz4 \
  pool04 \
    raidz2 \
      ata-ST4000NM0035-1V4107_ZC11AHH9 \
      ata-ST4000NM0035-1V4107_ZC116F11 \
      ata-ST4000NM0035-1V4107_ZC1195V5 \
      ata-ST4000NM0035-1V4107_ZC11CDMB \
      ata-ST4000NM0035-1V4107_ZC1195PR \
      ata-ST4000NM0024-1HT178_Z4F164WG \
      ata-ST4000NM0024-1HT178_Z4F17SJK \
      ata-ST4000NM0024-1HT178_Z4F17M6B \
      ata-ST4000NM0024-1HT178_Z4F18FZE \
      ata-ST4000NM0024-1HT178_Z4F18G35

Hot and Distributed Spares

Spares vs Parity

You may not be reach a location quickly when a disk fails. In such a case, is it better to have a Z3 filesystem run in degraded performance mode (i.e. calculating parity the whole time) or a Z2 system that replaces the failed disk automatically?

It’s better to have a more parity until you go past the guidelines of 15 drives in a Z3 config. If you have 16 bays, add a hot spare.

Distributed vs Dedicated

A distributed spare is a newer feature that allows you reserve space on all of your disks, rather than just one. That allows resilvering to go much faster as you’re no longer limited by the speed of one disk. Here’s an example of such a pool that has 16 total devices.

# This pool has 3 parity, 12 data, 16 total count, with 1 spare
```bash
zpool create -f pool02 \
  draid3:12d:16c:1s \
    ata-ST4000NM000A-2HZ100_WJG04M27 \
    ata-ST4000NM000A-2HZ100_WJG09BH7 \
    ata-ST4000NM000A-2HZ100_WJG0QJ7X \
    ata-ST4000NM000A-2HZ100_WS20ECCD \
    ata-ST4000NM000A-2HZ100_WS20ECFH \
    ata-ST4000NM000A-2HZ100_WS20JXTA \
    ata-ST4000NM0024-1HT178_Z4F14K76 \
    ata-ST4000NM0024-1HT178_Z4F17SJK \
    ata-ST4000NM0024-1HT178_Z4F17YBP \
    ata-ST4000NM0024-1HT178_Z4F1BJR1 \
    ata-ST4000NM002A-2HZ101_WJG0GBXB \
    ata-ST4000NM002A-2HZ101_WJG11NGC \
    ata-ST4000NM0035-1V4107_ZC1168N3 \
    ata-ST4000NM0035-1V4107_ZC116F11 \
    ata-ST4000NM0035-1V4107_ZC116MSW \
    ata-ST4000NM0035-1V4107_ZC116NZM \

References

11.6 - Pool Testing

Best Practices

Best practice from Oracle says a VDev should be less than 9 disks¹. So given 24 disks, you should have 3 VDevs. They further recommend the following amount of parity vs data:

single-parity starting at 3 disks (2+1)
double-parity starting at 6 disks (4+2)
triple-parity starting at 9 disks (6+3)

It is not recommended to create a zpool with a single large vdev, say 20 disks, because write IOPS performance will be that of a single disk, which also means that resilver time will be very long (possibly weeks with future large drives).

Reasons For These Practices

I interpret this as meaning that when a single IO write operation is given to the VDev, it won’t write anything else until it’s done. But if you have multiple VDevs, you can hand out a writes to other VDevs while you’re waiting on the first. Reading is probably unaffected, but writes will be faster with more VDevs.

Also, when resilvering the array, you have to read from each of the drives in the VDev to calculate the parity bit. If you have 24 drives in a VDev, then you have to read a block of data from all 24 drives to produce the parity bit. If you have only 8, then you have only 1/3 as much data to read. Meanwhile, the rest of the VDevs are available for real work.

Rebuilding the array also introduces stress which can cause other disks to fail, so it’s best to limit that to a smaller set of drives. I’ve heard many times of resilvering causing sister drives that were already on the edge, to go over and fail the array.

Calculating Failure Rates

You can calculate the failure rates of different configurations with an on-line tool². The chart scales the X axis by 50, so the differences in failure rates are not as large as it would seem, but if they didn’t you wouldn’t be able to see the lines. But in most cases, there’s not a large difference say between a 4x9 and a 3x12.

When To Use a Hot Spare

Given 9 disks where one fails, is it better to drop from 3 parity to 2 and run in degraded mode, or 2 parity that drops to 1 and a spare that recovers without intervention. The math² says its better to have parity. But what about speed? When you loose a disk, 1 out of every 9 IOPS requires that you reconstruct it from parity. Anecdotally, observed performance penalties are minor. So the only times to use a hot spare is:

When you have unused capacity in RAIDZ3 (i.e. almost never)
When IOPS require a mirror pool

Say you have 16 bays of 4TB Drives. A 2x8 Z2 config gives you 48TB but you only want 32TB. Change that to a 2x8 Z3 and get 40TB. Still only need 32 TB? Change that to a 2x7 Z3 with 2 hot spares. Now you have 32TB with the maximum protection and the insurance of an automatic replacement.

Or maybe you have a 37 bay system. You do something that equals 36 plus a spare.

The other case is when your IOPS demands push past what RAIDZ can do an you must use a mirror pool. A failure there looses all redundancy and a hot spare is your only option.

When To Use a Distributed Spare

A distributed spare recovers in half the time³ from a disk loss, and is always better than a dedicated spare - though you should almost never use a spare anyway. The only time to use a normal hot spare is when you have a single global spare.

Testing Speed

The speed difference isn’t charted. So let’s test that some.

Given 24 disks, and deciding to live dangerously, should you should have a single, 24 disk vdev with three parity disks, or three VDevs with a single parity disk each? The reason for the 1st case is better resiliency, and the latter better write speed and recovery from disk failures.

Build a 3-Wide RAIDZ1

Create the pool across 24 disks

zpool create \
-f -m /srv srv \
raidz sdb sdc sdd sde sdf sdg sdh sdi \
raidz sdj sdk sdl sdm sdn sdo sdp sdq \
raidz sdr sds sdt sdu sdv sdw sdx sdy

Now copy a lot of random data to it

#!/bin/bash

no_of_files=1000
counter=0
while [[ $counter -le $no_of_files ]]
 do echo Creating file no $counter
   touch random-file.$counter
   shred -n 1 -s 1G random-file.$counter
   let "counter += 1"
 done

Now yank (literally) one of the physical disks and replace it

allen@server:~$ sudo zpool status  
                                                                             
  pool: srv                                                                                              
 state: DEGRADED                                                                                         
status: One or more devices could not be used because the label is missing or                            
        invalid.  Sufficient replicas exist for the pool to continue                                     
        functioning in a degraded state.                                                                 
action: Replace the device using 'zpool replace'.                                                        
   see: http://zfsonlinux.org/msg/ZFS-8000-4J                                                            
  scan: none requested                                                                                   
config:                                                                                                  
                                                                                                         
        NAME                     STATE     READ WRITE CKSUM                                              
        srv                      DEGRADED     0     0     0                                              
          raidz1-0               DEGRADED     0     0     0                                              
            sdb                  ONLINE       0     0     0                                              
            6847353731192779603  UNAVAIL      0     0     0  was /dev/sdc1                               
            sdd                  ONLINE       0     0     0                                              
            sde                  ONLINE       0     0     0                                              
            sdf                  ONLINE       0     0     0                                              
            sdg                  ONLINE       0     0     0                                              
            sdh                  ONLINE       0     0     0                                              
            sdi                  ONLINE       0     0     0                                              
          raidz1-1               ONLINE       0     0     0                                              
            sdr                  ONLINE       0     0     0                                              
            sds                  ONLINE       0     0     0                                              
            sdt                  ONLINE       0     0     0                                              
            sdu                  ONLINE       0     0     0                                              
            sdj                  ONLINE       0     0     0                                              
            sdk                  ONLINE       0     0     0                                              
            sdl                  ONLINE       0     0     0                                              
            sdm                  ONLINE       0     0     0  
          raidz1-2               ONLINE       0     0     0  
            sdv                  ONLINE       0     0     0  
            sdw                  ONLINE       0     0     0  
            sdx                  ONLINE       0     0     0  
            sdy                  ONLINE       0     0     0  
            sdn                  ONLINE       0     0     0  
            sdo                  ONLINE       0     0     0  
            sdp                  ONLINE       0     0     0  
            sdq                  ONLINE       0     0     0             
                                                                           
errors: No known data errors

allen@server:~$ lsblk                                                                                    
NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT                                                              
sda      8:0    0 465.8G  0 disk                                                                         
├─sda1   8:1    0 449.9G  0 part /                                                                       
├─sda2   8:2    0     1K  0 part                                                                         
└─sda5   8:5    0  15.9G  0 part [SWAP]                                                                  
sdb      8:16   1 931.5G  0 disk                                                                         
├─sdb1   8:17   1 931.5G  0 part                                                                         
└─sdb9   8:25   1     8M  0 part                                                                         
sdc      8:32   1 931.5G  0 disk                                                                         
sdd      8:48   1 931.5G  0 disk                                                                         
├─sdd1   8:49   1 931.5G  0 part                                                                         
└─sdd9   8:57   1     8M  0 part    
...


sudo zpool replace srv 6847353731192779603 /dev/sdc -f

allen@server:~$ sudo zpool status
  pool: srv
 state: DEGRADED
status: One or more devices is currently being resilvered.  The pool will
        continue to function, possibly in a degraded state.
action: Wait for the resilver to complete.
  scan: resilver in progress since Fri Mar 22 15:50:21 2019
    131G scanned out of 13.5T at 941M/s, 4h7m to go
    5.40G resilvered, 0.95% done
config:

        NAME                       STATE     READ WRITE CKSUM
        srv                        DEGRADED     0     0     0
          raidz1-0                 DEGRADED     0     0     0
            sdb                    ONLINE       0     0     0
            replacing-1            OFFLINE      0     0     0
              6847353731192779603  OFFLINE      0     0     0  was /dev/sdc1/old
              sdc                  ONLINE       0     0     0  (resilvering)
            sdd                    ONLINE       0     0     0
            sde                    ONLINE       0     0     0
            sdf                    ONLINE       0     0     0
...

A few hours later…

$ sudo zpool status


  pool: srv
 state: DEGRADED
status: One or more devices has experienced an error resulting in data corruption.  Applications may be affected.
action: Restore the file in question if possible.  Otherwise restore the entire pool from backup.
   see: http://zfsonlinux.org/msg/ZFS-8000-8A
  scan: resilvered 571G in 5h16m with 2946 errors on Fri Mar 22 21:06:48 2019
config:

    NAME                       STATE     READ WRITE CKSUM
    srv                        DEGRADED   208     0 2.67K
        raidz1-0               DEGRADED   208     0 5.16K
        sdb                    ONLINE       0     0     0
        replacing-1            OFFLINE      0     0     0
          6847353731192779603  OFFLINE      0     0     0  was /dev/sdc1/old
        sdc                    ONLINE       0     0     0
        sdd                    ONLINE     208     0     1
        sde                    ONLINE       0     0     0
        sdf                    ONLINE       0     0     0
        sdg                    ONLINE       0     0     0
        sdh                    ONLINE       0     0     0
        sdi                    ONLINE       0     0     0
      raidz1-1                 ONLINE       0     0     0
        sdr                    ONLINE       0     0     0
        sds                    ONLINE       0     0     0
        sdt                    ONLINE       0     0     0
        sdu                    ONLINE       0     0     0
        sdj                    ONLINE       0     0     1
        sdk                    ONLINE       0     0     1
        sdl                    ONLINE       0     0     0
        sdm                    ONLINE       0     0     0
      raidz1-2                 ONLINE       0     0     0
        sdv                    ONLINE       0     0     0
        sdw                    ONLINE       0     0     0
        sdx                    ONLINE       0     0     0
        sdy                    ONLINE       0     0     0
        sdn                    ONLINE       0     0     0
        sdo                    ONLINE       0     0     0
        sdp                    ONLINE       0     0     0
        sdq                    ONLINE       0     0     0

The time was 5h16m. But notice the error - during resilvering drive sdd had 208 read errors and data was lost. This is the classic RAID situation where resilvering stresses the drives, another goes bad and you can’t restore.

It’s somewhat questionable if this is a valid test as the affect of the error on resilvering duration is unknown. But on with the test.

Let’s wipe that away and create a raidz3

sudo zpool destroy srv


zpool create \
-f -m /srv srv \
raidz3 \
 sdb sdc sdd sde sdf sdg sdh sdi \
 sdj sdk sdl sdm sdn sdo sdp sdq \
 sdr sds sdt sdu sdv sdw sdx sdy



zdb
zpool offline srv 15700807100581040709
sudo zpool replace srv 15700807100581040709 sdc




allen@server:~$ sudo zpool status
  pool: srv
 state: DEGRADED
status: One or more devices is currently being resilvered.  The pool will continue to function, possibly in a degraded state.
action: Wait for the resilver to complete.
  scan: resilver in progress since Sun Mar 24 10:07:18 2019
    27.9G scanned out of 9.14T at 362M/s, 7h19m to go
    1.21G resilvered, 0.30% done
config:

    NAME             STATE     READ WRITE CKSUM
    srv              DEGRADED     0     0     0
      raidz3-0       DEGRADED     0     0     0
        sdb          ONLINE       0     0     0
        replacing-1  OFFLINE      0     0     0
          sdd        OFFLINE      0     0     0
          sdc        ONLINE       0     0     0  (resilvering)
        sde          ONLINE       0     0     0
        sdf          ONLINE       0     0     0
        ...


allen@server:~$ sudo zpool status

  pool: srv
 state: ONLINE
  scan: resilvered 405G in 6h58m with 0 errors on Sun Mar 24 17:05:50 2019
config:

    NAME        STATE     READ WRITE CKSUM
    srv         ONLINE       0     0     0
      raidz3-0  ONLINE       0     0     0
        sdb     ONLINE       0     0     0
        sdc     ONLINE       0     0     0
        sde     ONLINE       0     0     0
        sdf     ONLINE       0     0     0
        sdg     ONLINE       0     0     0
        ...

The time? 6h58m. Longer, but safer.

11.7 - ZFS Cache

Metadata Cache

There is a lot out there about ZFS cache config. I’ve found the most significant feature to be putting your metadata on a dedicated NVMe devices. This is noted as a ‘Special’ VDev. Here’s an example of a draid with such a device at the end.

Note: A 2x18 is bad practice - just more fun than a 3x12 with no spares.

zpool create -f pool02 \
    draid3:14d:18c:1s \
        ata-ST4000NM000A-2HZ100_WJG04M27 \
        ata-ST4000NM000A-2HZ100_WJG09BH7 \
        ata-ST4000NM000A-2HZ100_WJG0QJ7X \
        ata-ST4000NM000A-2HZ100_WS20ECCD \
        ata-ST4000NM000A-2HZ100_WS20ECFH \
        ata-ST4000NM000A-2HZ100_WS20JXTA \
        ata-ST4000NM0024-1HT178_Z4F14K76 \
        ata-ST4000NM0024-1HT178_Z4F17SJK \
        ata-ST4000NM0024-1HT178_Z4F17YBP \
        ata-ST4000NM0024-1HT178_Z4F1BJR1 \
        ata-ST4000NM002A-2HZ101_WJG0GBXB \
        ata-ST4000NM002A-2HZ101_WJG11NGC \
        ata-ST4000NM0035-1V4107_ZC1168N3 \
        ata-ST4000NM0035-1V4107_ZC116F11 \
        ata-ST4000NM0035-1V4107_ZC116MSW \
        ata-ST4000NM0035-1V4107_ZC116NZM \
        ata-ST4000NM0035-1V4107_ZC118WV5 \
        ata-ST4000NM0035-1V4107_ZC118WW0 \
    draid3:14d:18c:1s \
        ata-ST4000NM0035-1V4107_ZC118X74 \
        ata-ST4000NM0035-1V4107_ZC118X90 \
        ata-ST4000NM0035-1V4107_ZC118XBS \
        ata-ST4000NM0035-1V4107_ZC118Z23 \
        ata-ST4000NM0035-1V4107_ZC11907W \
        ata-ST4000NM0035-1V4107_ZC1192GG \
        ata-ST4000NM0035-1V4107_ZC1195PR \
        ata-ST4000NM0035-1V4107_ZC1195V5 \
        ata-ST4000NM0035-1V4107_ZC1195ZJ \
        ata-ST4000NM0035-1V4107_ZC11AHH9 \
        ata-ST4000NM0035-1V4107_ZC11CDD0 \
        ata-ST4000NM0035-1V4107_ZC11CE77 \
        ata-ST4000NM0035-1V4107_ZC11CV5E \
        ata-ST4000NM0035-1V4107_ZC11D2AQ \
        ata-ST4000NM0035-1V4107_ZC11HRGR \
        ata-ST4000NM0035-1V4107_ZC1B200R \
        ata-ST4000NM0035-1V4107_ZC1CBXEH \
        ata-ST4000NM0035-1V4107_ZC1DC98B \
    special mirror \
        ata-MICRON_M510DC_MTFDDAK960MBP_164614A1DBC4 \
        ata-MICRON_M510DC_MTFDDAK960MBP_170615BD4A74

zfs set special_small_blocks=64K pool02

Metadata is stores automatically on the special device but there’s a benefit in also directing the pool to use the special vdev for small files as well.

Sources

https://forum.level1techs.com/t/zfs-metadata-special-device-z/159954

11.8 - ZFS Encryption

You might want to store data such that it’s encrypted at rest. Or replicate data to such as system. ZFS offers this on a per-dataset option.

Create an Encrypted Fileset

Let’s assume that you’re at a remote site and want to create an encrypted fileset to receive your replications.

zfs create -o encryption=on -o keylocation=prompt -o keyformat=passphrase pool02/encrypted

Replicating to an Encrypted Fileset

This example uses mbuffer and assumes a secure VPN. Replace with SSH as needed.

# On the receiving side
sudo zfs load-key -r pool02/encrypted
mbuffer -4 -s 128k -m 1G -I 8990 | sudo zfs receive -s -F pool02/encrypted

# On the sending side
zfs send -i pool01/archive@snap1 pool01/archived@snap2 | mbuffer -s 128k -m 1G -O some.server:8990

11.9 - VDev Sizing

Best practice from Oracle says a VDev should be less than 9 disks¹. So given 24 disks you should have 3 VDevs. However, when using RAIDZ, the math shows they should be as large as possible with multiple parity disks². I.e. with 24 disks you should have a single, 24 disk VDev.

The reason for the best practice seems to be about the speed of writing and recovering from disk failures.

It is not recommended to create a zpool with a single large vdev, say 20 disks, because write IOPS performance will be that of a single disk, which also means that resilver time will be very long (possibly weeks with future large drives).

With a single VDev, you break up the data to send a chunk to each drive, then wait for them all to finish writing before you send the next. With several VDevs, you can move on to the next while you wait for the others to finish.

Build a 3-Wide RAIDZ1

Create the pool across 24 disks

#
# -O is the pool's root dataset. Lowercase letter -o is for pool properties
# sudo zfs get compression to check. lz4 is now prefered
#
zpool create \
-m /srv srv \
-O compression=lz4 \
raidz sdb sdc sdd sde sdf sdg sdh sdi \
raidz sdj sdk sdl sdm sdn sdo sdp sdq \
raidz sdr sds sdt sdu sdv sdw sdx sdy -f

Copy a lot of random data to it.

#!/bin/bash

no_of_files=1000
counter=0
while [[ $counter -le $no_of_files ]]
 do echo Creating file no $counter
   touch random-file.$counter
   shred -n 1 -s 1G random-file.$counter
   let "counter += 1"
 done

Yank out (literally) one of the physical disks and replace it.

sudo zpool status                                                               [433/433]

  pool: srv                                                                                              
 state: DEGRADED                                                                                         
status: One or more devices could not be used because the label is missing or                            
        invalid.  Sufficient replicas exist for the pool to continue                                     
        functioning in a degraded state.                                                                 
action: Replace the device using 'zpool replace'.                                                        
   see: http://zfsonlinux.org/msg/ZFS-8000-4J                                                            
  scan: none requested                                                                                   
config:                                                                                                  
                                                                                                         
        NAME                     STATE     READ WRITE CKSUM                                              
        srv                      DEGRADED     0     0     0                                              
          raidz1-0               DEGRADED     0     0     0                                              
            sdb                  ONLINE       0     0     0                                              
            6847353731192779603  UNAVAIL      0     0     0  was /dev/sdc1                               
            sdd                  ONLINE       0     0     0                                              
            sde                  ONLINE       0     0     0                                              
            sdf                  ONLINE       0     0     0                                              
            sdg                  ONLINE       0     0     0                                              
            sdh                  ONLINE       0     0     0                                              
            sdi                  ONLINE       0     0     0                                              
          raidz1-1               ONLINE       0     0     0                                              
            sdr                  ONLINE       0     0     0                                              
            sds                  ONLINE       0     0     0                                              
            sdt                  ONLINE       0     0     0                                              
            sdu                  ONLINE       0     0     0                                              
            sdj                  ONLINE       0     0     0                                              
            sdk                  ONLINE       0     0     0                                              
            sdl                  ONLINE       0     0     0                                              
            sdm                  ONLINE       0     0     0  
          raidz1-2               ONLINE       0     0     0  
            sdv                  ONLINE       0     0     0  
            sdw                  ONLINE       0     0     0  
            sdx                  ONLINE       0     0     0  
            sdy                  ONLINE       0     0     0  
            sdn                  ONLINE       0     0     0  
            sdo                  ONLINE       0     0     0  
            sdp                  ONLINE       0     0     0  
            sdq                  ONLINE       0     0     0             
                                                                           
errors: No known data errors

Insert a new disk and replace the missing one.

allen@server:~$ lsblk                                                                                    
NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT                                                              
sda      8:0    0 465.8G  0 disk                                                                         
├─sda1   8:1    0 449.9G  0 part /                                                                       
├─sda2   8:2    0     1K  0 part                                                                         
└─sda5   8:5    0  15.9G  0 part [SWAP]                                                                  
sdb      8:16   1 931.5G  0 disk                                                                         
├─sdb1   8:17   1 931.5G  0 part                                                                         
└─sdb9   8:25   1     8M  0 part                                                                         
sdc      8:32   1 931.5G  0 disk                       # <-- new disk showed up here
sdd      8:48   1 931.5G  0 disk                                                                         
├─sdd1   8:49   1 931.5G  0 part                                                                         
└─sdd9   8:57   1     8M  0 part    
...


sudo zpool replace srv 6847353731192779603 /dev/sdc -f

sudo zpool status

  pool: srv
 state: DEGRADED
status: One or more devices is currently being resilvered.  The pool will
        continue to function, possibly in a degraded state.
action: Wait for the resilver to complete.
  scan: resilver in progress since Fri Mar 22 15:50:21 2019
    131G scanned out of 13.5T at 941M/s, 4h7m to go
    5.40G resilvered, 0.95% done
config:

        NAME                       STATE     READ WRITE CKSUM
        srv                        DEGRADED     0     0     0
          raidz1-0                 DEGRADED     0     0     0
            sdb                    ONLINE       0     0     0
            replacing-1            OFFLINE      0     0     0
              6847353731192779603  OFFLINE      0     0     0  was /dev/sdc1/old
              sdc                  ONLINE       0     0     0  (resilvering)
            sdd                    ONLINE       0     0     0
            sde                    ONLINE       0     0     0
            sdf                    ONLINE       0     0     0
...

We can see it’s running at 941M/s. Not too bad.

Build a 1-Wide RAIDZ3

sudo zpool destroy srv

zpool create \
-m /srv srv \
-O compression=lz4 \
raidz3 \
 sdb sdc sdd sde sdf sdg sdh sdi \
 sdj sdk sdl sdm sdn sdo sdp sdq \
 sdr sds sdt sdu sdv sdw sdx sdy -f

Copy a lot of random data to it again (as above)

Replace a disk (as above)

allen@server:~$ sudo zpool status
  pool: srv
 state: DEGRADED
status: One or more devices is currently being resilvered.  The pool will
 continue to function, possibly in a degraded state.
action: Wait for the resilver to complete.
  scan: resilver in progress since Sun Mar 24 10:07:18 2019
    27.9G scanned out of 9.14T at 362M/s, 7h19m to go
    1.21G resilvered, 0.30% done
config:

 NAME             STATE     READ WRITE CKSUM
 srv              DEGRADED     0     0     0
   raidz3-0       DEGRADED     0     0     0
     sdb          ONLINE       0     0     0
     replacing-1  OFFLINE      0     0     0
       sdd        OFFLINE      0     0     0
       sdc        ONLINE       0     0     0  (resilvering)
     sde          ONLINE       0     0     0
     sdf          ONLINE       0     0     0

So that’s running quite a bit slower. Not exactly 1/3, but closer to it than not.

Surprise Ending

That was all about speed. What about reliability?

Our first resilver was going a lot faster, but it ended badly. Other errors popped up, some on the same VDev as was being resilvered, and so it failed.

sudo zpool status

  pool: srv
 state: DEGRADED
status: One or more devices has experienced an error resulting in data
 corruption.  Applications may be affected.
action: Restore the file in question if possible.  Otherwise restore the
 entire pool from backup.
   see: http://zfsonlinux.org/msg/ZFS-8000-8A
  scan: resilvered 571G in 5h16m with 2946 errors on Fri Mar 22 21:06:48 2019
config:

 NAME                       STATE     READ WRITE CKSUM
 srv                        DEGRADED   208     0 2.67K
   raidz1-0                 DEGRADED   208     0 5.16K
     sdb                    ONLINE       0     0     0
     replacing-1            OFFLINE      0     0     0
       6847353731192779603  OFFLINE      0     0     0  was /dev/sdc1/old
       sdc                  ONLINE       0     0     0
     sdd                    ONLINE     208     0     1
     sde                    ONLINE       0     0     0
     sdf                    ONLINE       0     0     0
     sdg                    ONLINE       0     0     0
     sdh                    ONLINE       0     0     0
     sdi                    ONLINE       0     0     0
   raidz1-1                 ONLINE       0     0     0
     sdr                    ONLINE       0     0     0
     sds                    ONLINE       0     0     0
     sdt                    ONLINE       0     0     0
     sdu                    ONLINE       0     0     0
     sdj                    ONLINE       0     0     1
     sdk                    ONLINE       0     0     1
     sdl                    ONLINE       0     0     0
     sdm                    ONLINE       0     0     0
   raidz1-2                 ONLINE       0     0     0
     sdv                    ONLINE       0     0     0
     sdw                    ONLINE       0     0     0
     sdx                    ONLINE       0     0     0
     sdy                    ONLINE       0     0     0
     sdn                    ONLINE       0     0     0
     sdo                    ONLINE       0     0     0
     sdp                    ONLINE       0     0     0
     sdq                    ONLINE       0     0     0

errors: 2946 data errors, use '-v' for a list

Our second resilver was going very slowly, but did slow and sure when the race? It did, but very very slowly.

allen@server:~$ sudo zpool status
[sudo] password for allen: 
  pool: srv
 state: ONLINE
  scan: resilvered 405G in 6h58m with 0 errors on Sun Mar 24 17:05:50 2019
config:

 NAME        STATE     READ WRITE CKSUM
 srv         ONLINE       0     0     0
   raidz3-0  ONLINE       0     0     0
     sdb     ONLINE       0     0     0
     sdc     ONLINE       0     0     0
     sde     ONLINE       0     0     0
     sdf     ONLINE       0     0     0
     sdg     ONLINE       0     0     0
      ...
      ...

It slowed even further down, as 400G in 7 hours is something like a 16M/s. I didn’t see any checksum errors this time, but that time is abysmal.

Though, to paraphrase Livy, better late than never.

11.10 - ZFS Replication Script

#!/usr/bin/env bash
#
# zfs-pull.sh
#
# Pulls incremental ZFS snapshots from a remote (source) server to the local (destination) server.
# Uses snapshots made by zfs-auto-snapshot. Locates the latest snapshot common to both sides
# to perform an incremental replication; if none is found, it does a full send.
#
# Usage: replicate-zfs-pull.sh <SOURCE_HOST> <SOURCE_DATASET> <DEST_DATASET>
#
# Example:
#   ./replicate-zfs-pull.sh mysourcehost tank/mydata tank/backup/mydata
#
# Assumptions/Notes:
#   - The local server is the destination. The remote server is the source.
#   - We're using "zfs recv -F" locally, which can forcibly roll back the destination 
#     dataset if it has diverging snapshots. Remove or change -F as desired.
#   - This script is minimal and doesn't handle advanced errors or timeouts gracefully.
#   - Key-based SSH authentication should be set up so that `ssh <SOURCE_HOST>` doesn't require a password prompt.
#

set -euo pipefail

##############################################################################
# 1. Parse command-line arguments
##############################################################################
if [[ $# -ne 3 ]]; then
  echo "Usage: $0 <SOURCE_HOST> <SOURCE_DATASET> <DEST_DATASET>"
  exit 1
fi

SOURCE_HOST="$1"
SOURCE_DATASET="$2"
DEST_DATASET="$3"

##############################################################################
# 2. Gather snapshot lists
#
#    The command zfs list -H -t snapshot -o name -S creation -d 1
#          -H           : Output without headers for script-friendliness
#          -t snapshot  : Only list snapshots
#          -o name      : Only list the name
#          -d 1         : Only descend one level - i.e. don't tree out child datasets
##############################################################################
# - Remote (source) snapshots: via SSH to the remote host
# - Local (destination) snapshots: from the local ZFS

echo "Collecting snapshots from remote source: ${SOURCE_HOST}:${SOURCE_DATASET}..."
REMOTE_SNAPSHOTS=$(ssh "${SOURCE_HOST}" zfs list -H -t snapshot -o name -d 1 "${SOURCE_DATASET}" 2>/dev/null \
  | grep "${SOURCE_DATASET}@" \
  | awk -F'@' '{print $2}' || true)

echo "Collecting snapshots from local destination: ${DEST_DATASET}..."
LOCAL_SNAPSHOTS=$(zfs list -H -t snapshot -o name -d 1 "${DEST_DATASET}" 2>/dev/null \
  | grep "${DEST_DATASET}@" \
  | awk -F'@' '{print $2}' || true)

##############################################################################
# 3. Find the latest common snapshot
#
#   The snapshots names have prefixes like "zfs-auto-snap_daily" and "zfs-auto-snap_hourly"
#   that confuse sorting for the linux comm program, so we strip the prefix with sed before 
#   using 'comm -12' to find common elements of input 1 and 2, and tail to get the last one.
#  
COMMON_SNAPSHOT=$(comm -12 <(echo "$REMOTE_SNAPSHOTS" | sed 's/zfs-auto-snap_\w*-//' | sort) <(echo "$LOCAL_SNAPSHOTS" | sed 's/zfs-auto-snap_\w*-//' | sort) | tail -n 1)

# We need the full name back for the transfer, so grep it out of the local list. Make sure to quote the variable sent to grep or you'll loose the newlines.
COMMON_SNAPSHOT=$(echo "$LOCAL_SNAPSHOTS" | grep $COMMON_SNAPSHOT)

if [[ -n "$COMMON_SNAPSHOT" ]]; then
  echo "Found common snapshot: $COMMON_SNAPSHOT"
else
  echo "No common snapshot found—will perform a full send."
fi


##############################################################################
# 4. Identify the most recent snapshot on the remote source
#
#   This works because we zfs list'ed the snapshots originally in order
#   so we can just take the first line with 'head -n 1'
##############################################################################
LATEST_REMOTE_SNAPSHOT=$(echo "$REMOTE_SNAPSHOTS" | head -n 1)

if [[ -z "$LATEST_REMOTE_SNAPSHOT" ]]; then
  echo "No snapshots found on the remote source. Check if zfs-auto-snapshot is enabled there."
  exit 1
fi

##############################################################################
# 5. Perform replication
##############################################################################
echo "Starting pull-based replication from ${SOURCE_HOST}:${SOURCE_DATASET} to local ${DEST_DATASET}..."

if [[ -n "$COMMON_SNAPSHOT" ]]; then
  echo "Performing incremental replication from @$COMMON_SNAPSHOT up to @$LATEST_REMOTE_SNAPSHOT."
  ssh "${SOURCE_HOST}" zfs send -I "${SOURCE_DATASET}@${COMMON_SNAPSHOT}" "${SOURCE_DATASET}@${LATEST_REMOTE_SNAPSHOT}" \
    | zfs recv -F "${DEST_DATASET}"
else
  echo "Performing full replication of @$LATEST_REMOTE_SNAPSHOT."
  ssh "${SOURCE_HOST}" zfs send "${SOURCE_DATASET}@${LATEST_REMOTE_SNAPSHOT}" \
    | zfs recv -F "${DEST_DATASET}"
fi

echo "Replication completed successfully!"