debian-mirror-gitlab/doc/administration/nfs.md

457 lines
22 KiB
Markdown
Raw Normal View History

2020-10-24 23:57:45 +05:30
---
2022-07-23 23:45:48 +05:30
stage: Systems
2021-02-22 17:27:13 +05:30
group: Distribution
2022-11-25 23:54:43 +05:30
info: To determine the technical writer assigned to the Stage/Group associated with this page, see https://about.gitlab.com/handbook/product/ux/technical-writing/#assignments
2020-10-24 23:57:45 +05:30
type: reference
---
2021-09-04 01:27:46 +05:30
# Using NFS with GitLab **(FREE SELF)**
2020-10-24 23:57:45 +05:30
NFS can be used as an alternative for object storage but this isn't typically
2021-09-04 01:27:46 +05:30
recommended for performance reasons.
2020-10-24 23:57:45 +05:30
2023-01-13 00:05:48 +05:30
For data objects such as LFS, Uploads, and Artifacts, an [Object Storage service](object_storage.md)
2020-10-24 23:57:45 +05:30
is recommended over NFS where possible, due to better performance.
2022-03-02 08:16:31 +05:30
When eliminating the usage of NFS, there are [additional steps you need to take](object_storage.md#other-alternatives-to-file-system-storage)
in addition to moving to Object Storage.
2020-10-24 23:57:45 +05:30
2021-03-11 19:13:27 +05:30
File system performance can impact overall GitLab performance, especially for
2021-01-29 00:20:46 +05:30
actions that read or write to Git repositories. For steps you can use to test
2021-03-11 19:13:27 +05:30
file system performance, see
2021-09-04 01:27:46 +05:30
[File System Performance Benchmarking](operations/filesystem_benchmarking.md).
2020-10-24 23:57:45 +05:30
2023-03-04 22:38:38 +05:30
## Gitaly with NFS not supported
2021-04-29 21:17:54 +05:30
2023-03-04 22:38:38 +05:30
Technical and engineering support for using NFS to store Git repository data is officially at end-of-life. No product
changes or troubleshooting is provided through engineering, security or paid support channels.
2021-11-18 22:05:49 +05:30
2023-03-04 22:38:38 +05:30
If an issue is reproducible, or if it happens intermittently but regularly, GitLab Support can investigate providing the
issue can be reproduced without NFS. To reproduce without NFS, migrate the affected repositories to a different Gitaly
shard. For example, a Gitaly Cluster or a standalone Gitaly VM, backed with block storage.
2021-04-29 21:17:54 +05:30
2020-10-24 23:57:45 +05:30
## Known kernel version incompatibilities
RedHat Enterprise Linux (RHEL) and CentOS v7.7 and v7.8 ship with kernel
2022-08-27 11:52:29 +05:30
version `3.10.0-1127`, which [contains a bug](https://bugzilla.redhat.com/show_bug.cgi?id=1783554) that causes
2020-10-24 23:57:45 +05:30
[uploads to fail to copy over NFS](https://gitlab.com/gitlab-org/gitlab/-/issues/218999). The
following GitLab versions include a fix to work properly with that
kernel version:
2020-11-24 15:15:51 +05:30
- [12.10.12](https://about.gitlab.com/releases/2020/06/25/gitlab-12-10-12-released/)
- [13.0.7](https://about.gitlab.com/releases/2020/06/25/gitlab-13-0-7-released/)
- [13.1.1](https://about.gitlab.com/releases/2020/06/24/gitlab-13-1-1-released/)
- 13.2 and up
2020-10-24 23:57:45 +05:30
If you are using that kernel version, be sure to upgrade GitLab to avoid
errors.
## Fast lookup of authorized SSH keys
The [fast SSH key lookup](operations/fast_ssh_key_lookup.md) feature can improve
performance of GitLab instances even if they're using block storage.
[Fast SSH key lookup](operations/fast_ssh_key_lookup.md) is a replacement for
`authorized_keys` (in `/var/opt/gitlab/.ssh`) using the GitLab database.
NFS increases latency, so fast lookup is recommended if `/var/opt/gitlab`
is moved to NFS.
We are investigating the use of
[fast lookup as the default](https://gitlab.com/groups/gitlab-org/-/epics/3104).
2021-12-11 22:18:48 +05:30
## Improving NFS performance with GitLab
NFS performance with GitLab can in some cases be improved with
2022-10-11 01:57:18 +05:30
[direct Git access](gitaly/index.md#direct-access-to-git-in-gitlab) using [Rugged](https://github.com/libgit2/rugged).
2021-12-11 22:18:48 +05:30
2022-10-11 01:57:18 +05:30
Depending on the GitLab version, GitLab [automatically detects](gitaly/index.md#automatic-detection) if Rugged can and should
be used per storage.
2022-08-27 11:52:29 +05:30
2022-10-11 01:57:18 +05:30
If the Rugged feature flag is explicitly set to either `true` or `false`, GitLab uses the value explicitly set. If you
previously enabled Rugged using the feature flag and you want to use automatic detection instead, you must unset
the feature flag:
2021-12-11 22:18:48 +05:30
```shell
sudo gitlab-rake gitlab:features:unset_rugged
```
2022-10-11 01:57:18 +05:30
From GitLab 12.7, Rugged is only automatically enabled for use with Puma if the
[Puma thread count is set to `1`](../install/requirements.md#puma-settings). To use Rugged with a Puma thread count of
more than `1`, enable Rugged using the [feature flag](../development/gitaly.md#legacy-rugged-code).
2021-12-11 22:18:48 +05:30
2020-10-24 23:57:45 +05:30
## NFS server
Installing the `nfs-kernel-server` package allows you to share directories with
the clients running the GitLab application:
```shell
sudo apt-get update
sudo apt-get install nfs-kernel-server
```
### Required features
**File locking**: GitLab **requires** advisory file locking, which is only
supported natively in NFS version 4. NFSv3 also supports locking as long as
Linux Kernel 2.6.5+ is used. We recommend using version 4 and do not
specifically test NFSv3.
### Recommended options
When you define your NFS exports, we recommend you also add the following
options:
2023-04-23 21:23:45 +05:30
- `no_root_squash` - NFS usually changes the `root` user to `nobody`. This is
2021-06-08 01:23:25 +05:30
a good security measure when NFS shares are accessed by many different
users. However, in this case only GitLab uses the NFS share so it
2020-10-24 23:57:45 +05:30
is safe. GitLab recommends the `no_root_squash` setting because we need to
manage file permissions automatically. Without the setting you may receive
2021-10-27 15:23:28 +05:30
errors when the Omnibus package tries to alter permissions. GitLab
2020-10-24 23:57:45 +05:30
and other bundled components do **not** run as `root` but as non-privileged
users. The recommendation for `no_root_squash` is to allow the Omnibus package
to set ownership and permissions on files, as needed. In some cases where the
`no_root_squash` option is not available, the `root` flag can achieve the same
result.
- `sync` - Force synchronous behavior. Default is asynchronous and under certain
circumstances it could lead to data loss if a failure occurs before data has
synced.
Due to the complexities of running Omnibus with LDAP and the complexities of
maintaining ID mapping without LDAP, in most cases you should enable numeric UIDs
and GIDs (which is off by default in some cases) for simplified permission
management between systems:
- [NetApp instructions](https://library.netapp.com/ecmdocs/ECMP1401220/html/GUID-24367A9F-E17B-4725-ADC1-02D86F56F78E.html)
2021-09-30 23:02:18 +05:30
- For non-NetApp devices, disable NFSv4 `idmapping` by performing opposite of [enable NFSv4 idmapper](https://wiki.archlinux.org/title/NFS#Enabling_NFSv4_idmapping)
2020-10-24 23:57:45 +05:30
### Disable NFS server delegation
We recommend that all NFS users disable the NFS server delegation feature. This
is to avoid a [Linux kernel bug](https://bugzilla.redhat.com/show_bug.cgi?id=1552203)
which causes NFS clients to slow precipitously due to
[excessive network traffic from numerous `TEST_STATEID` NFS messages](https://gitlab.com/gitlab-org/gitlab-foss/-/issues/52017).
To disable NFS server delegation, do the following:
1. On the NFS server, run:
```shell
echo 0 > /proc/sys/fs/leases-enable
sysctl -w fs.leases-enable=0
```
1. Restart the NFS server process. For example, on CentOS run `service nfs restart`.
2021-02-22 17:27:13 +05:30
NOTE:
2020-10-24 23:57:45 +05:30
The kernel bug may be fixed in
[more recent kernels with this commit](https://github.com/torvalds/linux/commit/95da1b3a5aded124dd1bda1e3cdb876184813140).
Red Hat Enterprise 7 [shipped a kernel update](https://access.redhat.com/errata/RHSA-2019:2029)
on August 6, 2019 that may also have resolved this problem.
You may not need to disable NFS server delegation if you know you are using a version of
the Linux kernel that has been fixed. That said, GitLab still encourages instance
administrators to keep NFS server delegation disabled.
## NFS client
The `nfs-common` provides NFS functionality without installing server components which
we don't need running on the application nodes.
```shell
apt-get update
apt-get install nfs-common
```
### Mount options
Here is an example snippet to add to `/etc/fstab`:
```plaintext
2021-04-29 21:17:54 +05:30
10.1.0.1:/var/opt/gitlab/.ssh /var/opt/gitlab/.ssh nfs4 defaults,vers=4.1,hard,rsize=1048576,wsize=1048576,noatime,nofail,_netdev,lookupcache=positive 0 2
10.1.0.1:/var/opt/gitlab/gitlab-rails/uploads /var/opt/gitlab/gitlab-rails/uploads nfs4 defaults,vers=4.1,hard,rsize=1048576,wsize=1048576,noatime,nofail,_netdev,lookupcache=positive 0 2
10.1.0.1:/var/opt/gitlab/gitlab-rails/shared /var/opt/gitlab/gitlab-rails/shared nfs4 defaults,vers=4.1,hard,rsize=1048576,wsize=1048576,noatime,nofail,_netdev,lookupcache=positive 0 2
10.1.0.1:/var/opt/gitlab/gitlab-ci/builds /var/opt/gitlab/gitlab-ci/builds nfs4 defaults,vers=4.1,hard,rsize=1048576,wsize=1048576,noatime,nofail,_netdev,lookupcache=positive 0 2
10.1.0.1:/var/opt/gitlab/git-data /var/opt/gitlab/git-data nfs4 defaults,vers=4.1,hard,rsize=1048576,wsize=1048576,noatime,nofail,_netdev,lookupcache=positive 0 2
2020-10-24 23:57:45 +05:30
```
2020-11-24 15:15:51 +05:30
You can view information and options set for each of the mounted NFS file
systems by running `nfsstat -m` and `cat /etc/fstab`.
2020-10-24 23:57:45 +05:30
Note there are several options that you should consider using:
| Setting | Description |
| ------- | ----------- |
| `vers=4.1` |NFS v4.1 should be used instead of v4.0 because there is a Linux [NFS client bug in v4.0](https://gitlab.com/gitlab-org/gitaly/-/issues/1339) that can cause significant problems due to stale data.
| `nofail` | Don't halt boot process waiting for this mount to become available
| `lookupcache=positive` | Tells the NFS client to honor `positive` cache results but invalidates any `negative` cache results. Negative cache results cause problems with Git. Specifically, a `git push` can fail to register uniformly across all NFS clients. The negative cache causes the clients to 'remember' that the files did not exist previously.
| `hard` | Instead of `soft`. [Further details](#soft-mount-option).
2021-04-29 21:17:54 +05:30
| `cto` | `cto` is the default option, which you should use. Do not use `nocto`. [Further details](#nocto-mount-option).
2021-06-08 01:23:25 +05:30
| `_netdev` | Wait to mount file system until network is online. See also the [`high_availability['mountpoint']`](https://docs.gitlab.com/omnibus/settings/configuration.html#only-start-omnibus-gitlab-services-after-a-given-file-system-is-mounted) option.
2020-10-24 23:57:45 +05:30
#### `soft` mount option
It's recommended that you use `hard` in your mount options, unless you have a specific
reason to use `soft`.
2021-12-11 22:18:48 +05:30
When GitLab.com used NFS, we used `soft` because there were times when we had NFS servers
2020-10-24 23:57:45 +05:30
reboot and `soft` improved availability, but everyone's infrastructure is different.
If your NFS is provided by on-premise storage arrays with redundant controllers,
for example, you shouldn't need to worry about NFS server availability.
The NFS man page states:
> "soft" timeout can cause silent data corruption in certain cases
Read the [Linux man page](https://linux.die.net/man/5/nfs) to understand the difference,
and if you do use `soft`, ensure that you've taken steps to mitigate the risks.
If you experience behavior that might have been caused by
writes to disk on the NFS server not occurring, such as commits going missing,
use the `hard` option, because (from the man page):
> use the soft option only when client responsiveness is more important than data integrity
Other vendors make similar recommendations, including
2022-06-21 17:19:12 +05:30
[System Applications and Products in Data Processing (SAP)](https://wiki.scn.sap.com/wiki/display/Basis/Recommended+mount+options+for+read-write+directories) and NetApp's
2020-10-24 23:57:45 +05:30
[knowledge base](https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/What_are_the_differences_between_hard_mount_and_soft_mount),
they highlight that if the NFS client driver caches data, `soft` means there is no certainty if
writes by GitLab are actually on disk.
Mount points set with the option `hard` may not perform as well, and if the
2021-06-08 01:23:25 +05:30
NFS server goes down, `hard` causes processes to hang when interacting with
2020-10-24 23:57:45 +05:30
the mount point. Use `SIGKILL` (`kill -9`) to deal with hung processes.
The `intr` option
[stopped working in the 2.6 kernel](https://access.redhat.com/solutions/157873).
2021-04-29 21:17:54 +05:30
#### `nocto` mount option
Do not use `nocto`. Instead, use `cto`, which is the default.
When using `nocto`, the dentry cache is always used, up to `acdirmax` seconds (attribute cache time) from the time it's created.
This results in stale dentry cache issues with multiple clients, where each client can see a different (cached)
version of a directory.
From the [Linux man page](https://linux.die.net/man/5/nfs), the important parts:
2022-08-13 15:12:31 +05:30
> If the `nocto` option is specified, the client uses a non-standard heuristic to determine when files on the server have changed.
2021-04-29 21:17:54 +05:30
>
2022-08-13 15:12:31 +05:30
> Using the `nocto` option may improve performance for read-only mounts, but should be used only if the data on the server changes only occasionally.
2021-04-29 21:17:54 +05:30
We have noticed this behavior in an issue about [refs not found after a push](https://gitlab.com/gitlab-org/gitlab/-/issues/326066),
where newly added loose refs can be seen as missing on a different client with a local dentry cache, as
[described in this issue](https://gitlab.com/gitlab-org/gitlab/-/issues/326066#note_539436931).
2020-10-24 23:57:45 +05:30
### A single NFS mount
It's recommended to nest all GitLab data directories within a mount, that allows automatic
restore of backups without manually moving existing data.
```plaintext
mountpoint
└── gitlab-data
├── builds
├── git-data
├── shared
└── uploads
```
2021-06-08 01:23:25 +05:30
To do so, configure Omnibus with the paths to each directory nested
2020-10-24 23:57:45 +05:30
in the mount point as follows:
Mount `/gitlab-nfs` then use the following Omnibus
configuration to move each data location to a subdirectory:
```ruby
git_data_dirs({"default" => { "path" => "/gitlab-nfs/gitlab-data/git-data"} })
gitlab_rails['uploads_directory'] = '/gitlab-nfs/gitlab-data/uploads'
gitlab_rails['shared_path'] = '/gitlab-nfs/gitlab-data/shared'
gitlab_ci['builds_directory'] = '/gitlab-nfs/gitlab-data/builds'
```
2021-01-29 00:20:46 +05:30
Run `sudo gitlab-ctl reconfigure` to start using the central location. Be aware
2021-06-08 01:23:25 +05:30
that if you had existing data, you need to manually copy or rsync it to
2021-01-29 00:20:46 +05:30
these new locations, and then restart GitLab.
2020-10-24 23:57:45 +05:30
### Bind mounts
Alternatively to changing the configuration in Omnibus, bind mounts can be used
to store the data on an NFS mount.
Bind mounts provide a way to specify just one NFS mount and then
bind the default GitLab data locations to the NFS mount. Start by defining your
2023-04-23 21:23:45 +05:30
single NFS mount point as you typically would in `/etc/fstab`. Let's assume your
2020-10-24 23:57:45 +05:30
NFS mount point is `/gitlab-nfs`. Then, add the following bind mounts in
`/etc/fstab`:
```shell
/gitlab-nfs/gitlab-data/git-data /var/opt/gitlab/git-data none bind 0 0
/gitlab-nfs/gitlab-data/.ssh /var/opt/gitlab/.ssh none bind 0 0
/gitlab-nfs/gitlab-data/uploads /var/opt/gitlab/gitlab-rails/uploads none bind 0 0
/gitlab-nfs/gitlab-data/shared /var/opt/gitlab/gitlab-rails/shared none bind 0 0
/gitlab-nfs/gitlab-data/builds /var/opt/gitlab/gitlab-ci/builds none bind 0 0
```
2021-06-08 01:23:25 +05:30
Using bind mounts requires you to manually make sure the data directories
2020-10-24 23:57:45 +05:30
are empty before attempting a restore. Read more about the
[restore prerequisites](../raketasks/backup_restore.md).
### Multiple NFS mounts
2021-06-08 01:23:25 +05:30
When using default Omnibus configuration you need to share 4 data locations
2020-10-24 23:57:45 +05:30
between all GitLab cluster nodes. No other locations should be shared. The
following are the 4 locations need to be shared:
| Location | Description | Default configuration |
| -------- | ----------- | --------------------- |
2021-06-08 01:23:25 +05:30
| `/var/opt/gitlab/git-data` | Git repository data. This accounts for a large portion of your data | `git_data_dirs({"default" => { "path" => "/var/opt/gitlab/git-data"} })`
2020-10-24 23:57:45 +05:30
| `/var/opt/gitlab/gitlab-rails/uploads` | User uploaded attachments | `gitlab_rails['uploads_directory'] = '/var/opt/gitlab/gitlab-rails/uploads'`
2023-01-13 00:05:48 +05:30
| `/var/opt/gitlab/gitlab-rails/shared` | Objects such as build artifacts, GitLab Pages, LFS objects, and temp files. If you're using LFS this may also account for a large portion of your data | `gitlab_rails['shared_path'] = '/var/opt/gitlab/gitlab-rails/shared'`
2020-10-24 23:57:45 +05:30
| `/var/opt/gitlab/gitlab-ci/builds` | GitLab CI/CD build traces | `gitlab_ci['builds_directory'] = '/var/opt/gitlab/gitlab-ci/builds'`
Other GitLab directories should not be shared between nodes. They contain
node-specific files and GitLab code that does not need to be shared. To ship
logs to a central location consider using remote syslog. Omnibus GitLab packages
provide configuration for [UDP log shipping](https://docs.gitlab.com/omnibus/settings/logs.html#udp-log-shipping-gitlab-enterprise-edition-only).
2021-06-08 01:23:25 +05:30
Having multiple NFS mounts requires you to manually make sure the data directories
2020-10-24 23:57:45 +05:30
are empty before attempting a restore. Read more about the
[restore prerequisites](../raketasks/backup_restore.md).
2021-01-03 14:25:43 +05:30
## Testing NFS
2023-01-13 00:05:48 +05:30
When you've set up the NFS server and client, you can verify NFS is configured correctly
2021-01-03 14:25:43 +05:30
by testing the following commands:
```shell
sudo mkdir /gitlab-nfs/test-dir
sudo chown git /gitlab-nfs/test-dir
sudo chgrp root /gitlab-nfs/test-dir
2021-01-29 00:20:46 +05:30
sudo chmod 0700 /gitlab-nfs/test-dir
sudo chgrp gitlab-www /gitlab-nfs/test-dir
sudo chmod 0751 /gitlab-nfs/test-dir
sudo chgrp git /gitlab-nfs/test-dir
sudo chmod 2770 /gitlab-nfs/test-dir
2021-01-03 14:25:43 +05:30
sudo chmod 2755 /gitlab-nfs/test-dir
sudo -u git mkdir /gitlab-nfs/test-dir/test2
sudo -u git chmod 2755 /gitlab-nfs/test-dir/test2
sudo ls -lah /gitlab-nfs/test-dir/test2
sudo -u git rm -r /gitlab-nfs/test-dir
```
Any `Operation not permitted` errors means you should investigate your NFS server export options.
2020-10-24 23:57:45 +05:30
## NFS in a Firewalled Environment
2022-07-16 23:28:13 +05:30
If the traffic between your NFS server and NFS clients is subject to port filtering
2021-06-08 01:23:25 +05:30
by a firewall, then you need to reconfigure that firewall to allow NFS communication.
2020-10-24 23:57:45 +05:30
2021-11-18 22:05:49 +05:30
[This guide from The Linux Documentation Project (TDLP)](https://tldp.org/HOWTO/NFS-HOWTO/security.html#FIREWALLS)
2020-10-24 23:57:45 +05:30
covers the basics of using NFS in a firewalled environment. Additionally, we encourage you to
search for and review the specific documentation for your operating system or distribution and your firewall software.
Example for Ubuntu:
Check that NFS traffic from the client is allowed by the firewall on the host by running
the command: `sudo ufw status`. If it's being blocked, then you can allow traffic from a specific
client with the command below.
```shell
sudo ufw allow from <client_ip_address> to any port nfs
```
## Known issues
2021-01-03 14:25:43 +05:30
### Upgrade to Gitaly Cluster or disable caching if experiencing data loss
2021-02-22 17:27:13 +05:30
WARNING:
2023-03-04 22:38:38 +05:30
Engineering support for NFS for Git repositories
[is unavailable](../update/removals.md#nfs-as-git-repository-storage-is-no-longer-supported).
2021-01-03 14:25:43 +05:30
Customers and users have reported data loss on high-traffic repositories when using NFS for Git repositories.
2021-04-29 21:17:54 +05:30
For example, we have seen:
- [Inconsistent updates after a push](https://gitlab.com/gitlab-org/gitaly/-/issues/2589).
- `git ls-remote` [returning the wrong (or no branches)](https://gitlab.com/gitlab-org/gitaly/-/issues/3083)
causing Jenkins to intermittently re-run all pipelines for a repository.
The problem may be partially mitigated by adjusting caching using the following NFS client mount options:
2021-01-03 14:25:43 +05:30
| Setting | Description |
| ------- | ----------- |
| `lookupcache=positive` | Tells the NFS client to honor `positive` cache results but invalidates any `negative` cache results. Negative cache results cause problems with Git. Specifically, a `git push` can fail to register uniformly across all NFS clients. The negative cache causes the clients to 'remember' that the files did not exist previously.
| `actimeo=0` | Sets the time to zero that the NFS client caches files and directories before requesting fresh information from a server. |
| `noac` | Tells the NFS client not to cache file attributes and forces application writes to become synchronous so that local changes to a file become visible on the server immediately. |
2021-02-22 17:27:13 +05:30
WARNING:
2021-01-03 14:25:43 +05:30
The `actimeo=0` and `noac` options both result in a significant reduction in performance, possibly leading to timeouts.
You may be able to avoid timeouts and data loss using `actimeo=0` and `lookupcache=positive` _without_ `noac`, however
2021-06-08 01:23:25 +05:30
we expect the performance reduction is still significant. Upgrade to
2021-01-03 14:25:43 +05:30
[Gitaly Cluster](gitaly/praefect.md) as soon as possible.
2021-04-17 20:07:23 +05:30
### Avoid using cloud-based file systems
2020-10-24 23:57:45 +05:30
2021-04-17 20:07:23 +05:30
GitLab strongly recommends against using cloud-based file systems such as:
2020-10-24 23:57:45 +05:30
2021-04-17 20:07:23 +05:30
- AWS Elastic File System (EFS).
- Google Cloud Filestore.
- Azure Files.
2020-10-24 23:57:45 +05:30
2021-04-17 20:07:23 +05:30
Our support team cannot assist with performance issues related to cloud-based file system access.
Customers and users have reported that these file systems don't perform well for
the file system access GitLab requires. Workloads where many small files are written in
a serialized manner, like `git`, are not well suited to cloud-based file systems.
If you do choose to use these, avoid storing GitLab log files (for example, those in `/var/log/gitlab`)
2022-01-26 12:08:38 +05:30
there because this also affects performance. We recommend that the log files be
2020-10-24 23:57:45 +05:30
stored on a local volume.
2021-04-17 20:07:23 +05:30
For more details on the experience of using a cloud-based file systems with GitLab,
see this [Commit Brooklyn 2019 video](https://youtu.be/K6OS8WodRBQ?t=313).
2020-10-24 23:57:45 +05:30
### Avoid using CephFS and GlusterFS
GitLab strongly recommends against using CephFS and GlusterFS.
2022-01-26 12:08:38 +05:30
These distributed file systems are not well-suited for the GitLab input/output access patterns because Git uses many small files and access times and file locking times to propagate makes Git activity very slow.
2020-10-24 23:57:45 +05:30
### Avoid using PostgreSQL with NFS
GitLab strongly recommends against running your PostgreSQL database
2022-01-26 12:08:38 +05:30
across NFS. The GitLab support team is not able to assist on performance issues related to
2020-10-24 23:57:45 +05:30
this configuration.
Additionally, this configuration is specifically warned against in the
[PostgreSQL Documentation](https://www.postgresql.org/docs/current/creating-cluster.html#CREATING-CLUSTER-NFS):
>PostgreSQL does nothing special for NFS file systems, meaning it assumes NFS behaves exactly like
>locally-connected drives. If the client or server NFS implementation does not provide standard file
>system semantics, this can cause reliability problems. Specifically, delayed (asynchronous) writes
>to the NFS server can cause data corruption problems.
2021-01-29 00:20:46 +05:30
For supported database architecture, see our documentation about
[configuring a database for replication and failover](postgresql/replication_and_failover.md).
2020-10-24 23:57:45 +05:30
2021-02-22 17:27:13 +05:30
## Troubleshooting
2020-10-24 23:57:45 +05:30
2021-02-22 17:27:13 +05:30
### Finding the requests that are being made to NFS
2020-10-24 23:57:45 +05:30
2021-02-22 17:27:13 +05:30
In case of NFS-related problems, it can be helpful to trace
2021-03-11 19:13:27 +05:30
the file system requests that are being made by using `perf`:
2021-02-22 17:27:13 +05:30
```shell
2021-09-04 01:27:46 +05:30
sudo perf trace -e 'nfs4:*' -p $(pgrep -fd ',' puma)
2021-02-22 17:27:13 +05:30
```
On Ubuntu 16.04, use:
```shell
2021-09-04 01:27:46 +05:30
sudo perf trace --no-syscalls --event 'nfs4:*' -p $(pgrep -fd ',' puma)
2021-02-22 17:27:13 +05:30
```
2022-07-16 23:28:13 +05:30
### Warnings `garbage found: .../repositories/@hashed/...git/objects/pack/.nfs...` in Gitaly logs
2022-08-27 11:52:29 +05:30
If you find any warnings like `garbage found: .../repositories/@hashed/...git/objects/pack/.nfs...` in [Gitaly logs](logs/index.md#gitaly-logs),
2022-07-16 23:28:13 +05:30
this problem occurs if `lookupcache=positive` is not set, which we recommend as an
[NFS mount option](#mount-options).
See [Gitaly issue #3175](https://gitlab.com/gitlab-org/gitaly/-/issues/3175) for more details.