Upgrade GitLab by using the GitLab package (FREE SELF)
You can upgrade GitLab to a new version by using the GitLab package.
Prerequisites
- Decide when to upgrade by viewing the supported upgrade paths. You can't directly skip major versions (for example, go from 10.3 to 12.7 in one step).
- If you are upgrading from a non-package installation to a GitLab package installation, see Upgrading from a non-package installation to a GitLab package installation.
- Ensure that any background migrations are fully completed. Upgrading before background migrations have finished can lead to data corruption. We recommend performing upgrades between major and minor releases no more than once per week, to allow time for background migrations to finish.
- Gitaly servers must be upgraded to the newer version prior to upgrading the application server. This prevents the gRPC client on the application server from sending RPCs that the old Gitaly version does not support.
Downtime
- For single node installations, GitLab is not available to users while an
upgrade is in progress. The user's web browser shows a
Deploy in progress
message or a502
error. - For multi-node installations, see how to perform zero downtime upgrades.
- Upgrades to multi-node installations can also be performed with downtime.
Version-specific changes
Upgrading versions might need some manual intervention. For more information, check the version your are upgrading to:
Earlier GitLab versions
For version-specific information for earlier GitLab versions, see the documentation archives. The versions of the documentation in the archives contain version-specific information for even earlier versions of GitLab.
For example, the documentation for GitLab 15.11 contains information on versions back to GitLab 11.
Back up before upgrading
The GitLab database is backed up before installing a newer GitLab version. You
may skip this automatic database backup by creating an empty file
at /etc/gitlab/skip-auto-backup
:
sudo touch /etc/gitlab/skip-auto-backup
Nevertheless, it is highly recommended to maintain a full up-to-date backup on your own.
Upgrade using the official repositories
All GitLab packages are posted to the GitLab package server. Five repositories are maintained:
-
gitlab/gitlab-ee
: The full GitLab package that contains all the Community Edition features plus the Enterprise Edition ones. -
gitlab/gitlab-ce
: A stripped down package that contains only the Community Edition features. -
gitlab/unstable
: Release candidates and other unstable versions. -
gitlab/nightly-builds
: Nightly builds. -
gitlab/raspberry-pi2
: Official Community Edition releases built for Raspberry Pi packages.
If you have installed GitLab Community Edition or Enterprise Edition, then the official GitLab repository should have already been set up for you.
Upgrade to the latest version using the official repositories
If you upgrade GitLab regularly, for example once a month, you can upgrade to the latest version by using your package manager.
To upgrade to the latest GitLab version:
# Ubuntu/Debian
sudo apt update && sudo apt install gitlab-ee
# RHEL/CentOS 6 and 7
sudo yum install gitlab-ee
# RHEL/CentOS 8
sudo dnf install gitlab-ee
# SUSE
sudo zypper install gitlab-ee
NOTE:
For the GitLab Community Edition, replace gitlab-ee
with
gitlab-ce
.
Upgrade to a specific version using the official repositories
Linux package managers default to installing the latest available version of a package for installation and upgrades. Upgrading directly to the latest major version can be problematic for older GitLab versions that require a multi-stage upgrade path. An upgrade path can span multiple versions, so you must specify the specific GitLab package with each upgrade.
To specify the intended GitLab version number in your package manager's install or upgrade command:
-
Identify the version number of the installed package:
# Ubuntu/Debian sudo apt-cache madison gitlab-ee # RHEL/CentOS 6 and 7 yum --showduplicates list gitlab-ee # RHEL/CentOS 8 dnf --showduplicates list gitlab-ee # SUSE zypper search -s gitlab-ee
-
Install the specific
gitlab-ee
package by using one of the following commands and replacing<version>
with the next supported version you would like to install (make sure to review the upgrade path to confirm the version you're installing is part of a supported path):# Ubuntu/Debian sudo apt install gitlab-ee=<version> # RHEL/CentOS 6 and 7 yum install gitlab-ee-<version> # RHEL/CentOS 8 dnf install gitlab-ee-<version> # SUSE zypper install gitlab-ee=<version>
NOTE:
For the GitLab Community Edition, replace gitlab-ee
with
gitlab-ce
.
Upgrade using a manually-downloaded package
NOTE: The package repository is recommended over a manual installation.
If for some reason you don't use the official repositories, you can download the package and install it manually. This method can be used to either install GitLab for the first time or upgrade it.
To download and install GitLab:
-
Visit the official repository of your package.
-
Filter the list by searching for the version you want to install (for example 14.1.8). Multiple packages may exist for a single version, one for each supported distribution and architecture. Next to the filename is a label indicating the distribution, as the filenames may be the same.
-
Find the package version you wish to install, and select the filename from the list.
-
In the upper-right corner, select Download.
-
After the package is downloaded, install it by using one of the following commands and replacing
<package_name>
with the package name you downloaded:# Debian/Ubuntu dpkg -i <package_name> # RHEL/CentOS 6 and 7 rpm -Uvh <package_name> # RHEL/CentOS 8 dnf install <package_name> # SUSE zypper install <package_name>
NOTE:
For the GitLab Community Edition, replace gitlab-ee
with
gitlab-ce
.
Upgrade the product documentation
This is an optional step. If you installed the product documentation, see how to upgrade to a later version.
Troubleshooting
Get the status of a GitLab installation
sudo gitlab-ctl status
sudo gitlab-rake gitlab:check SANITIZE=true
- Information on using
gitlab-ctl
to perform maintenance tasks. - Information on using
gitlab-rake
to check the configuration.
RPM 'package is already installed' error
If you are using RPM and you are upgrading from GitLab Community Edition to GitLab Enterprise Edition you may get an error like this:
package gitlab-7.5.2_omnibus.5.2.1.ci-1.el7.x86_64 (which is newer than gitlab-7.5.2_ee.omnibus.5.2.1.ci-1.el7.x86_64) is already installed
You can override this version check with the --oldpackage
option:
sudo rpm -Uvh --oldpackage gitlab-7.5.2_ee.omnibus.5.2.1.ci-1.el7.x86_64.rpm
Package obsoleted by installed package
CE and EE packages are marked as obsoleting and replacing each other so that both aren't installed and running at the same time.
If you are using local RPM files to switch from CE to EE or vice versa, use rpm
for installing the package rather than yum
. If you try to use yum, then you may get an error like this:
Cannot install package gitlab-ee-11.8.3-ee.0.el6.x86_64. It is obsoleted by installed package gitlab-ce-11.8.3-ce.0.el6.x86_64
To avoid this issue, either:
- Use the same instructions provided in the Upgrade using a manually-downloaded package section.
- Temporarily disable this checking in yum by adding
--setopt=obsoletes=0
to the options given to the command.
500 error when accessing Project > Settings > Repository
This error occurs when GitLab is converted from CE > EE > CE, and then back to EE. When viewing a project's repository settings, you can view this error in the logs:
Processing by Projects::Settings::RepositoryController#show as HTML
Parameters: {"namespace_id"=>"<namespace_id>", "project_id"=>"<project_id>"}
Completed 500 Internal Server Error in 62ms (ActiveRecord: 4.7ms | Elasticsearch: 0.0ms | Allocations: 14583)
NoMethodError (undefined method `commit_message_negative_regex' for #<PushRule:0x00007fbddf4229b8>
Did you mean? commit_message_regex_change):
This error is caused by an EE feature being added to a CE instance on the initial move to EE.
After the instance is moved back to CE and then is upgraded to EE again, the
push_rules
table already exists in the database. Therefore, a migration is
unable to add the commit_message_regex_change
column.
This results in the backport migration of EE tables not working correctly. The backport migration assumes that certain tables in the database do not exist when running CE.
To fix this issue:
-
Start a database console:
In GitLab 14.2 and later:
sudo gitlab-rails dbconsole --database main
In GitLab 14.1 and earlier:
sudo gitlab-rails dbconsole
-
Manually add the missing
commit_message_negative_regex
column:ALTER TABLE push_rules ADD COLUMN commit_message_negative_regex VARCHAR; # Exit psql \q
-
Restart GitLab:
sudo gitlab-ctl restart
Failed to connect to the internal GitLab API
on a separate GitLab Pages server
Error See GitLab Pages administration troubleshooting.
An error occurred during the signature verification
when running apt-get update
Error To update the GPG key of the GitLab packages server run:
curl --silent "https://packages.gitlab.com/gpg.key" | apt-key add -
apt-get update
Mixlib::ShellOut::CommandTimeout: rails_migration[gitlab-rails] [..] Command timed out after 3600s
If database schema and data changes (database migrations) must take more than one hour to run,
upgrades fail with a timed out
error:
FATAL: Mixlib::ShellOut::CommandTimeout: rails_migration[gitlab-rails] (gitlab::database_migrations line 51)
had an error: Mixlib::ShellOut::CommandTimeout: bash[migrate gitlab-rails database]
(/opt/gitlab/embedded/cookbooks/cache/cookbooks/gitlab/resources/rails_migration.rb line 16)
had an error: Mixlib::ShellOut::CommandTimeout: Command timed out after 3600s:
To fix this error:
-
Run the remaining database migrations:
sudo gitlab-rake db:migrate
This command may take a very long time to complete. Use
screen
or some other mechanism to ensure the program is not interrupted if your SSH session drops. -
Complete the upgrade:
sudo gitlab-ctl reconfigure
-
Hot reload
puma
andsidekiq
services:sudo gitlab-ctl hup puma sudo gitlab-ctl restart sidekiq
Missing asset files
Following an upgrade, GitLab might not be correctly serving up assets such as images, JavaScript, and style sheets. It might be generating 500 errors, or the web UI may be failing to render properly.
In a scaled out GitLab environment, if one web server behind the load balancer is demonstrating this issue, the problem occurs intermittently.
The Rake task to recompile the
assets doesn't apply to an Omnibus installation which serves
pre-compiled assets from /opt/gitlab/embedded/service/gitlab-rails/public/assets
.
Potential causes and fixes:
- Ensure no old processes are running.
- Remove duplicate sprockets files
- The installation is incomplete
- NGINX Gzip support is disabled
Old processes
The most likely cause is that an old Puma process is running, instructing clients to request asset files from a previous release of GitLab. As the files no longer exist, HTTP 404 errors are returned.
A reboot is the best way to ensure these old Puma processes are no longer running.
Alternatively:
-
Stop Puma:
gitlab-ctl stop puma
-
Check for any remaining Puma processes, and kill them:
ps -ef | egrep 'puma[: ]' kill <processid>
-
Verify with
ps
that the Puma processes have stopped running. -
Start Puma
gitlab-ctl start puma
Duplicate sprockets files
The compiled asset files have unique file names in each release. The sprockets files provide a mapping from the filenames in the application code to the unique filenames.
/opt/gitlab/embedded/service/gitlab-rails/public/assets/.sprockets-manifest*.json
Make sure there's only one sprockets file. Rails uses the first one.
A check for duplicate sprockets files runs during Omnibus GitLab upgrades:
GitLab discovered stale file(s) from the previous install that need to be cleaned up.
The following files need to be removed:
/opt/gitlab/embedded/service/gitlab-rails/public/assets/.sprockets-manifest-e16fdb7dd73cfdd64ed9c2cc0e35718a.json
Options for resolving this include:
-
If you have the output from the package upgrade, remove the specified files. Then restart Puma:
gitlab-ctl restart puma
-
If you don't have the message, perform a reinstall (see incomplete installation below for more details) to generate it again.
-
Remove all the sprockets files and then follow the instructions for an incomplete installation.
Incomplete installation
An incomplete installation could be the cause of this issue.
Verify the package to determine if this is the problem:
-
For Debian distributions:
apt-get install debsums debsums -c gitlab-ee
-
For Red Hat/SUSE (RPM) distributions:
rpm -V gitlab-ee
To reinstall the package to fix an incomplete installation:
-
Check the installed version
-
For Debian distributions:
apt --installed list gitlab-ee
-
For Red Hat/SUSE (RPM) distributions:
rpm -qa gitlab-ee
-
-
Reinstall the package, specifying the installed version. For example 14.4.0 Enterprise Edition:
-
For Debian distributions:
apt-get install --reinstall gitlab-ee=14.4.0-ee.0
-
For Red Hat/SUSE (RPM) distributions:
yum reinstall gitlab-ee-14.4.0
-
NGINX Gzip support
Check whether nginx['gzip_enabled']
has been disabled:
grep gzip /etc/gitlab/gitlab.rb
This might prevent some assets from being served. Read more in one of the related issues.