If replicating over SSH, ensure the host key of the remote system(s) is already in the Gerrit user's ~/.ssh/known_hosts
file. The easiest way to add the host key is to connect once by hand with the command line:
sudo su -c 'ssh mirror1.us.some.org echo' gerrit2
[remote "host-two"] url = gerrit2@host-two.example.com:/some/path/${name}.git apiUrl = http://host-two.example.com:8080 fetch = +refs/*:refs/* [remote "pubmirror"] url = mirror1.us.some.org:/pub/git/${name}.git url = mirror2.us.some.org:/pub/git/${name}.git url = mirror3.us.some.org:/pub/git/${name}.git fetch = +refs/heads/*:refs/heads/* fetch = +refs/tags/*:refs/tags/* threads = 3 authGroup = Public Mirror Group authGroup = Second Public Mirror Group
And make sure that instanceId is setup in $site_path/etc/gerrit.config
:
[gerrit] instanceId = host-one
Then reload the replication plugin to pick up the new configuration:
ssh -p 29418 localhost gerrit plugin reload @PLUGIN@
To manually trigger replication at runtime, see SSH command start.
gerrit.config
cache.@PLUGIN@-apply_objects.maxAge : Maximum age to keep history of the latest successful apply-object refs. Values should use common unit suffixes to express their setting:
s, sec, second, seconds m, min, minute, minutes h, hr, hour, hours d, day, days w, week, weeks (1 week is treated as 7 days) mon, month, months (1 month is treated as 30 days) y, year, years (1 year is treated as 365 days) If a unit suffix is not specified, seconds is assumed. If 0 is supplied, the maximum age is infinite and items are never purged except when the cache is full. Default is 60s.
cache.@PLUGIN@-apply_objects.memoryLimit : The maximum number of apply-object refs retained in memory.
Default is 1024.
cache.projects.refreshAfterWrite : The Gerrit configuration of the projects
cache, as documented
Needs to be set to a relatively low value (e.g. 1 min) for allowing the project settings cache to be kept relatively up-to-date (e.g. within 1 mins) with the incoming replication tasks updating it.
cache.project_list.refreshAfterWrite : The Gerrit configuration of the project_list
cache, as documented
Needs to be set to a relatively low value (e.g. 5 min) for allowing the creation, removal and hiding of projects performed by incoming replication tasks to be reflected relavitely soon (e.g. within 5 mins) in the list of projects.
@PLUGIN@.config
The optional file $site_path/etc/replication.config
is a Git-style config file that controls the replication settings for the replication plugin.
The file is composed of one or more remote
sections, each remote section provides common configuration settings for one or more source URLs.
Each remote section uses its own thread pool. If fetching from multiple remotes, over differing types of network connections (e.g. LAN and also public Internet), its a good idea to put them into different remote sections, so that replication to the slower connection does not starve out the faster local one. The example file above does this.
In the keys below, the NAME
portion is unused by this plugin, but must be unique to distinguish the different sections if more than one remote section appears in the file.
gerrit.replicateOnStartup : If true, replicates from all remotes on startup to ensure they are in-sync with this server. By default, false.
gerrit.autoReload : If true, automatically reloads replication sources and settings after replication.config
file is updated, without the need to restart the replication plugin. When the reload takes place, pending replication events based on old settings are discarded. By default, false.
gerrit.defaultForceUpdate : If true, the default fetch refspec will be set to use forced update to the local repository when no refspec is given. By default, false.
replication.lockErrorMaxRetries : Number of times to retry a replication operation if a lock error is detected.
If two or more replication operations (to the same GIT and Ref) are scheduled at approximately the same time (and end up on different replication threads), there is a large probability that the last fetch to complete will fail with a remote "failure to lock" error. This option allows Gerrit to retry the replication fetch when the "failure to lock" error is detected. A good value would be 3 retries or less, depending on how often you see lockError collisions in your server logs. A too highly set value risks keeping around the replication operations in the queue for a long time, and the number of items in the queue will increase with time. Normally Gerrit will succeed with the replication during its first retry, but in certain edge cases (e.g. a mirror introduces a ref namespace with the same name as a branch on the master) the retry will never succeed. The issue can also be mitigated somewhat by increasing the replicationDelay. Default: 0 (disabled, i.e. never retry)
replication.maxRetries : Maximum number of times to retry a fetch operation that previously failed.
When a fetch operation reaches its maximum number of retries, the replication event is discarded from the queue. Can be overridden at remote-level by setting replicationMaxRetries. By default, fetches are retried indefinitely. Note that only transient errors will be retried, whilst persistent errors will cause a terminal failure, and the fetch will not be scheduled again. This is only supported for JGit, not cGit. Currently, only the following failures are considered permanent: - UnknownHostKey: thrown by Jsch when establishing an SSH connection for an unknown host. - cGit / JGit transport exception when the remote ref does not exist. The assumption here is that the remote ref does not exist so it is not worth retrying. If the exception arisen as a consequence of some ACLs (mis)configuration, then after fixing the ACLs, an explicit replication must be manually triggered.
replication.instanceLabel : Remote configuration name of the current server. This label is passed as a part of the payload to notify other servers to fetch specified ref-update object id from the url provided in the remote configuration section which name is equal to instanceLabel.
Deprecated: This property is kept for backward compatibility and will be removed in the future release. Use [gerrit.instanceId](https://gerrit-review.googlesource.com/Documentation/config-gerrit.html#gerrit.instanceId) instead.
replication.consumeStreamEvents : Use stream events to trigger pull-replication actions alongside the REST approach. This mechanism is useful together with event-broker and multi-site to provide backfill mechanism when a node has to catch up with the events after being unreachable.
When `consumeStreamEvents` is enabled gerrit.instanceId instead of [replication.instanceLabel](https://gerrit.googlesource.com/plugins/pull-replication/+/refs/heads/stable-3.4/src/main/resources/Documentation/config.md#replication.instanceLabel) must be used. Using pull-replication standalone with a broker is also possible. Check the replication.eventBrokerTopic parameter. Default: false
replication.eventBrokerTopic : Topic to consumer stream events from.
Pull-replication can receive stream events and use it as notification mechanism as alternative to REST API notifications. It can work in standalone, not necessarily with the multi-site plugin (check replication.consumerStreamEvents if you need to use it with multi-site). This parameter is used to define the topic to consume stream events messages from when using the pull-replication in standalone with a broker. When `eventBrokerTopic` is enabled gerrit.instanceId instead of [replication.instanceLabel](https://gerrit.googlesource.com/plugins/pull-replication/+/refs/heads/stable-3.4/src/main/resources/Documentation/config.md#replication.instanceLabel) must be used. Bear in mind that if consumeStreamEvents is set to true this parameter will be ignored. Default: unset
replication.eventBrokerGroupId : Optional group id for consumers. Used for keeping the offset of the the last acked message on the topic with a specific group id; supported only if the event-broker plugin supports the ExtendedBrokerApi interface. When not set, the messages received are acked using a global group id shared by all subscribers in Gerrit.
Default: unset
replication.maxConnectionsPerRoute : Maximum number of HTTP connections per one HTTP route.
Default: 100
replication.maxConnections : Total number of HTTP connections pool.
Default: 2 * replication.maxConnectionsPerRoute
replication.useCGitClient : By default Gerrit uses JGit library to execute all git protocol command. By setting this property to true all git fetch operation are going to be executed using CGit client instead of JGit.
Default: false
replication.refsBatchSize : Number of refs that are fetched in a single fetch call. If number of refs to fetch is greater then this param, refs are going to be split into a separate fetch operations.
Value must be greater than zero. Default: 50
replication.excludeRefs : Specify which refs should be excluded from git fetch calls. It can be provided more than once, and supports three formats: regular expressions, wildcard matching, and single ref matching. All three formats match are case-sensitive.
Values starting with a caret `^` are treated as regular expressions. For the regular expressions details please follow official [java documentation](https://docs.oracle.com/javase/tutorial/essential/regex/). Please note that regular expressions could also be used with inverse match. Values that are not regular expressions and end in `*` are treated as wildcard matches. Wildcards match refs whose name agrees from the beginning until the trailing `*`. So `foo/b*` would match the refs `foo/b`, `foo/bar`, and `foo/baz`, but neither `foobar`, nor `bar/foo/baz`. Values that are neither regular expressions nor wildcards are treated as single ref matches. So `foo/bar` matches only the ref `foo/bar`, but no other refs. By default, the following refs are excluded from the git fetch calls: - refs/cache-automerge/* - refs/starred-changes/* Note that if you are using @PLUGIN@ together with multi-site, you should explicitly exclude `refs/multi-site/version` from being replicated.
replication.syncRefs : Specify for which refs git fetch calls should be executed synchronously. It can be provided more than once, and supports three formats: regular expressions, wildcard matching, and single ref matching. All three formats match are case-sensitive.
Values starting with a caret `^` are treated as regular expressions. For the regular expressions details please follow official [java documentation](https://docs.oracle.com/javase/tutorial/essential/regex/). Please note that regular expressions could also be used with inverse match. Values that are not regular expressions and end in `*` are treated as wildcard matches. Wildcards match refs whose name agrees from the beginning until the trailing `*`. So `foo/b*` would match the refs `foo/b`, `foo/bar`, and `foo/baz`, but neither `foobar`, nor `bar/foo/baz`. Values that are neither regular expressions nor wildcards are treated as single ref matches. So `foo/bar` matches only the ref `foo/bar`, but no other refs. By default, set to '*' (all refs are replicated synchronously).
replication.fallbackToApplyObjectsRefs : Specify for which refs will fallback to apply-objects REST-API that is capable of applying an entire chain of commits on the ref chain so that the Git fetch can be avoided altogether. It can be provided more than once, and supports three formats: regular expressions, wildcard matching, and single ref matching. All three formats matches are case-sensitive.
Values starting with a caret `^` are treated as regular expressions. For the regular expressions details please follow official [java documentation](https://docs.oracle.com/javase/tutorial/essential/regex/). Please note that regular expressions could also be used with inverse match. Values that are not regular expressions and end in `*` are treated as wildcard matches. Wildcards match refs whose name agrees from the beginning until the trailing `*`. So `foo/b*` would match the refs `foo/b`, `foo/bar`, and `foo/baz`, but neither `foobar`, nor `bar/foo/baz`. Values that are neither regular expressions nor wildcards are treated as single ref matches. So `foo/bar` matches only the ref `foo/bar`, but no other refs. By default, set to 'refs/changes/**/meta'. Note that the refs/changes/**/meta always fallback to apply-objects REST-API
replication.maxApiPayloadSize : Maximum size in bytes of the ref to be sent as a REST Api call payload. For refs larger than threshold git fetch operation will be used.
Default: 10000
replication.shutDownDrainTimeout : Maximum duration to wait for pending, retrying and executing fetch tasks to complete after a request of plugin shutdown. Values should use common unit suffixes to express their setting:
s, sec, second, seconds m, min, minute, minutes h, hr, hour, hours d, day, days w, week, weeks (1 week is treated as 7 days) mon, month, months (1 month is treated as 30 days) y, year, years (1 year is treated as 365 days) If a unit suffix is not specified, seconds is assumed. If 0 is supplied, the maximum age is infinite and items are never purged except when the cache is full. Default: 5 minutes
remote.NAME.url : Address of the remote server to fetch from. Single URL can be specified within a single remote block. A remote node can request a fetch from a source. However, it will need to be sure that the source corresponds to only him configured as remote. Fetching from multiple sources at the same time, whilst it was useful in a push scenario, doesn‘t make so much sense in a pull-replication flow. Within each URL value the magic placeholder ${name}
is replaced with the Gerrit project name. This is a Gerrit specific extension to the otherwise standard Git URL syntax and it must be included in each URL so that Gerrit can figure out where each project needs to be replicated. ${name}
may only be omitted if the remote refers to a single repository (i.e.: Exactly one remote.NAME.projects and that name’s value is a single project match.).
See [git fetch][1] for details on Git URL syntax.
remote.NAME.apiUrl : Address of the rest api endpoint of the remote server to fetch from. Multiple URLs may be specified within a single remote block, listing different destinations which share the same settings. Gerrit calls all URLs in sequence.
remote.NAME.connectionTimeout : Defines the socket timeout ({@code SO_TIMEOUT}) in milliseconds, which is the timeout for waiting for data or, put differently, a maximum period inactivity between two consecutive data packets.
Default: 5000
remote.NAME.idleTimeout : Defines period of inactivity in milliseconds after which persistent connections must be re-validated prior to being leased to the consumer. Non-positive value disables connection validation. This check helps detect connections that have become stale (half-closed) while kept inactive in the pool.
Default: 10000
remote.NAME.uploadpack : Path of the git-upload-pack
executable on the remote system, if using the SSH transport.
Defaults to `git-upload-pack`.
remote.NAME.fetch : Standard Git refspec denoting what should be replicated. Setting this to +refs/heads/*:refs/heads/*
would mirror only the active branches, but not the change refs under refs/changes/
, or the tags under refs/tags/
.
Multiple fetch keys can be supplied, to specify multiple patterns to match against. In the [example above][2], remote "pubmirror" uses two fetch keys to match both `refs/heads/*` and `refs/tags/*`, but excludes all others, including `refs/changes/*`. Note that the `refs/meta/config` branch is only replicated when `replicatePermissions` is true, even if the push refspec is 'all refs'.
NOTE: NOT having a
fetch
configuration for a remote, causes the remote to be ignored.
remote.NAME.timeout : Number of seconds to wait for a network read or write to complete before giving up and declaring the remote side is not responding. If 0, there is no timeout, and the push client waits indefinitely.
A timeout should be large enough to mostly transfer the objects from the other side. 1 second may be too small for larger projects, especially over a WAN link, while 10-30 seconds is a much more reasonable timeout value. Defaults to 0 seconds, wait indefinitely.
remote.NAME.replicationDelay : Time to wait before scheduling an asynchronous remote fetch operation. Setting the delay to 0 effectively disables the delay, causing the fetch to start as soon as possible.
This is a Gerrit specific extension to the Git remote block. By default for asynchronous fetch, 4 seconds. For a synchronous fetch replicationDelay is zero.
remote.NAME.rescheduleDelay : Delay when rescheduling a fetch operation due to an in-flight fetch running for the same project.
Cannot be set to a value lower than 3 seconds to avoid a tight loop of schedule/run which could cause 1K+ retries per second. A configured value lower than 3 seconds will be rounded to 3 seconds. By default, 3 seconds.
remote.NAME.replicationRetry : Time to wait before scheduling a remote fetch operation previously failed due to a remote server error.
If a remote fetch operation fails because a remote server was offline, all fetch operations from the same source URL are blocked, and the remote fetch is continuously retried unless the replicationMaxRetries value is set. This is a Gerrit specific extension to the Git remote block. By default, 1 minute.
remote.NAME.replicationMaxRetries : Maximum number of times to retry a fetch operation that previously failed.
When a fetch operation reaches its maximum number of retries the replication event is discarded from the queue. This is a Gerrit specific extension to the Git remote block. By default, use replication.maxRetries. Note that not all fetch failures are retriable. Please refer to `replication.maxRetries` for more information on this.
remote.NAME.threads : Number of worker threads to dedicate to fetching to the repositories described by this remote. Each thread can fetch one project at a time, from one source URL. Scheduling within the thread pool is done on a per-project basis. If a remote block describes 4 URLs, allocating 4 threads in the pool will permit some level of parallel fetching.
By default, 1 thread.
remote.NAME.createMissingRepositories : Replicate newly created repositories.
By default, true.
remote.NAME.replicateProjectDeletions : If true, project deletions will also be replicated to the remote site.
By default, false, do *not* replicate project deletions.
remote.NAME.replicateHiddenProjects : If true, hidden projects will also be replicated to the remote site.
By default, false, do *not* replicate hidden projects.
remote.NAME.mirror : If true, replication will remove local branches and tags that are absent remotely or invisible to the replication (for example read access denied via authGroup
option).
By default, false, do not remove remote branches or tags.
remote.NAME.authGroup : Specifies the name of a group that the remote should use to access the repositories. Multiple authGroups may be specified within a single remote block to signify a wider access right. In the project administration web interface the read access can be specified for this group to control if a project should be replicated or not to the remote.
By default, replicates without group control, i.e. replicates everything from all remotes.
remote.NAME.remoteNameStyle : Provides possibilities to influence the name of the source repository, e.g. by replacing slashes in the ${name}
placeholder.
Github and Gitorious do not permit slashes "/" in repository names and will change them to dashes "-" at repository creation time. If this setting is set to "dash", slashes will be replaced with dashes in the remote repository name. If set to "underscore", slashes will be replaced with underscores in the repository name. Option "basenameOnly" makes `${name}` to be only the basename (the part after the last slash) of the repository path on the Gerrit server, e.g. `${name}` of `foo/bar/my-repo.git` would be `my-repo`. By default, "slash", i.e. remote names will contain slashes as they do in Gerrit.
remote.NAME.projects : Specifies which repositories should be replicated from the remote. It can be provided more than once, and supports three formats: regular expressions, wildcard matching, and single project matching. All three formats match case-sensitive.
Values starting with a caret `^` are treated as regular expressions. `^foo/(bar|baz)` would match the projects `foo/bar`, and `foo/baz`. Regular expressions have to fully match the project name. So the above example would not match `foo/bar2`, while `^foo/(bar|baz).*` would. Projects may be excluded from replication by using a regular expression with inverse match. `^(?:(?!PATTERN).)*$` will exclude any project that matches. Values that are not regular expressions and end in `*` are treated as wildcard matches. Wildcards match projects whose name agrees from the beginning until the trailing `*`. So `foo/b*` would match the projects `foo/b`, `foo/bar`, and `foo/baz`, but neither `foobar`, nor `bar/foo/baz`. Values that are neither regular expressions nor wildcards are treated as single project matches. So `foo/bar` matches only the project `foo/bar`, but no other project. By default, replicates without matching, i.e. replicates everything from all remotes.
remote.NAME.enableBatchedRefs : Keep the synchronous replication actions together as a single call with all the refs included in the batch ref update.
NOTE: if any ref from a single batch matches
replication.syncRefs
filter, all refs in that batch are going to be fetched synchronously as a single git fetch operation.
NOTE: Should ref deletions over apply-object/HTTP, they will be attempted over fetch/HTTP endpoint only when
enableBatchedRefs
is enabled.By default, true.
remote.NAME.fetchEvery : Fetch the ref-spec remote.NAME.fetch
from the remote repository defined at remote.NAME.url
for remote changes every n
seconds. By default set to 0s
which means that polling is disabled. Note that larger time units (m
, h
, etc...) can be used to specify it conveniently.
Setting `fetchEvery` to value greater than `0` means that It will, periodically, invoke git fetch to detect and retrieve new data from the remote.
NOTE: it is meant for the remote that doesn't offer events or webhooks that could be used otherwise for new data detection and cannot be configured together with
remote.NAME.apiUrl
; such configuration is considered invalid and prevents the plugin from starting
NOTE: Periodic fetches are scheduled using a dedicated (single threaded) pool, called
PeriodicallyFetchFromSources
. It is created only when there is at least one remote configured to fetch periodically.
NOTE: Scheduling is performed on plugin start therefore one needs to reload plugin when configuration gets changed.
replication
The optional directory $site_path/etc/replication
contains Git-style config files that controls the replication settings for the pull replication plugin. When present all remote
sections from replication.config
file are ignored.
Files are composed of one remote
section. Multiple remote
sections or any other section makes the file invalid and skipped by the pull replication plugin. File name defines remote section name. Each section provides common configuration settings for one or more destination URLs. For more details how to setup remote
sections please refer to the replication.config
section.
Static configuration in $site_path/etc/replication.config
:
[gerrit] autoReload = true replicateOnStartup = false [replication] lockErrorMaxRetries = 5 maxRetries = 5
Remote sections in $site_path/etc/replication
directory:
$site_path/etc/replication/host-two.config
[remote] url = gerrit2@host-two.example.com:/some/path/${name}.git apiUrl = http://host-two fetch = +refs/*:refs/*
$site_path/etc/replication/host-three.config
[remote] url = mirror1.host-three:/pub/git/${name}.git url = mirror2.host-three:/pub/git/${name}.git url = mirror3.host-three:/pub/git/${name}.git apiUrl = http://host-three fetch = +refs/heads/*:refs/heads/* fetch = +refs/tags/*:refs/tags/*
Pull replication plugin resolves config files to the following configuration:
[gerrit] autoReload = true replicateOnStartup = false [replication] lockErrorMaxRetries = 5 maxRetries = 5 [remote "host-two"] url = gerrit2@host-two.example.com:/some/path/${name}.git apiUrl = http://host-two fetch = +refs/*:refs/* [remote "host-three"] url = mirror1.host-three:/pub/git/${name}.git url = mirror2.host-three:/pub/git/${name}.git url = mirror3.host-three:/pub/git/${name}.git apiUrl = http://host-three fetch = +refs/heads/*:refs/heads/* fetch = +refs/tags/*:refs/tags/*
secure.config
The optional file $site_path/secure.config
is a Git-style config file that provides secure values that should not be world-readable, such as passwords. The HTTP authentication can be configured in 2 different flavours:
HTTP Bearer Token Authentication
auth.bearerToken : shared secret configured on all replication endpoints.
HTTP Basic Authentication
remote.NAME.username : Username to use for HTTP authentication on this remote, if not given in the URL.
remote.NAME.password : Password to use for HTTP authentication on this remote.
In both cases, the Global Capability Access Database
1 needs to be allowed in order to permit all All-Users
' refs to be replicated. When basic auth is used, the capability must be assigned to the remote.NAME.username
used in configuration, whilst for bearer token, it needs to be assigned to the Pull-replication Internal User
user.
1 https://gerrit-review.googlesource.com/Documentation/access-control.html#capability_accessDatabase
~/.ssh/config
If present, Gerrit reads and caches ~/.ssh/config
at startup, and supports most SSH configuration options. For example:
Host host-one.example.com IdentityFile ~/.ssh/id_hostone PreferredAuthentications publickey Host mirror*.us.some.org User mirror-updater IdentityFile ~/.ssh/id_pubmirror PreferredAuthentications publickey
Supported options:
SSH authentication must be by passwordless public key, as there is no facility to read passphrases on startup or passwords during the SSH connection setup, and SSH agents are not supported from Java.
Host keys for any destination SSH servers must appear in the user's ~/.ssh/known_hosts
file, and must be added in advance, before Gerrit starts. If a host key is not listed, Gerrit will be unable to connect to that destination, and replication to that URL will fail.