1--- 2stage: Enablement 3group: Geo 4info: To determine the technical writer assigned to the Stage/Group associated with this page, see https://about.gitlab.com/handbook/engineering/ux/technical-writing/#assignments 5type: howto 6--- 7 8# Disaster Recovery (Geo) **(PREMIUM SELF)** 9 10Geo replicates your database, your Git repositories, and few other assets, 11but there are some [limitations](../index.md#limitations). 12 13WARNING: 14Disaster recovery for multi-secondary configurations is in **Alpha**. 15For the latest updates, check the [Disaster Recovery epic for complete maturity](https://gitlab.com/groups/gitlab-org/-/epics/3574). 16Multi-secondary configurations require the complete re-synchronization and re-configuration of all non-promoted secondaries and 17causes downtime. 18 19## Promoting a **secondary** Geo site in single-secondary configurations 20 21We don't currently provide an automated way to promote a Geo replica and do a 22failover, but you can do it manually if you have `root` access to the machine. 23 24This process promotes a **secondary** Geo site to a **primary** site. To regain 25geographic redundancy as quickly as possible, you should add a new **secondary** site 26immediately after following these instructions. 27 28### Step 1. Allow replication to finish if possible 29 30If the **secondary** site is still replicating data from the **primary** site, follow 31[the planned failover docs](planned_failover.md) as closely as possible in 32order to avoid unnecessary data loss. 33 34### Step 2. Permanently disable the **primary** site 35 36WARNING: 37If the **primary** site goes offline, there may be data saved on the **primary** site 38that have not been replicated to the **secondary** site. This data should be treated 39as lost if you proceed. 40 41If an outage on the **primary** site happens, you should do everything possible to 42avoid a split-brain situation where writes can occur in two different GitLab 43instances, complicating recovery efforts. So to prepare for the failover, we 44must disable the **primary** site. 45 46- If you have SSH access: 47 48 1. SSH into the **primary** site to stop and disable GitLab: 49 50 ```shell 51 sudo gitlab-ctl stop 52 ``` 53 54 1. Prevent GitLab from starting up again if the server unexpectedly reboots: 55 56 ```shell 57 sudo systemctl disable gitlab-runsvdir 58 ``` 59 60- If you do not have SSH access to the **primary** site, take the machine offline and 61 prevent it from rebooting by any means at your disposal. 62 You might need to: 63 64 - Reconfigure the load balancers. 65 - Change DNS records (for example, point the primary DNS record to the 66 **secondary** site to stop usage of the **primary** site). 67 - Stop the virtual servers. 68 - Block traffic through a firewall. 69 - Revoke object storage permissions from the **primary** site. 70 - Physically disconnect a machine. 71 72 If you plan to [update the primary domain DNS record](#step-4-optional-updating-the-primary-domain-dns-record), 73 you may wish to lower the TTL now to speed up propagation. 74 75### Step 3. Promoting a **secondary** site 76 77WARNING: 78In GitLab 13.2 and 13.3, promoting a secondary site to a primary while the 79secondary is paused fails. Do not pause replication before promoting a 80secondary. If the secondary site is paused, be sure to resume before promoting. 81This issue has been fixed in GitLab 13.4 and later. 82 83Note the following when promoting a secondary: 84 85- If replication was paused on the secondary site (for example as a part of 86 upgrading, while you were running a version of GitLab earlier than 13.4), you 87 _must_ [enable the site by using the database](../replication/troubleshooting.md#message-activerecordrecordinvalid-validation-failed-enabled-geo-primary-node-cannot-be-disabled) 88 before proceeding. If the secondary site 89 [has been paused](../../geo/index.md#pausing-and-resuming-replication), the promotion 90 performs a point-in-time recovery to the last known state. 91 Data that was created on the primary while the secondary was paused is lost. 92- A new **secondary** should not be added at this time. If you want to add a new 93 **secondary**, do this after you have completed the entire process of promoting 94 the **secondary** to the **primary**. 95- If you encounter an `ActiveRecord::RecordInvalid: Validation failed: Name has already been taken` 96 error message during this process, for more information, see this 97 [troubleshooting advice](../replication/troubleshooting.md#fixing-errors-during-a-failover-or-when-promoting-a-secondary-to-a-primary-node). 98- If you run into errors when using `--force` or `--skip-preflight-checks` before 13.5 during this process, 99 for more information, see this 100 [troubleshooting advice](../replication/troubleshooting.md#errors-when-using---skip-preflight-checks-or---force). 101 102#### Promoting a **secondary** site running on a single node running GitLab 14.5 and later 103 1041. SSH in to your **secondary** node and execute: 105 106 - To promote the secondary node to primary: 107 108 ```shell 109 sudo gitlab-ctl geo promote 110 ``` 111 112 - To promote the secondary node to primary **without any further confirmation**: 113 114 ```shell 115 sudo gitlab-ctl geo promote --force 116 ``` 117 1181. Verify you can connect to the newly-promoted **primary** site using the URL used 119 previously for the **secondary** site. 1201. If successful, the **secondary** site is now promoted to the **primary** site. 121 122#### Promoting a **secondary** site running on a single node running GitLab 14.4 and earlier 123 124WARNING: 125The `gitlab-ctl promote-to-primary-node` and `gitlab-ctl promoted-db` commands are 126deprecated in GitLab 14.5 and later, and are scheduled to [be removed in GitLab 15.0](https://gitlab.com/gitlab-org/gitlab/-/issues/345207). 127Use `gitlab-ctl geo promote` instead. 128 1291. SSH in to your **secondary** node and login as root: 130 131 ```shell 132 sudo -i 133 ``` 134 1351. If you're using GitLab 13.5 and later, skip this step. If not, edit 136 `/etc/gitlab/gitlab.rb` and remove any of the following lines that 137 might be present: 138 139 ```ruby 140 geo_secondary_role['enable'] = true 141 roles ['geo_secondary_role'] 142 ``` 143 1441. Promote the **secondary** site to the **primary** site: 145 146 - To promote the secondary node to primary along with [preflight checks](planned_failover.md#preflight-checks): 147 148 ```shell 149 gitlab-ctl promote-to-primary-node 150 ``` 151 152 - If you have already run the preflight checks separately or don't want to run them, 153 you can skip them with: 154 155 ```shell 156 gitlab-ctl promote-to-primary-node --skip-preflight-checks 157 ``` 158 159 NOTE: 160 In GitLab 13.7 and earlier, if you have a data type with zero items to sync 161 and don't skip the preflight checks, promoting the secondary reports 162 `ERROR - Replication is not up-to-date` even if replication is actually 163 up-to-date. If replication and verification output 164 shows that it is complete, you can skip the preflight checks to make the 165 command complete promotion. This bug was fixed in GitLab 13.8 and later. 166 167 - To promote the secondary node to primary **without any further confirmation**, 168 even when preflight checks fail: 169 170 ```shell 171 gitlab-ctl promote-to-primary-node --force 172 ``` 173 1741. Verify you can connect to the newly-promoted **primary** site using the URL used 175 previously for the **secondary** site. 1761. If successful, the **secondary** site is now promoted to the **primary** site. 177 178#### Promoting a **secondary** site with multiple nodes running GitLab 14.5 and later 179 1801. SSH to every Sidekiq, PostgresSQL, and Gitaly node in the **secondary** site and run one of the following commands: 181 182 - To promote the secondary node to primary: 183 184 ```shell 185 sudo gitlab-ctl geo promote 186 ``` 187 188 - To promote the secondary node to primary **without any further confirmation**: 189 190 ```shell 191 sudo gitlab-ctl geo promote --force 192 ``` 193 1941. SSH into each Rails node on your **secondary** site and run one of the following commands: 195 196 - To promote the secondary node to primary: 197 198 ```shell 199 sudo gitlab-ctl geo promote 200 ``` 201 202 - To promote the secondary node to primary **without any further confirmation**: 203 204 ```shell 205 sudo gitlab-ctl geo promote --force 206 ``` 207 2081. Verify you can connect to the newly-promoted **primary** site using the URL used 209 previously for the **secondary** site. 2101. If successful, the **secondary** site is now promoted to the **primary** site. 211 212#### Promoting a **secondary** site with multiple nodes running GitLab 14.4 and earlier 213 214WARNING: 215The `gitlab-ctl promote-to-primary-node` and `gitlab-ctl promoted-db` commands are 216deprecated in GitLab 14.5 and later, and are scheduled to [be removed in GitLab 15.0](https://gitlab.com/gitlab-org/gitlab/-/issues/345207). 217Use `gitlab-ctl geo promote` instead. 218 219The `gitlab-ctl promote-to-primary-node` command cannot be used yet in 220conjunction with multiple nodes, as it can only perform changes on 221a **secondary** with only a single node. Instead, you must 222do this manually. 223 2241. SSH in to the database node in the **secondary** site and trigger PostgreSQL to 225 promote to read-write: 226 227 ```shell 228 sudo gitlab-ctl promote-db 229 ``` 230 231 In GitLab 12.8 and earlier, see [Message: `sudo: gitlab-pg-ctl: command not found`](../replication/troubleshooting.md#message-sudo-gitlab-pg-ctl-command-not-found). 232 2331. Edit `/etc/gitlab/gitlab.rb` on every node in the **secondary** site to 234 reflect its new status as **primary** by removing any of the following 235 lines that might be present: 236 237 ```ruby 238 geo_secondary_role['enable'] = true 239 roles ['geo_secondary_role'] 240 ``` 241 242 After making these changes, [reconfigure GitLab](../../restart_gitlab.md#omnibus-gitlab-reconfigure) 243 on each machine so the changes take effect. 244 2451. Promote the **secondary** to **primary**. SSH into a single application 246 server and execute: 247 248 ```shell 249 sudo gitlab-rake geo:set_secondary_as_primary 250 ``` 251 2521. Verify you can connect to the newly-promoted **primary** using the URL used 253 previously for the **secondary**. 2541. If successful, the **secondary** site is now promoted to the **primary** site. 255 256#### Promoting a **secondary** site with a Patroni standby cluster running GitLab 14.5 and later 257 2581. SSH to every Sidekiq, PostgresSQL, and Gitaly node in the **secondary** site and run one of the following commands: 259 260 - To promote the secondary node to primary: 261 262 ```shell 263 sudo gitlab-ctl geo promote 264 ``` 265 266 - To promote the secondary node to primary **without any further confirmation**: 267 268 ```shell 269 sudo gitlab-ctl geo promote --force 270 ``` 271 2721. SSH into each Rails node on your **secondary** site and run one of the following commands: 273 274 - To promote the secondary node to primary: 275 276 ```shell 277 sudo gitlab-ctl geo promote 278 ``` 279 280 - To promote the secondary node to primary **without any further confirmation**: 281 282 ```shell 283 sudo gitlab-ctl geo promote --force 284 ``` 285 2861. Verify you can connect to the newly-promoted **primary** site using the URL used 287 previously for the **secondary** site. 2881. If successful, the **secondary** site is now promoted to the **primary** site. 289 290#### Promoting a **secondary** site with a Patroni standby cluster running GitLab 14.4 and earlier 291 292WARNING: 293The `gitlab-ctl promote-to-primary-node` and `gitlab-ctl promoted-db` commands are 294deprecated in GitLab 14.5 and later, and are scheduled to [be removed in GitLab 15.0](https://gitlab.com/gitlab-org/gitlab/-/issues/345207). 295Use `gitlab-ctl geo promote` instead. 296 297The `gitlab-ctl promote-to-primary-node` command cannot be used yet in 298conjunction with a Patroni standby cluster, as it can only perform changes on 299a **secondary** with only a single node. Instead, you must do this manually. 300 3011. SSH in to the Standby Leader database node in the **secondary** site and trigger PostgreSQL to 302 promote to read-write: 303 304 ```shell 305 sudo gitlab-ctl promote-db 306 ``` 307 3081. Edit `/etc/gitlab/gitlab.rb` on every application and Sidekiq nodes in the secondary to reflect its new status as primary by removing any of the following lines that might be present: 309 310 ```ruby 311 geo_secondary_role['enable'] = true 312 roles ['geo_secondary_role'] 313 ``` 314 3151. Edit `/etc/gitlab/gitlab.rb` on every Patroni node in the secondary to disable the standby cluster: 316 317 ```ruby 318 patroni['standby_cluster']['enable'] = false 319 ``` 320 3211. Reconfigure GitLab on each machine for the changes to take effect: 322 323 ```shell 324 sudo gitlab-ctl reconfigure 325 ``` 326 3271. Promote the **secondary** to **primary**. SSH into a single application server and execute: 328 329 ```shell 330 sudo gitlab-rake geo:set_secondary_as_primary 331 ``` 332 3331. Verify you can connect to the newly-promoted **primary** using the URL used 334 previously for the **secondary**. 3351. If successful, the **secondary** site is now promoted to the **primary** site. 336 337#### Promoting a **secondary** site with an external PostgreSQL database running GitLab 14.5 and later 338 339The `gitlab-ctl geo promote` command can be used in conjunction with 340an external PostgreSQL database, but it can only perform changes on 341a **secondary** PostgreSQL database managed by Omnibus. 342You must promote the replica database associated with the **secondary** 343site first. 344 3451. Promote the replica database associated with the **secondary** site. This 346 sets the database to read-write. The instructions vary depending on where your database is hosted: 347 - [Amazon RDS](https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_ReadRepl.html#USER_ReadRepl.Promote) 348 - [Azure PostgreSQL](https://docs.microsoft.com/en-us/azure/postgresql/howto-read-replicas-portal#stop-replication) 349 - [Google Cloud SQL](https://cloud.google.com/sql/docs/mysql/replication/manage-replicas#promote-replica) 350 - For other external PostgreSQL databases, save the following script in your 351 secondary node, for example `/tmp/geo_promote.sh`, and modify the connection 352 parameters to match your environment. Then, execute it to promote the replica: 353 354 ```shell 355 #!/bin/bash 356 357 PG_SUPERUSER=postgres 358 359 # The path to your pg_ctl binary. You may need to adjust this path to match 360 # your PostgreSQL installation 361 PG_CTL_BINARY=/usr/lib/postgresql/10/bin/pg_ctl 362 363 # The path to your PostgreSQL data directory. You may need to adjust this 364 # path to match your PostgreSQL installation. You can also run 365 # `SHOW data_directory;` from PostgreSQL to find your data directory 366 PG_DATA_DIRECTORY=/etc/postgresql/10/main 367 368 # Promote the PostgreSQL database and allow read/write operations 369 sudo -u $PG_SUPERUSER $PG_CTL_BINARY -D $PG_DATA_DIRECTORY promote 370 ``` 371 3721. SSH to every Sidekiq, PostgresSQL, and Gitaly node in the **secondary** site and run one of the following commands: 373 374 - To promote the secondary node to primary: 375 376 ```shell 377 sudo gitlab-ctl geo promote 378 ``` 379 380 - To promote the secondary node to primary **without any further confirmation**: 381 382 ```shell 383 sudo gitlab-ctl geo promote --force 384 ``` 385 3861. SSH into each Rails node on your **secondary** site and run one of the following commands: 387 388 - To promote the secondary node to primary: 389 390 ```shell 391 sudo gitlab-ctl geo promote 392 ``` 393 394 - To promote the secondary node to primary **without any further confirmation**: 395 396 ```shell 397 sudo gitlab-ctl geo promote --force 398 ``` 399 4001. Verify you can connect to the newly-promoted **primary** site using the URL used 401 previously for the **secondary** site. 4021. If successful, the **secondary** site is now promoted to the **primary** site. 403 404#### Promoting a **secondary** site with an external PostgreSQL database running GitLab 14.4 and earlier 405 406WARNING: 407The `gitlab-ctl promote-to-primary-node` and `gitlab-ctl promoted-db` commands are 408deprecated in GitLab 14.5 and later, and are scheduled to [be removed in GitLab 15.0](https://gitlab.com/gitlab-org/gitlab/-/issues/345207). 409Use `gitlab-ctl geo promote` instead. 410 411The `gitlab-ctl promote-to-primary-node` command cannot be used in conjunction with 412an external PostgreSQL database, as it can only perform changes on a **secondary** 413node with GitLab and the database on the same machine. As a result, a manual process is 414required: 415 4161. Promote the replica database associated with the **secondary** site. This 417 sets the database to read-write. The instructions vary depending on where your database is hosted: 418 - [Amazon RDS](https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_ReadRepl.html#USER_ReadRepl.Promote) 419 - [Azure PostgreSQL](https://docs.microsoft.com/en-us/azure/postgresql/howto-read-replicas-portal#stop-replication) 420 - [Google Cloud SQL](https://cloud.google.com/sql/docs/mysql/replication/manage-replicas#promote-replica) 421 - For other external PostgreSQL databases, save the following script in your 422 secondary node, for example `/tmp/geo_promote.sh`, and modify the connection 423 parameters to match your environment. Then, execute it to promote the replica: 424 425 ```shell 426 #!/bin/bash 427 428 PG_SUPERUSER=postgres 429 430 # The path to your pg_ctl binary. You may need to adjust this path to match 431 # your PostgreSQL installation 432 PG_CTL_BINARY=/usr/lib/postgresql/10/bin/pg_ctl 433 434 # The path to your PostgreSQL data directory. You may need to adjust this 435 # path to match your PostgreSQL installation. You can also run 436 # `SHOW data_directory;` from PostgreSQL to find your data directory 437 PG_DATA_DIRECTORY=/etc/postgresql/10/main 438 439 # Promote the PostgreSQL database and allow read/write operations 440 sudo -u $PG_SUPERUSER $PG_CTL_BINARY -D $PG_DATA_DIRECTORY promote 441 ``` 442 4431. Edit `/etc/gitlab/gitlab.rb` on every node in the **secondary** site to 444 reflect its new status as **primary** by removing any of the following 445 lines that might be present: 446 447 ```ruby 448 geo_secondary_role['enable'] = true 449 roles ['geo_secondary_role'] 450 ``` 451 452 After making these changes [Reconfigure GitLab](../../restart_gitlab.md#omnibus-gitlab-reconfigure) 453 on each node so the changes take effect. 454 4551. Promote the **secondary** to **primary**. SSH into a single secondary application 456 node and execute: 457 458 ```shell 459 sudo gitlab-rake geo:set_secondary_as_primary 460 ``` 461 4621. Verify you can connect to the newly-promoted **primary** using the URL used 463 previously for the **secondary**. 4641. If successful, the **secondary** site is now promoted to the **primary** site. 465 466### Step 4. (Optional) Updating the primary domain DNS record 467 468Updating the DNS records for the primary domain to point to the **secondary** site 469to prevent the need to update all references to the primary domain to the 470secondary domain, like changing Git remotes and API URLs. 471 4721. SSH into the **secondary** site and login as root: 473 474 ```shell 475 sudo -i 476 ``` 477 4781. Update the primary domain's DNS record. After updating the primary domain's 479 DNS records to point to the **secondary** site, edit `/etc/gitlab/gitlab.rb` on the 480 **secondary** site to reflect the new URL: 481 482 ```ruby 483 # Change the existing external_url configuration 484 external_url 'https://<new_external_url>' 485 ``` 486 487 NOTE: 488 Changing `external_url` does not prevent access via the old secondary URL, as 489 long as the secondary DNS records are still intact. 490 4911. Reconfigure the **secondary** site for the change to take effect: 492 493 ```shell 494 gitlab-ctl reconfigure 495 ``` 496 4971. Execute the command below to update the newly promoted **primary** site URL: 498 499 ```shell 500 gitlab-rake geo:update_primary_node_url 501 ``` 502 503 This command uses the changed `external_url` configuration defined 504 in `/etc/gitlab/gitlab.rb`. 505 5061. For GitLab 11.11 through 12.7 only, you may need to update the **primary** 507 node's name in the database. This bug has been fixed in GitLab 12.8. 508 509 To determine if you need to do this, search for the 510 `gitlab_rails["geo_node_name"]` setting in your `/etc/gitlab/gitlab.rb` 511 file. If it is commented out with `#` or not found at all, then you 512 need to update the **primary** site's name in the database. You can search for it 513 like so: 514 515 ```shell 516 grep "geo_node_name" /etc/gitlab/gitlab.rb 517 ``` 518 519 To update the **primary** site's name in the database: 520 521 ```shell 522 gitlab-rails runner 'Gitlab::Geo.primary_node.update!(name: GeoNode.current_node_name)' 523 ``` 524 5251. Verify you can connect to the newly promoted **primary** using its URL. 526 If you updated the DNS records for the primary domain, these changes may 527 not have yet propagated depending on the previous DNS records TTL. 528 529### Step 5. (Optional) Add **secondary** Geo site to a promoted **primary** site 530 531Promoting a **secondary** site to **primary** site using the process above does not enable 532Geo on the new **primary** site. 533 534To bring a new **secondary** site online, follow the [Geo setup instructions](../index.md#setup-instructions). 535 536### Step 6. (Optional) Removing the secondary's tracking database 537 538Every **secondary** has a special tracking database that is used to save the status of the synchronization of all the items from the **primary**. 539Because the **secondary** is already promoted, that data in the tracking database is no longer required. 540 541The data can be removed with the following command: 542 543```shell 544sudo rm -rf /var/opt/gitlab/geo-postgresql 545``` 546 547If you have any `geo_secondary[]` configuration options enabled in your `gitlab.rb` 548file, these can be safely commented out or removed, and then [reconfigure GitLab](../../restart_gitlab.md#omnibus-gitlab-reconfigure) 549for the changes to take effect. 550 551## Promoting secondary Geo replica in multi-secondary configurations 552 553If you have more than one **secondary** site and you need to promote one of them, we suggest you follow 554[Promoting a **secondary** Geo site in single-secondary configurations](#promoting-a-secondary-geo-site-in-single-secondary-configurations) 555and after that you also need two extra steps. 556 557### Step 1. Prepare the new **primary** site to serve one or more **secondary** sites 558 5591. SSH into the new **primary** site and login as root: 560 561 ```shell 562 sudo -i 563 ``` 564 5651. Edit `/etc/gitlab/gitlab.rb`: 566 567 ```ruby 568 ## Enable a Geo Primary role (if you haven't yet) 569 roles ['geo_primary_role'] 570 571 ## 572 # Allow PostgreSQL client authentication from the primary and secondary IPs. These IPs may be 573 # public or VPC addresses in CIDR format, for example ['198.51.100.1/32', '198.51.100.2/32'] 574 ## 575 postgresql['md5_auth_cidr_addresses'] = ['<primary_node_ip>/32', '<secondary_node_ip>/32'] 576 577 # Every secondary server needs to have its own slot so specify the number of secondary nodes you're going to have 578 postgresql['max_replication_slots'] = 1 579 580 ## 581 ## Disable automatic database migrations temporarily 582 ## (until PostgreSQL is restarted and listening on the private address). 583 ## 584 gitlab_rails['auto_migrate'] = false 585 ``` 586 587 (For more details about these settings you can read [Configure the primary server](../setup/database.md#step-1-configure-the-primary-server)) 588 5891. Save the file and reconfigure GitLab for the database listen changes and 590 the replication slot changes to be applied: 591 592 ```shell 593 gitlab-ctl reconfigure 594 ``` 595 596 Restart PostgreSQL for its changes to take effect: 597 598 ```shell 599 gitlab-ctl restart postgresql 600 ``` 601 6021. Re-enable migrations now that PostgreSQL is restarted and listening on the 603 private address. 604 605 Edit `/etc/gitlab/gitlab.rb` and **change** the configuration to `true`: 606 607 ```ruby 608 gitlab_rails['auto_migrate'] = true 609 ``` 610 611 Save the file and reconfigure GitLab: 612 613 ```shell 614 gitlab-ctl reconfigure 615 ``` 616 617### Step 2. Initiate the replication process 618 619Now we need to make each **secondary** site listen to changes on the new **primary** site. To do that you need 620to [initiate the replication process](../setup/database.md#step-3-initiate-the-replication-process) again but this time 621for another **primary** site. All the old replication settings are overwritten. 622 623## Promoting a secondary Geo cluster in GitLab Cloud Native Helm Charts 624 625When updating a Cloud Native Geo deployment, the process for updating any node that is external to the secondary Kubernetes cluster does not differ from the non Cloud Native approach. As such, you can always defer to [Promoting a secondary Geo site in single-secondary configurations](#promoting-a-secondary-geo-site-in-single-secondary-configurations) for more information. 626 627The following sections assume you are using the `gitlab` namespace. If you used a different namespace when setting up your cluster, you should also replace `--namespace gitlab` with your namespace. 628 629WARNING: 630In GitLab 13.2 and 13.3, promoting a secondary site to a primary while the 631secondary is paused fails. Do not pause replication before promoting a 632secondary. If the site is paused, be sure to resume before promoting. This 633issue has been fixed in GitLab 13.4 and later. 634 635### Step 1. Permanently disable the **primary** cluster 636 637WARNING: 638If the **primary** site goes offline, there may be data saved on the **primary** site 639that has not been replicated to the **secondary** site. This data should be treated 640as lost if you proceed. 641 642If an outage on the **primary** site happens, you should do everything possible to 643avoid a split-brain situation where writes can occur in two different GitLab 644instances, complicating recovery efforts. So to prepare for the failover, you 645must disable the **primary** site: 646 647- If you have access to the **primary** Kubernetes cluster, connect to it and disable the GitLab webservice and Sidekiq pods: 648 649 ```shell 650 kubectl --namespace gitlab scale deploy gitlab-geo-webservice-default --replicas=0 651 kubectl --namespace gitlab scale deploy gitlab-geo-sidekiq-all-in-1-v1 --replicas=0 652 ``` 653 654- If you do not have access to the **primary** Kubernetes cluster, take the cluster offline and 655 prevent it from coming back online by any means at your disposal. 656 You might need to: 657 658 - Reconfigure the load balancers. 659 - Change DNS records (for example, point the primary DNS record to the 660 **secondary** site to stop usage of the **primary** site). 661 - Stop the virtual servers. 662 - Block traffic through a firewall. 663 - Revoke object storage permissions from the **primary** site. 664 - Physically disconnect a machine. 665 666### Step 2. Promote all **secondary** sites external to the cluster 667 668WARNING: 669If the secondary site [has been paused](../../geo/index.md#pausing-and-resuming-replication), this performs 670a point-in-time recovery to the last known state. 671Data that was created on the primary while the secondary was paused is lost. 672 673If you are running GitLab 14.5 and later: 674 6751. SSH to every Sidekiq, PostgresSQL, and Gitaly node in the **secondary** site and run one of the following commands: 676 677 - To promote the secondary node to primary: 678 679 ```shell 680 sudo gitlab-ctl geo promote 681 ``` 682 683 - To promote the secondary node to primary **without any further confirmation**: 684 685 ```shell 686 sudo gitlab-ctl geo promote --force 687 ``` 688 6891. SSH into each Rails node on your **secondary** site and run one of the following commands: 690 691 - To promote the secondary node to primary: 692 693 ```shell 694 sudo gitlab-ctl geo promote 695 ``` 696 697 - To promote the secondary node to primary **without any further confirmation**: 698 699 ```shell 700 sudo gitlab-ctl geo promote --force 701 ``` 702 703If you are running GitLab 14.4 and earlier: 704 7051. SSH in to the database node in the **secondary** and trigger PostgreSQL to 706 promote to read-write: 707 708 ```shell 709 sudo gitlab-ctl promote-db 710 ``` 711 712 In GitLab 12.8 and earlier, see [Message: `sudo: gitlab-pg-ctl: command not found`](../replication/troubleshooting.md#message-sudo-gitlab-pg-ctl-command-not-found). 713 7141. Edit `/etc/gitlab/gitlab.rb` on the database node in the **secondary** site to 715 reflect its new status as **primary** by removing any lines that enabled the 716 `geo_secondary_role`: 717 718 NOTE: 719 Depending on your architecture, these steps need to run on any GitLab node that is external to the **secondary** Kubernetes cluster. 720 721 ```ruby 722 ## In pre-11.5 documentation, the role was enabled as follows. Remove this line. 723 geo_secondary_role['enable'] = true 724 725 ## In 11.5+ documentation, the role was enabled as follows. Remove this line. 726 roles ['geo_secondary_role'] 727 ``` 728 729 After making these changes, [reconfigure GitLab](../../restart_gitlab.md#omnibus-gitlab-reconfigure) on the database node. 730 7311. Find the task runner pod: 732 733 ```shell 734 kubectl --namespace gitlab get pods -lapp=task-runner 735 ``` 736 7371. Promote the secondary: 738 739 ```shell 740 kubectl --namespace gitlab exec -ti gitlab-geo-task-runner-XXX -- gitlab-rake geo:set_secondary_as_primary 741 ``` 742 743### Step 3. Promote the **secondary** cluster 744 7451. Update the existing cluster configuration. 746 747 You can retrieve the existing configuration with Helm: 748 749 ```shell 750 helm --namespace gitlab get values gitlab-geo > gitlab.yaml 751 ``` 752 753 The existing configuration contains a section for Geo that should resemble: 754 755 ```yaml 756 geo: 757 enabled: true 758 role: secondary 759 nodeName: secondary.example.com 760 psql: 761 host: geo-2.db.example.com 762 port: 5431 763 password: 764 secret: geo 765 key: geo-postgresql-password 766 ``` 767 768 To promote the **secondary** cluster to a **primary** cluster, update `role: secondary` to `role: primary`. 769 770 If the cluster remains as a primary site, you can remove the entire `psql` section; it refers to the tracking database and is ignored whilst the cluster is acting as a primary site. 771 772 Update the cluster with the new configuration: 773 774 ```shell 775 helm upgrade --install --version <current Chart version> gitlab-geo gitlab/gitlab --namespace gitlab -f gitlab.yaml 776 ``` 777 7781. Verify you can connect to the newly promoted primary using the URL used previously for the secondary. 779 7801. Success! The secondary has now been promoted to primary. 781 782## Troubleshooting 783 784This section was moved to [another location](../replication/troubleshooting.md#fixing-errors-during-a-failover-or-when-promoting-a-secondary-to-a-primary-node). 785