Cluster Upgrades

Verify that you can upgrade.

To upgrade to a new major version, you must first be on a production release of the previous version. The release does not need to be the latest production release of the previous version, but it must be a production release and not a testing release (alpha/beta).

Therefore, in order to upgrade to v23.2, you must be on a production release of v23.1.

1. If you are upgrading to v23.2 from a production release earlier than v23.1, or from a testing release (alpha/beta), first upgrade to a production release of v23.1. Be sure to complete all the steps.
2. Then return to this page and perform a second upgrade to v23.2.
3. If you are upgrading from a production release of v23.1, or from any earlier v23.2 patch release, you do not have to go through intermediate releases; continue to step 2.

Verify the overall health of your cluster using the DB Console. On the Overview:

• Under Node Status, make sure all nodes that should be live are listed as such. If any nodes are unexpectedly listed as suspect or dead, identify why the nodes are offline and either restart them or decommission them before beginning your upgrade. If there are dead and non-decommissioned nodes in your cluster, it will not be possible to finalize the upgrade (either automatically or manually).
• Under Replication Status, make sure there are 0 under-replicated and unavailable ranges. Otherwise, performing a rolling upgrade increases the risk that ranges will lose a majority of their replicas and cause cluster unavailability. Therefore, it's important to identify and resolve the cause of range under-replication and/or unavailability before beginning your upgrade.
• In the Node List:
  • Make sure all nodes are on the same version. If not all nodes are on the same version, upgrade them to the cluster's highest current version first, and then start this process over.
  • Make sure capacity and memory usage are reasonable for each node. Nodes must be able to tolerate some increase in case the new version uses more resources for your workload. Also go to Metrics > Dashboard: Hardware and make sure CPU percent is reasonable across the cluster. If there's not enough headroom on any of these metrics, consider adding nodes to your cluster before beginning your upgrade.

Review the backward-incompatible changes in v23.2 and deprecated features. If any affect your deployment, make the necessary changes before starting the rolling upgrade to v23.2.

Change the desired Docker image in the custom resource:

image:
  name: cockroachdb/cockroach:v23.2.16

Apply the new settings to the cluster:

$ kubectl apply -f example.yaml

The Operator will perform the staged update.

To check the status of the rolling upgrade, run kubectl get pods. The pods are restarted one at a time with the new image.

Verify that all pods have been upgraded by running:

$ kubectl get pods \
-o jsonpath='{range .items[*]}{.metadata.name}{"\t"}{.spec.containers[0].image}{"\n"}'

You can also check the CockroachDB version of each node in the DB Console.

Monitor the stability and performance of your cluster until you are comfortable with the upgrade (generally at least a day).

If you decide to roll back the upgrade, revert the image name in the custom resource and apply the new value.

To finalize the upgrade, re-enable auto-finalization:

1. Start the CockroachDB built-in SQL client. For example, if you followed the steps in Deploy CockroachDB with Kubernetes to launch a secure client pod, get a shell into the cockroachdb-client-secure pod:

   $ kubectl exec -it cockroachdb-client-secure \-- ./cockroach sql \
   --certs-dir=/cockroach/cockroach-certs \
   --host={cluster-name}-public
   

2. Re-enable auto-finalization:

   > RESET CLUSTER SETTING cluster.preserve_downgrade_option;
   

   After the upgrade to v23.2 is finalized, you may notice an increase in compaction activity due to a background migration within the storage engine. To observe the migration's progress, check the Compactions section of the Storage Dashboard in the DB Console or monitor the storage.marked-for-compaction-files time-series metric. When the metric's value nears or reaches 0, the migration is complete and compaction activity will return to normal levels. By default, the storage engine uses a compaction concurrency of 3. If you have sufficient IOPS and CPU headroom, you can consider increasing this setting via the COCKROACH_COMPACTION_CONCURRENCY environment variable. This may help to reshape the LSM more quickly in inverted LSM scenarios; and it can lead to increased overall performance for some workloads. Cockroach Labs strongly recommends testing your workload against non-default values of this setting.

3. Exit the SQL shell and pod:

   > \q
   

Verify that you can upgrade.

To upgrade to a new major version, you must first be on a production release of the previous version. The release does not need to be the latest production release of the previous version, but it must be a production release and not a testing release (alpha/beta).

Therefore, in order to upgrade to v23.2, you must be on a production release of v23.1.

1. If you are upgrading to v23.2 from a production release earlier than v23.1, or from a testing release (alpha/beta), first upgrade to a production release of v23.1. Be sure to complete all the steps.
2. Then return to this page and perform a second upgrade to v23.2.
3. If you are upgrading from any production release of v23.1, or from any earlier v23.2 patch release, you do not have to go through intermediate releases; continue to step 2.

Verify the overall health of your cluster using the DB Console. On the Overview:

• Under Node Status, make sure all nodes that should be live are listed as such. If any nodes are unexpectedly listed as suspect or dead, identify why the nodes are offline and either restart them or decommission them before beginning your upgrade. If there are dead and non-decommissioned nodes in your cluster, it will not be possible to finalize the upgrade (either automatically or manually).
• Under Replication Status, make sure there are 0 under-replicated and unavailable ranges. Otherwise, performing a rolling upgrade increases the risk that ranges will lose a majority of their replicas and cause cluster unavailability. Therefore, it's important to identify and resolve the cause of range under-replication and/or unavailability before beginning your upgrade.
• In the Node List:
  • Make sure all nodes are on the same version. If not all nodes are on the same version, upgrade them to the cluster's highest current version first, and then start this process over.
  • Make sure capacity and memory usage are reasonable for each node. Nodes must be able to tolerate some increase in case the new version uses more resources for your workload. Also go to Metrics > Dashboard: Hardware and make sure CPU percent is reasonable across the cluster. If there's not enough headroom on any of these metrics, consider adding nodes to your cluster before beginning your upgrade.

Review the backward-incompatible changes in v23.2 and deprecated features. If any affect your deployment, make the necessary changes before starting the rolling upgrade to v23.2.

Decide how the upgrade will be finalized.

By default, after all nodes are running the new version, the upgrade process will be auto-finalized. This will enable certain [features and performance improvements introduced in v23.2. After finalization, however, it will no longer be possible to perform a downgrade to v23.1. In the event of a catastrophic failure or corruption, the only option is to start a new cluster using the old binary and then restore from a backup created prior to the upgrade. For this reason, we recommend disabling auto-finalization so you can monitor the stability and performance of the upgraded cluster before finalizing the upgrade, but note that you will need to follow all of the subsequent directions, including the manual finalization in a later step.

1. Start the CockroachDB built-in SQL client. For example, if you followed the steps in Deploy CockroachDB with Kubernetes to launch a secure client pod, get a shell into the cockroachdb-client-secure pod:

   $ kubectl exec -it cockroachdb-client-secure \-- ./cockroach sql \
   --certs-dir=/cockroach-certs \
   --host=cockroachdb-public
   

2. Set the cluster.preserve_downgrade_option cluster setting to the version you are upgrading from:

   > SET CLUSTER SETTING cluster.preserve_downgrade_option = '["23.1"]';
   

3. Exit the SQL shell and delete the temporary pod:

   > \q
   

Add a partition to the update strategy defined in the StatefulSet. Only the pods numbered greater than or equal to the partition value will be updated. For a cluster with 3 pods (e.g., cockroachdb-0, cockroachdb-1, cockroachdb-2) the partition value should be 2:

$ kubectl patch statefulset cockroachdb \
-p='{"spec":{"updateStrategy":{"type":"RollingUpdate","rollingUpdate":{"partition":2}}}}'

statefulset.apps/cockroachdb patched

Kick off the upgrade process by changing the Docker image used in the CockroachDB StatefulSet:

$ kubectl patch statefulset cockroachdb \
--type='json' \
-p='[{"op": "replace", "path": "/spec/template/spec/containers/0/image", "value":"cockroachdb/cockroach:v23.2.16"}]'

statefulset.apps/cockroachdb patched

Check the status of your cluster's pods. You should see one of them being restarted:

$ kubectl get pods

NAME            READY     STATUS        RESTARTS   AGE
cockroachdb-0   1/1       Running       0          2m
cockroachdb-1   1/1       Running       0          2m
cockroachdb-2   0/1       Terminating   0          1m
...

After the pod has been restarted with the new image, start the CockroachDB built-in SQL client:

$ kubectl exec -it cockroachdb-client-secure \-- ./cockroach sql \
--certs-dir=/cockroach-certs \
--host=cockroachdb-public

Run the following SQL query to verify that the number of underreplicated ranges is zero:

SELECT sum((metrics->>'ranges.underreplicated')::DECIMAL)::INT AS ranges_underreplicated FROM crdb_internal.kv_store_status;

  ranges_underreplicated
--------------------------
                       0
(1 row)

This indicates that it is safe to proceed to the next pod.

Exit the SQL shell:

> \q

Decrement the partition value by 1 to allow the next pod in the cluster to update:

$ kubectl patch statefulset cockroachdb \
-p='{"spec":{"updateStrategy":{"type":"RollingUpdate","rollingUpdate":{"partition":1}}}}'

statefulset.apps/cockroachdb patched

Repeat steps 4-8 until all pods have been restarted and are running the new image (the final partition value should be 0).

Check the image of each pod to confirm that all have been upgraded:

$ kubectl get pods \
-o jsonpath='{range .items[*]}{.metadata.name}{"\t"}{.spec.containers[0].image}{"\n"}'

cockroachdb-0   cockroachdb/cockroach:v23.2.16
cockroachdb-1   cockroachdb/cockroach:v23.2.16
cockroachdb-2   cockroachdb/cockroach:v23.2.16
...

You can also check the CockroachDB version of each node in the DB Console.

If you disabled auto-finalization earlier, monitor the stability and performance of your cluster until you are comfortable with the upgrade (generally at least a day).

If you decide to roll back the upgrade, repeat the rolling restart procedure with the old binary.

To finalize the upgrade, re-enable auto-finalization:

1. Start the CockroachDB built-in SQL client:

   $ kubectl exec -it cockroachdb-client-secure \
   -- ./cockroach sql \
   --certs-dir=/cockroach-certs \
   --host=cockroachdb-public
   

2. Re-enable auto-finalization:

   > RESET CLUSTER SETTING cluster.preserve_downgrade_option;
   

   After the upgrade to v23.2 is finalized, you may notice an increase in compaction activity due to a background migration within the storage engine. To observe the migration's progress, check the Compactions section of the Storage Dashboard in the DB Console or monitor the storage.marked-for-compaction-files time-series metric. When the metric's value nears or reaches 0, the migration is complete and compaction activity will return to normal levels.

3. Exit the SQL shell and delete the temporary pod:

   > \q
   

Verify that you can upgrade.

To upgrade to a new major version, you must first be on a production release of the previous version. The release does not need to be the latest production release of the previous version, but it must be a production release and not a testing release (alpha/beta).

Therefore, in order to upgrade to v23.2, you must be on a production release of v23.1.

1. If you are upgrading to v23.2 from a production release earlier than v23.1, or from a testing release (alpha/beta), first upgrade to a production release of v23.1. Be sure to complete all the steps.
2. Then return to this page and perform a second upgrade to v23.2.
3. If you are upgrading from any production release of v23.1, or from any earlier v23.2 patch release, you do not have to go through intermediate releases; continue to step 2.

Verify the overall health of your cluster using the DB Console. On the Overview:

• Under Node Status, make sure all nodes that should be live are listed as such. If any nodes are unexpectedly listed as suspect or dead, identify why the nodes are offline and either restart them or decommission them before beginning your upgrade. If there are dead and non-decommissioned nodes in your cluster, it will not be possible to finalize the upgrade (either automatically or manually).
• Under Replication Status, make sure there are 0 under-replicated and unavailable ranges. Otherwise, performing a rolling upgrade increases the risk that ranges will lose a majority of their replicas and cause cluster unavailability. Therefore, it's important to identify and resolve the cause of range under-replication and/or unavailability before beginning your upgrade.
• In the Node List:
  • Make sure all nodes are on the same version. If not all nodes are on the same version, upgrade them to the cluster's highest current version first, and then start this process over.
  • Make sure capacity and memory usage are reasonable for each node. Nodes must be able to tolerate some increase in case the new version uses more resources for your workload. Also go to Metrics > Dashboard: Hardware and make sure CPU percent is reasonable across the cluster. If there's not enough headroom on any of these metrics, consider adding nodes to your cluster before beginning your upgrade.

Review the backward-incompatible changes in v23.2 and deprecated features. If any affect your deployment, make the necessary changes before starting the rolling upgrade to v23.2.

Decide how the upgrade will be finalized.

By default, after all nodes are running the new version, the upgrade process will be auto-finalized. This will enable certain [features and performance improvements introduced in v23.2. After finalization, however, it will no longer be possible to perform a downgrade to v23.1. In the event of a catastrophic failure or corruption, the only option is to start a new cluster using the old binary and then restore from a backup created prior to the upgrade. For this reason, we recommend disabling auto-finalization so you can monitor the stability and performance of the upgraded cluster before finalizing the upgrade, but note that you will need to follow all of the subsequent directions, including the manual finalization in a later step.

1. Get a shell into the pod with the cockroach binary created earlier and start the CockroachDB built-in SQL client:

   $ kubectl exec -it cockroachdb-client-secure \
   -- ./cockroach sql \
   --certs-dir=/cockroach-certs \
   --host=my-release-cockroachdb-public
   

2. Set the cluster.preserve_downgrade_option cluster setting to the version you are upgrading from:

   > SET CLUSTER SETTING cluster.preserve_downgrade_option = '["23.1"]';
   

3. Exit the SQL shell and delete the temporary pod:

   > \q
   

Add a partition to the update strategy defined in the StatefulSet. Only the pods numbered greater than or equal to the partition value will be updated. For a cluster with 3 pods (e.g., cockroachdb-0, cockroachdb-1, cockroachdb-2) the partition value should be 2:

$ helm upgrade \
my-release \
cockroachdb/cockroachdb \
--set statefulset.updateStrategy.rollingUpdate.partition=2

Kick off the upgrade process by changing the Docker image used in the CockroachDB StatefulSet:

$ kubectl delete job my-release-cockroachdb-init

$ helm upgrade \
my-release \
cockroachdb/cockroachdb \
--set image.tag=v23.2.16 \
--reuse-values

Check the status of your cluster's pods. You should see one of them being restarted:

$ kubectl get pods

NAME                                READY     STATUS              RESTARTS   AGE
my-release-cockroachdb-0            1/1       Running             0          2m
my-release-cockroachdb-1            1/1       Running             0          3m
my-release-cockroachdb-2            0/1       ContainerCreating   0          25s
my-release-cockroachdb-init-nwjkh   0/1       ContainerCreating   0          6s
...

After the pod has been restarted with the new image, start the CockroachDB built-in SQL client:

$ kubectl exec -it cockroachdb-client-secure \
-- ./cockroach sql \
--certs-dir=/cockroach-certs \
--host=my-release-cockroachdb-public

Run the following SQL query to verify that the number of underreplicated ranges is zero:

SELECT sum((metrics->>'ranges.underreplicated')::DECIMAL)::INT AS ranges_underreplicated FROM crdb_internal.kv_store_status;

  ranges_underreplicated
--------------------------
                       0
(1 row)

This indicates that it is safe to proceed to the next pod.

Exit the SQL shell:

> \q

Decrement the partition value by 1 to allow the next pod in the cluster to update:

$ helm upgrade \
my-release \
cockroachdb/cockroachdb \
--set statefulset.updateStrategy.rollingUpdate.partition=1 \

Repeat steps 4-8 until all pods have been restarted and are running the new image (the final partition value should be 0).

Check the image of each pod to confirm that all have been upgraded:

$ kubectl get pods \
-o jsonpath='{range .items[*]}{.metadata.name}{"\t"}{.spec.containers[0].image}{"\n"}'

my-release-cockroachdb-0    cockroachdb/cockroach:v23.2.16
my-release-cockroachdb-1    cockroachdb/cockroach:v23.2.16
my-release-cockroachdb-2    cockroachdb/cockroach:v23.2.16
...

You can also check the CockroachDB version of each node in the DB Console.

If you disabled auto-finalization earlier, monitor the stability and performance of your cluster until you are comfortable with the upgrade (generally at least a day).

If you decide to roll back the upgrade, repeat the rolling restart procedure with the old binary.

To finalize the upgrade, re-enable auto-finalization:

1. Get a shell into the pod with the cockroach binary created earlier and start the CockroachDB built-in SQL client:

   $ kubectl exec -it cockroachdb-client-secure \
   -- ./cockroach sql \
   --certs-dir=/cockroach-certs \
   --host=my-release-cockroachdb-public
   

2. Re-enable auto-finalization:

   > RESET CLUSTER SETTING cluster.preserve_downgrade_option;
   

   After the upgrade to v23.2 is finalized, you may notice an increase in compaction activity due to a background migration within the storage engine. To observe the migration's progress, check the Compactions section of the Storage Dashboard in the DB Console or monitor the storage.marked-for-compaction-files time-series metric. When the metric's value nears or reaches 0, the migration is complete and compaction activity will return to normal levels.

3. Exit the SQL shell and delete the temporary pod:

   > \q