How to upgrade¶
Major version upgrade¶
This refers to an upgrade from MongoDB 6 to MongoDB 8.
After having upgraded your sharded cluster with (see MongoDB - How to upgrade), you can proceed to deploy a new Mongos VM application (see: Mongos - How to deploy)) and relate it with the config-server.
Please note that Mongos 6 and MongoDB 8 are not compatible, so you have to update them separately.
Please also note that Mongos 8 is released on Ubuntu 24.04 base, which means the host application has to be deployed as well using an Ubuntu 24.04 base.
Minor version upgrade¶
This refers to an upgrade between different charm revisions of Mongos 8.
Supported revisions¶
Only upgrade from these supported revisions in a production environment.
VM:
Revision 62
Revision 78
K8s:
Revision 54
Revision 69
Upgrade steps¶
Ensure your MongoDB deployment is healthy (active and idle), and not performing any long-running operations such as creating a backup.
Step 1: Pre-refresh check¶
Before refreshing, the charm needs to perform some preparatory tasks to define the refresh plan.
Run the pre-refresh-check action against the leader unit of your Mongos application:
juju run mongos/leader pre-refresh-check
juju run mongos-k8s/leader pre-refresh-check
Warning
Do not proceed if this action is unsuccessful.
If the action succeeds, copy down the rollback command. Keep the command available in case you need to roll back.
Step 2: Determine your current revision¶
Find your current revision with juju status:
juju status | grep mongos | head -1 | awk '{print $6}'
juju status | grep mongos-k8s | head -1 | awk '{print $6}'
Caution
If your current revision is not in the list, then upgrade is not supported.
In the case that your upgrade is not successful, you will need this revision number to roll back your application.
Step 3: Determine new revision¶
Choose a revision to upgrade to from the list of available revisions to upgrade to
Step 4: Upgrade the router¶
Using the revision from Step 3, upgrade your application
juju refresh mongos --revision=<revision number>
juju refresh mongos-k8s --revision=<revision number>
The upgrade should begin, and upgrade all deployed mongos routers. You can watch and wait for the upgrade to succeed with juju status --watch 1s.
After the first unit is upgraded, the refresh will be paused (pause-after-unit-refresh config option is set to first by default).
Run the following command to resume refresh:
juju run mongos/X resume-refresh # X is the next unit to upgrade.
juju run mongos-k8s/leader --revision=<revision number>
If all units show active and idle statuses after upgrading, then your upgrade succeeded.
If your upgrade failed, then juju status will show that the unit is blocked with the message Unhealthy after refresh. You can do the following:
Recommended: Wait for 5-10 minutes to see if the router can heal itself
Recommended: Roll back. Use the rollback command from the pre-refresh check.
Not Recommended: Force an upgrade by running
juju run mongos/leader force-refresh-start. Please note this can break your database and is not recommended.