Docs: Add note about needing to do plugin reload on each node (#8108)
* Add note about needing to do this on each node Specifically calling this out will heed off operators doing this on a single node and thinking it is a bug that it didn't propagate to the other nodes, secondaries, etc. * Updated to reflect not needing to do registration on each
This commit is contained in:
parent
12fba35a7c
commit
57a3b72cde
|
@ -35,4 +35,9 @@ with either the `plugin` or `mounts` parameter respectively.
|
|||
Until step 4, the mount will still use plugin_v1, and when the reload is
|
||||
triggered, Vault will kill plugin_v1’s process and start a plugin_v2 process.
|
||||
|
||||
-> **Important:** Plugin reload of a new plugin binary must be
|
||||
performed on each Vault instance. Performing a plugin upgrade on a single
|
||||
instance or through a load balancer can result in mismatched
|
||||
plugin binaries within a cluster.
|
||||
|
||||
[plugin_reload_api]: /api/system/plugins-reload-backend.html
|
Loading…
Reference in a new issue