juju-restore doesn't support HA
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
juju-core |
Fix Released
|
High
|
Horacio Durán |
Bug Description
Currently juju-restore preemptively says "if the state that was backed up was in HA, refuse to restore".
We should come up with some sort of answer for HA. If it just brings back up a single Non-HA state server, which we can then HA once it is back up and happy, that should be sufficient.
It is possible that we can just change the code that checks for if len(servers) > 1: error, into something that just tries to restore a single non-HA machine.
The assumption is that if you are using Restore, all state servers have died (otherwise you would just be running on one of the HA nodes.)
If it isn't possible to do restore at all after going HA, then we should refuse to run juju-backup unless there is *something* useful that can be done.
Related branches
- Juju Engineering: Pending requested
-
Diff: 290 lines (+124/-33)2 files modifiedcmd/plugins/juju-restore/restore.go (+71/-33)
doc/backup_and_restore.txt (+53/-0)
tags: | removed: 1.19.1 |
Changed in juju-core: | |
milestone: | none → 1.19.1 |
tags: | added: backup-restore |
Changed in juju-core: | |
assignee: | nobody → Horacio Durán (hduran-8) |
Changed in juju-core: | |
milestone: | 1.19.1 → 1.20.0 |
Changed in juju-core: | |
milestone: | 1.20.0 → next-stable |
Changed in juju-core: | |
status: | Triaged → In Progress |
Changed in juju-core: | |
status: | In Progress → Fix Committed |
Changed in juju-core: | |
milestone: | next-stable → 1.19.3 |
Changed in juju-core: | |
status: | Fix Committed → Fix Released |
I have a patch to propose here but I would like a CI test that takes this story in account so I am sure that my fix covers all possibilities.