documentation for power recovery
Bug #1507065 reported by
Kevin Fox
This bug affects 5 people
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
kolla |
Expired
|
Undecided
|
Unassigned |
Bug Description
There is presently no documented procedure for recovering the various clusters in Kolla when a power failure takes out all cluster nodes at once.
This includes a power back on of the following services:
* Galera/MariaDB
* RabbitMQ
For Galera, the data should be recovered from the last written server.
For RabbitMQ, if the OpenStack Services just resubmit the messages, they can be freshly initialized, or recovered, whichever is more reasonable.
Changed in kolla: | |
assignee: | nobody → Sam Yaple (s8m) |
status: | New → Triaged |
Changed in kolla: | |
milestone: | none → ocata-1 |
tags: | added: doc |
Changed in kolla: | |
milestone: | ocata-1 → ocata-2 |
Changed in kolla: | |
milestone: | ocata-2 → ocata-3 |
Changed in kolla: | |
milestone: | ocata-3 → ocata-rc1 |
Changed in kolla: | |
milestone: | ocata-rc1 → pike-1 |
Changed in kolla: | |
milestone: | pike-2 → pike-3 |
Changed in kolla: | |
milestone: | pike-3 → pike-rc1 |
Changed in kolla: | |
milestone: | pike-rc1 → queens-1 |
Changed in kolla: | |
milestone: | queens-2 → queens-3 |
Changed in kolla: | |
milestone: | queens-3 → queens-rc1 |
Changed in kolla: | |
milestone: | queens-rc1 → queens-rc2 |
Changed in kolla: | |
milestone: | queens-rc2 → rocky-1 |
To post a comment you must log in.
At the summit RabbitMQ was discussed and to support lights out changes we wil be backporting clusterer
For galera it will need ot be a bit different. The idea is it should be automated, but im not sure the implementation will allow that. I was tasked with figuring that out and ill work on that right now.