zeoraid won't start until all backends come up

Bug #465834 reported by ChrisW
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
gocept.zeoraid
Fix Released
High
Christian Theune

Bug Description

zeoraid won't start until all backend zeo servers start up.

This is bad, since you may need to restart zeoraid when one of the backends is unavailable.

Revision history for this message
ChrisW (chris-simplistix) wrote :

The workaround for this is to remove the failed backend from the zeoraid config when you need to start it up and then add it back after zeoraid has come up.

When the backend is running again, you can add the config section back, do a zeoraid reload for the storage concerned and then a zeoraid recover once it's showing as present-but-failed.

Changed in gocept.zeoraid:
assignee: nobody → Christian Theune (ct-gocept)
status: New → Confirmed
Revision history for this message
Christian Theune (ctheune) wrote :

I reworked the initialisation code a bit and allow opening a storage to fail gracefully without disrupting the RAID. I also added some documentation about how to configure your back-end storages to allow ZEORaid to start up with failed storages.

Changed in gocept.zeoraid:
status: Confirmed → In Progress
Revision history for this message
Christian Theune (ctheune) wrote :

Partially solved in r105746, missing a test.

Changed in gocept.zeoraid:
importance: Undecided → High
Revision history for this message
ChrisW (chris-simplistix) wrote :

Verified as fixed in b6 and with "wait false" in each of the zeoclient sections in the zeoraid's zeo.conf

Changed in gocept.zeoraid:
status: In Progress → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.