CirrusSearch: Figure out why reindexing can get stuck waiting on shards to allocate when the shards have already allocated
Closed, ResolvedPublic
Actions

Assigned To

Authored By

	• Manybubbles
	Nov 19 2013, 2:57 PM

Description

The relevant part of the output:

Validating number of replicas...is 0 but should be 2...corrected
Waiting for all shards to start...
        4 remaining
        12 remaining
        12 remaining
        12 remaining
        12 remaining
        12 remaining
        12 remaining
        12 remaining

This was for dewikivoyage_content_1384800915.

Version: unspecified
Severity: normal
Whiteboard: cirrus_reenable

Details

Reference: bz57247

Event Timeline

• bzimport raised the priority of this task from to High.Nov 22 2014, 2:23 AM

• bzimport added a project: CirrusSearch.

• bzimport set Reference to bz57247.

• Manybubbles created this task.Nov 19 2013, 2:57 PM

Change 96366 had a related patch set uploaded by Manybubbles:
Switch shard startup monitoring using health api

https://gerrit.wikimedia.org/r/96366

That commit simplifies the counting process and increases the logging. If this happens again we'll know why.

Change 96366 merged by jenkins-bot:
Switch shard startup monitoring using health api

https://gerrit.wikimedia.org/r/96366

• Deskana moved this task from Inbox to Resolved/Invalid/Declined/Legacy on the CirrusSearch board.Apr 20 2015, 4:09 AM

CirrusSearch: Figure out why reindexing can get stuck waiting on shards to allocate when the shards have already allocatedClosed, ResolvedPublicActions

Description

Details

Event Timeline

CirrusSearch: Figure out why reindexing can get stuck waiting on shards to allocate when the shards have already allocated
Closed, ResolvedPublic
Actions