Page MenuHomePhabricator

Scripts that wait for lagged database servers should ignore those out of rotation
Closed, ResolvedPublic

Description

  • s3 cluster **

10.0.0.234 db1 0
10.0.0.101 adler 0
10.0.0.238 db5 0
10.0.6.21 db11 0
10.0.6.28 db18 174366 ********
10.0.6.29 db19 0
10.0.6.31 db21 0
10.0.6.39 db29 1 *

Currently, scripts that wait for lagged servers to catch up before continuing operations don't distinguish between those in the system that will actually be used, and those which have been taken out of rotation (0 load placed on them).

To avoid disrupting activity when something needs a long catch-up, we currently have to remove it entirely from db.php -- this is pretty crappy because it then stops showing up in our regular lag tracking, so it's harder to see it when it catches up.

It would be nice to be able to have it in the system so it can be tracked, but still not actively used.


Version: unspecified
Severity: enhancement
Whiteboard: aklapper-moreinfo

Details

Reference
bz16993

Event Timeline

bzimport raised the priority of this task from to Low.Nov 21 2014, 10:30 PM
bzimport set Reference to bz16993.
bzimport added a subscriber: Unknown Object (MLST).

Five years later: Is this still relevant?

Which file is that about (if still in MediaWiki code base at all)?
core/includes/specials/SpecialLockdb.php and SpecialUnlockdb.php?
Or operations/mediawiki-config/docroot/noc/db.php ?

Five years later: Is this still relevant?

Which file is that about (if still in MediaWiki code base at all)?
core/includes/specials/SpecialLockdb.php and SpecialUnlockdb.php?
Or operations/mediawiki-config/docroot/noc/db.php ?

Aklapper changed the task status from Open to Stalled.Nov 25 2014, 7:47 PM
Aklapper lowered the priority of this task from Low to Lowest.Oct 16 2015, 1:24 PM
Aklapper added a subscriber: Aklapper.
aaron claimed this task.
aaron added a subscriber: aaron.

No longer seems to be an issue.