503614 - Re-index Sphinx on SUMO Staging Server

Status: RESOLVED FIXED

(mozilla.org Graveyard :: Server Operations, task)

Description

[E. Cooper [:ecooper]]

While investigating bug 485329, we discovered a possible code conflict. We need to confirm this could be the problem we're facing on production. 

On staging, we'll need to 1) make sure PHP error reporting is on 2) re-index staging as per the instructions at https://bugzilla.mozilla.org/show_bug.cgi?id=488374#c1 and 3) report back the results

Comment 1

[Dave Miller [:justdave]]

chizu's in IRC and on the case.

Comment 2

[chizu]

The reindex failed with the same error from the indexer binary. Output was identical to before I added error_reporting(E_ALL); to indexer.php. Is that the correct file to modify to enable error reporting during indexing? First four lines of the modified file:

<?php
set_time_limit(0);
error_reporting(E_ALL);
mb_internal_encoding("UTF-8");

Comment 3

[Laura Thomson :laura]

chizu, can you just check svn info on there?  The suggested fix is in r29609.

Comment 4

[chizu]

It's on r29609. I ran it once more, still throws the error.

Comment 5

[Laura Thomson :laura]

OK, so suggested fix not working (as I commented in https://bugzilla.mozilla.org/show_bug.cgi?id=485329#c59 )  Can you work out if we're hitting some kind of resource limit?

Comment 6

[E. Cooper [:ecooper]]

A patch landed in r29686 that resolved an error I was able to reproduce using the latest sumo dumps. See https://bugzilla.mozilla.org/show_bug.cgi?id=485329#c62

Indexing process shouldn't fail now (hopefully).

Comment 7

[chizu]

Reindexed, and it worked! Everything ran and restarted cleanly on svn revision 29686.