krinkle at deploy1001.eqiad.wmnet in /srv/mediawiki-staging (master) $ scap pull 00:15:54 Copying from deploy1001.eqiad.wmnet to deploy1001.eqiad.wmnet 00:15:54 Started rsync common 00:17:54 Finished rsync common (duration: 02m 00s) 00:17:54 Started scap-cdb-rebuild 00:17:55 Finished scap-cdb-rebuild (duration: 00m 00s) 00:17:55 Checking if php-fpm restart needed 00:17:55 Last output: sudo: a password is required 00:17:55 php-fpm restart failed!
The reason for running this command is that in order for a maintenance script to take into account staged changed, they must first be applied locally.
For example, running mwscript extensions/WikimediaMaintenance/dumpInterwiki.php.
For any other server, one typically uses scap pull for that. But while this worked, it seems to block for 2 minutes, and then report a php-fpm error.
I'm not sure what it is doing during these two minutes. Running scap pull on mwdebug1002 generally only takes a few seconds. I suspected that the sudo command (which one?) or the php-fpm failure is cause of it, but the timestamps don't agree. Maybe the timestamps are off?