Pat Allan on 25 Aug, 2015 11:11 PM
Good to know its up again, though certainly keen to stop this
recurring downtime. So I can debug further, can you let me know
which versions of thinking-sphinx and flying-sphinx you're using in
Pat Allan on 26 Aug, 2015 05:09 AM
Just further to this: it looks like you've got a scheduled job
running a rebuild every day - rebuilds are only required when you
change the structure of your Sphinx indices (new/different fields
or attributes, new/removed indices). Think of it as somewhat
equivalent to db:migrate.
Each time you run that, Sphinx is stopped while the indexing
takes place, and this is perhaps where the errors have been coming
from over time? Today's instance, for some reason, the daemon
didn't restart, hence the longer outage.
It's far better to just run the index command
instead of a rebuild - this will keep your indexed
data up-to-date, and avoids stopping/starting the daemon.