We have setup a rhel6 app server to test with (all the production app servers are rhel5). this server is app02.stg.
Mediawiki installs, starts and runs on this instance, but it's slow. It's so slow that haproxy removes it from the load and doesn't send traffic to it. :(
See: https://admin.stg.fedoraproject.org/haproxy/proxy1/
and look at app02 there.
Determine why mediawiki is so slow on app02.stg and fix it so it handles requests normally like app01.stg (which is rhel5).
We tracked this down to stg running memcached on app01.stg, but not having a firewall rule allowing that access.
We still need to figure out how to fix this in puppet.
This should be fixed on the app machines, closing now.
Login to comment on this ticket.