https://bugzilla.wikimedia.org/show_bug.cgi?id=72366
--- Comment #14 from Bryan Davis <[email protected]> --- (In reply to Greg Grossmeier from comment #13) > (In reply to Bryan Davis from comment #11) > > > > If restarting hhvm is the cause, the frequency would depend on the level of > > gerrit +2 activity, the number of tests running and the time needed for the > > hhvm fcgi process to restart. Mostly impossible to calculate before hand. > > Without a method to de-pool and drain an hhvm fcgi service before restarting > > it it will be quite likely that there are some failed requests. > > Is this (where "this" == "method to de-pool and drain an hhvm fcgi service > before restarting") something that is planned to be added to scap for > production? If so, when can we expect that to be in place? Giuseppe is interested in adding etcd support to pybal that could make this possible but I don't think there is a timeline for it. Also beta does not use pybal so that is more work as well. I'm a bit concerned about the possibility of 503 storms in prod as well. -- You are receiving this mail because: You are the assignee for the bug. You are on the CC list for the bug. _______________________________________________ Wikibugs-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikibugs-l
