Hi Stefan, On Fri, Sep 23, 2016 at 1:15 PM, Stefan Priebe - Profihost AG <[email protected]> wrote: > > a applied the latest patch (incl. fudge factor) to 1500 real life servers.
Thank you very much for testing in real conditions. > > Today i had the situation that no websites were delivered. The error.log > says: > > [Fri Sep 23 09:02:41.974245 2016] [mpm_event:error] [pid 6495:tid > 139689321596800] AH00485: scoreboard is full, not at MaxRequestWorkers > [Fri Sep 23 09:02:42.975267 2016] [mpm_event:error] [pid 6495:tid > 139689321596800] AH00485: scoreboard is full, not at MaxRequestWorkers > [Fri Sep 23 09:02:43.976217 2016] [mpm_event:error] [pid 6495:tid > 139689321596800] AH00485: scoreboard is full, not at MaxRequestWorkers > [Fri Sep 23 09:02:44.977276 2016] [mpm_event:error] [pid 6495:tid > 139689321596800] AH00485: scoreboard is full, not at MaxRequestWorkers > [Fri Sep 23 09:02:45.978314 2016] [mpm_event:error] [pid 6495:tid > 139689321596800] AH00485: scoreboard is full, not at MaxRequestWorkers > [Fri Sep 23 09:02:46.979222 2016] [mpm_event:error] [pid 6495:tid > 139689321596800] AH00485: scoreboard is full, not at MaxRequestWorkers > [Fri Sep 23 09:02:47.980304 2016] [mpm_event:error] [pid 6495:tid > 139689321596800] AH00485: scoreboard is full, not at MaxRequestWorkers > [Fri Sep 23 09:02:48.981213 2016] [mpm_event:error] [pid 6495:tid > 139689321596800] AH00485: scoreboard is full, not at MaxRequestWorkers > > Not sure if this is related to your patch. Hm, this never happened before you applied the patch? Did you observe a difference in the overall traffic (req/s, latency, ...) and/or CPU usage before it stopped responding? What's your MaxRequestsPerChild setting, and if it's zero, do you issue frequent graceful restarts? It could of course be caused by (a bug in )the patch, but I don't see for now how it could change the number of active/idle worker threads on its own, but with a different traffic. If it never happened before it may well be the case, though... It would help to know the (mod_)status of these workers. > I also found this one: > https://bz.apache.org/bugzilla/show_bug.cgi?id=53555 If you can try another (additive) patch on the faulty server, it would be awesome to test the one from [1], not sure it applies correctly with my 2.4.x patch though, possibly better with the trunk version over [1] (let me know otherwise). Regards, Yann. [1] https://bz.apache.org/bugzilla/show_bug.cgi?id=53555#c55
