Bug#513204: Progress?
clone 513204 -1 reassign -1 libmysqlclient15 found -1 5.0.51a-24 notfound -1 5.0.77-1 block 513204 by -1 thanks hi simon, On Tue, Apr 14, 2009 at 02:39:29PM +0100, Simon Waters wrote: > Any specific advice on how best to do this. As applying taskset to the apache > process forces all subprocesses including fastcgi processes to use the same > CPU, reducing the machine down to effectively a single CPU machine. (Note the yes, that's unfortunately the workaround at the moment. > I don't see a fix in the mysql client changelog for sid. Are there fixed > versions of the relevant packages available anywhere? i believe that sid is not affected by this due to later upstream code changes. > I'd have thought "unable to run LAMP stack reliably with PHP segfaulting" > would count as a major bug in terms of sending a fix out asap. Have you been > advised that it isn't suitable? yes but on the php side i don't think there's anything we can do. the mysql maintainer has been made aware of the problem but perhaps it's been dropped from his radar. i believe he's even found the problematic patch. i'm going to clone/reassign/block the bug to make sure that the problem stays visible. sean signature.asc Description: Digital signature
Bug#513204: Progress?
The workaround is described as taskset for PHP/Apache. Any specific advice on how best to do this. As applying taskset to the apache process forces all subprocesses including fastcgi processes to use the same CPU, reducing the machine down to effectively a single CPU machine. (Note the kernel doesn't reschedule processes to other processors lightly, so anyone seeing this issue had a need of more than one CPU for running the process in question at least once). I can probably force the FASTCGI processes to use additional processors which might resolve the issue on one of our servers, but down this road lies madness and micromanagement of which processor each process is using, and is unlikely to be applicable to all those affected. I don't see a fix in the mysql client changelog for sid. Are there fixed versions of the relevant packages available anywhere? I'd have thought "unable to run LAMP stack reliably with PHP segfaulting" would count as a major bug in terms of sending a fix out asap. Have you been advised that it isn't suitable? Simon -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org