Thanks! The trac<->git changeset browser obviously has to grind through the repo all the time, so this is the likely candidate for io load. But the repo isn't all that large all things considered, it should easily fit into memory. Some filesystem tuning might be really benefitial, in particular mounting everything with noatime so you don't need to *write* to inodes when reading through the git repo. One could also try writeback caching but then we should barely write anything to the trac instance.
On Friday, October 11, 2013 9:28:33 PM UTC+1, Keith Clawson wrote: > > The physical machine that Trac is running has been experiencing heavy disk > IO from different processes for the past 3 days, which is likely > contributing to the problems. The image file is also located on a ZFS > volume, and I've seen odd issues with ZFS recently. I think I will move > Trac to the new server tomorrow. This will require around 30-60 minutes of > downtime, and I'll post a notice before I do that. > > -Keith > > On Friday, October 11, 2013 8:32:15 PM UTC+1, William wrote: >>> >>> On Fri, Oct 11, 2013 at 11:55 AM, Volker Braun <vbrau...@gmail.com> >>> wrote: >>> > Trac isn't using any swap. It is pulling quite a lot of CPU load. >>> Looking at >>> > the log, it seems somebody is indexing trac and pulling up old >>> changesets >>> > and timeline (both are notoriously slow in trac). For example: >>> > >>> > >>> http://trac.sagemath.org/timeline?from=2009-05-06T13%3A17%3A39-07%3A00&precision=second >>> >>> > >>> > (every entry is a db query, duh). And we don't have a robots.txt. >>> >>> Heh, but I also search trac using google :-) >>> >>> So we should have a robots.txt, but only restrict it, not stop it >>> completely...? >>> >>> > >>> > >>> > On Friday, October 11, 2013 5:59:57 PM UTC+1, William wrote: >>> >> >>> >> On Fri, Oct 11, 2013 at 9:52 AM, William Stein <wst...@gmail.com> >>> wrote: >>> >> > On Fri, Oct 11, 2013 at 9:39 AM, Jean-Pierre Flori < >>> jpf...@gmail.com> >>> >> > wrote: >>> >> >> Is it down again ? >>> >> > >>> >> > It's down. At the moment, I think Keith is the only one who knows >>> how >>> >> > to fix it easily, so we should probably wait for him to wake up / >>> >> > check his email. I'm sure he'll take additional steps today to >>> make >>> >> > trac more robust (hint, hint). >>> >> >>> >> It's back now -- maybe it's just flaky/slow due to swapping, since >>> >> Keith did at 20GB of swap "just in case" last night. >>> >> >>> >> William >>> >> >>> >> > >>> >> > -- William >>> >> > >>> >> > >>> >> >> >>> >> >> -- >>> >> >> You received this message because you are subscribed to the Google >>> >> >> Groups >>> >> >> "sage-devel" group. >>> >> >> To unsubscribe from this group and stop receiving emails from it, >>> send >>> >> >> an >>> >> >> email to sage-devel+...@googlegroups.com. >>> >> >> To post to this group, send email to sage-...@googlegroups.com. >>> >> >> Visit this group at http://groups.google.com/group/sage-devel. >>> >> >> For more options, visit https://groups.google.com/groups/opt_out. >>> >> > >>> >> > >>> >> > >>> >> > -- >>> >> > William Stein >>> >> > Professor of Mathematics >>> >> > University of Washington >>> >> > http://wstein.org >>> >> >>> >> >>> >> >>> >> -- >>> >> William Stein >>> >> Professor of Mathematics >>> >> University of Washington >>> >> http://wstein.org >>> > >>> > -- >>> > You received this message because you are subscribed to the Google >>> Groups >>> > "sage-devel" group. >>> > To unsubscribe from this group and stop receiving emails from it, send >>> an >>> > email to sage-devel+...@googlegroups.com. >>> > To post to this group, send email to sage-...@googlegroups.com. >>> > Visit this group at http://groups.google.com/group/sage-devel. >>> > For more options, visit https://groups.google.com/groups/opt_out. >>> >>> >>> >>> -- >>> William Stein >>> Professor of Mathematics >>> University of Washington >>> http://wstein.org >>> >> -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
