Hello Jim, You issue sound a bit familiar to me ( it might not be the same issue), I once had a hanging mysql connection from a linux webserver to a mysql running on windows, for every new connection it always used to wait 5 seconds before returning any results.
I did a lot of testing and network debugging, and I fixed it by adding a reverse lookup entry for the linux server. It might sound estrange but rember how mysql authentication works, and you can see that you can define many different hostnames for every user account (like: localhost, web01, etc), and the way it resolves the hostnames are by doing a reverse lookup to the ip. Mysql in windows behaves a bit different because when the reverse lookup fails, it then tries to resolve the hostname by netbios (but we don't run netbios on any linux boxes), and that had a timeout of guess what? 5 seconds. Anyways good luck with your current problem, and I hope this might help. Sergio On Wed, Oct 13, 2010 at 1:29 PM, Jim Isaacs <[email protected]> wrote: > Ok, I have tested and tested check and rechecked the logs. > I can't find anything unusual in the cherokee logs. I assume this is > because the connection is never made because the hanging of the request. > I can't find anything in the mysqld.log either. > I don't know how to read the audit.log but I don't see anything unusual. > Just my own internal ssh sessions. > > What I did find out was from testing again and again. > > It seems that after I have authenticated though MySQL things are fine. I > can click around at a reasonable rate within the realm and things are find. > I can reload at a reasonable rate and things are fine. > > It is when I do anyone of these things at an unreasonable rate that the > request hangs. Like say I refresh twice one after another. > Or or when I click a link once after I have clicked another but the request > is still currently in transit. > Or when I simply hit enter in the url re-requesting faster than usual. > Finally, when I do these things and get a hang, I can do it again one more > time to get a successful response. > > After witnessing this behavior, I can only assume this has to do with > something on the process level. > Maybe because another mysql process is trying to start before the another > one has run it's course? > > I'm not a low lever programmer or anything, but this is just my guess. > > Let me know what you guys think? > > > > On Oct 13, 2010, at 9:50 AM, Juan J. Martínez wrote: > > > El mié, 13-10-2010 a las 09:33 -0700, Jim Isaacs escribió: > >> I filed a bug here. > >> http://code.google.com/p/cherokee/issues/detail?id=1021&start=300 > >> > >> Since it isn't occurring with "normal" text file based or fixed list > authentication, I can only assume that it has to do with Cherokee making a > connection to MySQL. > >> > >> What logs do you guys think I could look for these kinds of errors? > > > > You're running Fedora, so the best logs are: > > > > /var/log/audit/* -> BTW, do you have selinux enabled in enforcing mode? > > /var/log/mysqld.log > > > > And of course cherokee logs (error log is a safe bet). > > > > The idea it's to triage if it's something local to your configuration or > > if there is actually a bug. > > > > Cheers, > > > > Juanjo > > > > -- > > jjm's home: http://www.usebox.net/jjm/ > > blackshell: http://blackshell.usebox.net/ > > ramble on: http://rambleon.usebox.net/ > > > > > > _______________________________________________ > Cherokee mailing list > [email protected] > http://lists.octality.com/listinfo/cherokee >
_______________________________________________ Cherokee mailing list [email protected] http://lists.octality.com/listinfo/cherokee
