Hello, Over the last few weeks, I've been running Cherokee for a few very small PHP sites. I'm running a separate FastCGI information source for each site. These are practically identical to the one setup by the PHP wizard, apart from the port (eg. localhost:47992) and running FastCGI as the site's user and group.
Overall, it's been performing magnificently (cherokee really is awesome!). Occasionally though, it crashes or fails to start a FastCGI process. There seem to be two recurring problems. The first one is, I think, a known one. The spawner starts spewing error messages like "(warning) spawner.c:261 - Couldn't unlock spawning semaphore.." into the Virtual Server's error log. The main cherokee process and cherokee-worker start hogging CPU, and the only remedy seems to restart cherokee completely. Is there any known way to prevent this, aside from running cherokee-worker as root? The second seems a bit more rare, and leaves this in the Virtual Server's log: > Cannot bind/listen socket - [98] Address already in use. > Couldn't create FastCGI listen socket on port localhost:47992 > [27/08/2009 18:44:50.260] (error) handler_fcgi.c:92 - Parsing error: unknown > type > [27/08/2009 18:44:52.550] (critical) server.c:1806 - PANIC: re-panic: > '/usr/share/cherokee/cherokee-panic 824', status 255 This happens every now and then, sometimes even right after a reboot. Additionally, cherokee-panic does't seem to send any email. I've got postfix installed and /usr/sbin/sendmail is present and working (root can receive messages sent by my user ppvg). Does anyone know why this happening and/or how I can debug it? Thanks! Peter-Paul
cherokee.conf.excerpt
Description: Binary data
_______________________________________________ Cherokee mailing list [email protected] http://lists.octality.com/listinfo/cherokee
