On Wed, 25 Feb 2004, Matt Hoskins wrote:
> I'm running cyrus 2.2.3 murder. I've noticed on the frontend servers, processes
> will start to consume 99%
> of cpu. Mostly, it's the pop3d, pop3proxyd, imapd, imapproxyd. After a while, the
> process is killed
> (presumably by master), and in the logs I see a line like this:
>
> Feb 25 17:00:58 front1 master[3814]: service imap pid 29920 in BUSY state:
> terminated abnormally
>
> I did an strace on a process that was spinning out of control and this is what I see:
>
> select(11, [0 10], NULL, NULL, {60, 0}) = 1 (in [10], left {60, 0})
> time(NULL) = 1077749953
> open("/usr/cyrus/config/msg/shutdown", O_RDONLY) = -1 ENOENT (No such file or
> directory)
> time(NULL) = 1077749953
>
> Over and Over and Over and Over and Over again...It's in a infinite loop.
This is sort of curious, since it implies that it is just racing around
the main loop without ever stopping to get input...
I'd like to know more about what is going on with clients to these
connections, if that is possible. (I don't think I've ever seen this on
our solaris machines, though it is possible I've seen something similar on
our MX machines [with lmtpproxyd] which are linux, that I haven't looked
into that hard).
-Rob
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Rob Siemborski * Andrew Systems Group * Cyert Hall 207 * 412-268-7456
Research Systems Programmer * /usr/contributed Gatekeeper
---
Home Page: http://asg.web.cmu.edu/cyrus
Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
