Skip writes: > That looks like it is the problem.A I have sent BH an email asking them > about it.A By any chance do you know the name of the watchdog program > that they run to keep an eye on the user processes?A Or is it something > compiled into the kernel?A I have seen where sometimes depending on who > you get a hold of in tech support, they don't even know what their own > boxes are running and doing.
not a clue, I'm afraid. > Justin Mason wrote: > > Skip writes: > > > What do you know? I got permission from my web and email hosting > company (BlueHost) to run my own spamd process. Cool! Now I can have a > lot more control over the processing of my incoming mail, and I have > access to the logs! Well, after starting spamd, I was surprised after a > couple of minutes when it mysteriously wasn't running any more. After > running some experiments, it seems it is indeed stopping after just over > a minute. Here's the command line I'm using to start spamd: > > spamd -d -i 127.0.0.1 -p 6615 -C /home/xxxx/.spamassassin > --siteconfigpath=/home/xxxx/.spamassassin > --virtual-config-dir=/home/xxxx/.spamassassin/%l -s > /home/xxxx/.spamassassin/spamd.log --user-config -D -u xxxx > --pidfile=/home/xxxx/.spamassassin/spamd.pid --timeout-tcp=0 > --timeout-child=0 > > I tried it without the last two timeout parameters and they don't seem > to have any effect on this, and looking over the documentation, I > wouldn't have expected them to. > > Is this a normal behavior of spamd, that if it doesn't see any action > from spamc for a while, it just quits? By the way, I don't see anything > in the log that tells me spamd is shutting down or anything like that. > > I have been able to feed spamd some spam and it worked--I saw the scores > and everything, but again, a short time after I did the test, alas, > spamd shut down again. > > What did I miss? > > > that sounds a *lot* like Bluehost's automated CPU time limiting apps > shutting it down. Use "strace -p" to trace the process activity around > the 90 second mark, and see if it's getting a signal. > > --j.
