fetchmail/procmail forkbomb? [ X focus locking under fvwm (printer conflict?)]

fetchmail/procmail forkbomb? [ X focus locking under fvwm (printer conflict?)]

Post by hawk » Wed, 10 Aug 2005 01:18:43



Hmm, the freezing of the interface might be related to (or caused by)
another problem.

I'm using fetchmail to get mail from three different servers, which then
feeds to procmail and spamassassin. I get a great mainy instances of
procmail and the system becomes unresponsive.

I tried the procmail and fetchmail man pages, hoping to find a way to
limit the number of simultaneous processes, but found limits for just
about everything else instead.

hawk
--
Richard E. Hawkins, J.D. Ph.D., Esq. /"\ ASCII ribbon campaign
XXXX@XXXXX.COM (814) 375-4846 \ / against HTML mail
Find commentary on law, economics, and X and postings.
other issues of the day at dochawk.org! / \
 
 
 

fetchmail/procmail forkbomb? [ X focus locking under fvwm (printer conflict?)]

Post by Matthias B » Thu, 11 Aug 2005 07:06:13


is your DNS working correctly?

mkb.

 
 
 

fetchmail/procmail forkbomb? [ X focus locking under fvwm (printer conflict?)]

Post by hawk » Fri, 12 Aug 2005 00:57:29

In article < XXXX@XXXXX.COM >,



It seemst to be; I don't get unreasonable delays in general.

But could that cause dozens of parallel procmail's to start?

thanks

hawk
--
Richard E. Hawkins, J.D. Ph.D., Esq. /"\ ASCII ribbon campaign
XXXX@XXXXX.COM (702) 434-3557 \ / against HTML mail
Find commentary on law, economics, and X and postings.
other issues of the day at dochawk.org! / \
 
 
 

fetchmail/procmail forkbomb? [ X focus locking under fvwm (printer conflict?)]

Post by Matthias B » Fri, 12 Aug 2005 01:33:16


Does the mail get delivered at all, or do you have cycles in your
procmail config?

mkb.
 
 
 

fetchmail/procmail forkbomb? [ X focus locking under fvwm (printer conflict?)]

Post by hawk » Fri, 12 Aug 2005 07:08:41

In article < XXXX@XXXXX.COM >,



Yep, it gets delivered.

And to make things odder, today it gave me absolutely no problems at
all, and there were no procmail running when the fetch was done (i've
seen as high as 70-something, and a couple of dozen more than once).

hawk, scratching his head
--
Richard E. Hawkins, J.D. Ph.D., Esq. /"\ ASCII ribbon campaign
XXXX@XXXXX.COM (702) 434-3557 \ / against HTML mail
Find commentary on law, economics, and X and postings.
other issues of the day at dochawk.org! / \
 
 
 

fetchmail/procmail forkbomb? [ X focus locking under fvwm (printer conflict?)]

Post by Matthias B » Fri, 12 Aug 2005 10:16:01


Truly weird, I'm using fetchmail+procmail myself and haven't seen
such a thing yet. And fetching a couple mails (even if it's a couple
hundred) and piping them through procmail is hardly a heavy workload..
The only thing I could think of would be that certain procmail rules
somehow pipe mails back into itself thus forming a cycle.

mkb.
 
 
 

fetchmail/procmail forkbomb? [ X focus locking under fvwm (printer conflict?)]

Post by hawk » Thu, 18 Aug 2005 14:16:31

In article < XXXX@XXXXX.COM >,



All of the rules either file in a mailbox or cat to /dev/null./

It looks like there's a problem creating processes, as it hangs others
as well. Can this be some kind of race condition.

The rules were copied from a dual p4 with a psu.edu address, while I'm
now using an athlon with earthlink dsl. The router assings it a 10.
address.

Could procmail or spamassassain be having trouble contacting something
to confirm/verify addressess?

Overall, when the machine does this, it feels like linux under load, but
worse . . .

It is very slow to drop the number of processes at first. As the number
drops, it drops faster, leading me to suspect some kind of lock
contention, which decreases in severity as the number of competitors
drops . . .

hawk
--
Richard E. Hawkins, J.D. Ph.D., Esq. /"\ ASCII ribbon campaign
XXXX@XXXXX.COM (702) 434-3557 \ / against HTML mail
Find commentary on law, economics, and X and postings.
other issues of the day at dochawk.org! / \