On Sun, Jan 4, 2009 at 8:46 AM, Kenneth Pronovici <[email protected]> wrote: > On Sat, Jan 3, 2009 at 7:53 PM, Kenneth Pronovici <[email protected]> wrote: >> On Sat, Jan 3, 2009 at 4:06 PM, Michael Olson <[email protected]> wrote: >>> Davor Ocelic <[email protected]> writes: >>>> On Sat, 03 Jan 2009 15:15:07 -0800 >>>> Michael Olson <[email protected]> wrote: >>>>> Do you have logging turned on? If so, try turning it off. That made >>>>> a huge difference for me. >>>> >>>> I remember the thread where we discussed this and identified logging >>>> to have this negative side effect. >>>> >>>> Does someone remember the thread this was in, or knows keywords to >>>> look it up in the mailing list archives? (IIRC, the subject didn't >>>> mention "logging" directly). >>> >>> The long and short of it is that writing to the same AFS file in many >>> (potentially) overlapping processes is a performance nightmare. >> >> Ok, I set the following in my procmailrc: >> >> VERBOSE = no >> LOGABSTRACT = no >> >> Hopefully, that will solve the problem. I don't really need logging >> enabled, anyway. > > Well, that didn't seem to make a difference. Overnight, my procmail > log remained empty, but 15 messages got into ~/Maildir. Most of them > were spam, but one was a legitimate message. > > Any other suggestions for things that I should try, or look at?
I have come up with a temporary solution, until we can figure out what's going on. I've written a small cron job to "rescue" the stuck mail. The job looks through ~/Maildir/new/ and ~/Maildir/cur/ and runs 'procmail -m' on each message it finds. This isn't perfect, but at least I won't lose any other important messages. Thanks again... KEN _______________________________________________ HCoop-Help mailing list [email protected] https://lists.hcoop.net/listinfo/hcoop-help
