Re: beagled eating lots of VM again



Well, here we are, 3 days after beagled was started and 1 day and
3:05:22 of CPU time later and beagled is up in the 585M/364M of VM/RSS
neighborhood.

I am sure you mentioned this somewhere before, but what backends are
you running ? If you have everything on by default, what backends are
actually active: beagle-info --index-info (when beagled is running)
will print non-zero count for the active backends.

If you are using Thunderbird backend, I suggest you disable that - use
beagle-settings or run beagled with "--backend -Thunderbird". There
was some memory consumption optimization in the EvoMail backend - I
dont remember if that was prior to 0.2.14.

If you start with BEAGLE_EXERCISE_THE_DOG=1, and leave it on till the
indexing/crawling is over - do you get a high memory usage after that
? Yes - would mean, there is something wrong in the crawling/indexing
stage. No - would mean, we need to investigate the query system,
scheduler and index-readers (the long living systems).

I'd sure love to have beagle with heap-shot running this long to see if
it gets this big with it and where the memory is going but it crashes
too soon.

I think there should be a way to disable the auto-snapshotting and
only snapshot on demand. BTW, if you turn on heap-shot and it crashes,
if it crashes after about 100 MB RSS, that is still useful. Send
me/joe the snapshots in that case. Always, attach the log files (if
too large, first 1000 lines and last 1000 lines).

Which mono version are you using ? If I patch the source to disable
auto-snapshotting, I am thinking of sending you just the binary dll
file.

- dBera

--
-----------------------------------------------------
Debajyoti Bera @ http://dtecht.blogspot.com
beagle / KDE fan
Mandriva / Inspiron-1100 user



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]