Re: [Tracker] tracker-store leaking



On Tue, Oct 27, 2009 at 14:37, JÃrg Billeter<j bitron ch> wrote:
Hi,

On Tue, 2009-10-27 at 13:24 +0200, Tshepang Lekhonkhobe wrote:
Seems there's some bad leak in tracker-store; top claims it's upwards
of 300MB of RAM while valgrind output is summarised below. For a full
log, please find attached.

From looking at your log, it appears that you're running valgrind
without setting G_SLICE=always-malloc. Not setting this environment
variable may lead to bogus results. G_DEBUG=gc-friendly is recommended
to be set as well when using valgrind to further improve results.

If you see tracker-store using 300 MB private writable memory, there is
certainly something wrong, though. I see a couple small memory leaks
which I'll fix. My valgrind run doesn't report anything like the huge
memory leak you've been reported, however, I will continue to look into
this.

I'm still having trouble, and firing this:
G_DEBUG=gc-friendly G_SLICE=always-malloc valgrind --leak-check=full
--show-reachable=yes --leak-resolution=high
--log-file=/home/wena/valgrind.log /usr/libexec/tracker-store

gives the attached results which are summarised thus:
=9401== LEAK SUMMARY:
==9401==    definitely lost: 160 bytes in 2 blocks
==9401==    indirectly lost: 120 bytes in 10 blocks
==9401==      possibly lost: 345,371,316 bytes in 5,516,447 blocks
==9401==    still reachable: 28,586,598 bytes in 590,824 blocks
==9401==         suppressed: 4,912 bytes in 120 blocks


-- 
my place on the web:
floss-and-misc.blogspot.com

Attachment: valgrind.log.7z
Description: application/7z-compressed



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