On Wed, Nov 26, 2014 at 4:54 PM, James Lamanna <[email protected]> wrote:
> Hi, > I have filed a bug for this here: > https://issues.asterisk.org/jira/browse/ASTERISK-24555 > > It appears that something in Asterisk 11 is continuously allocating frames > and not freeing them. > This causes Asterisk 11.14.0 to allocate almost 4GB of memory within a > couple of days and ~15000 calls. > This PBX is primarily used for faxing (Asterisk + IAXModem & T.38 Gateway) > so I'm wondering if the issue is in the SLIN codec or the T.38 gateway code. > > I've taken a quick glance and I haven't seen any mismatched ast_frdup() / > ast_frfree() yet. > > Thanks. > > I did a quick valgrind run and attempted a couple of faxes. I found 7 or so entries in the valgrind log with: ==30640== 211,203 bytes in 279 blocks are definitely lost in loss record 1,772 of 1,789 ==30640== at 0x4C267CC: calloc (vg_replace_malloc.c:467) ==30640== by 0x4DC50E: ast_frdup (utils.h:523) ==30640== by 0x47125F: __ast_queue_frame (channel.c:1284) ==30640== by 0x1EF75589: __do_deliver (chan_iax2.c:3102) ==30640== by 0x1EF76C5A: schedule_delivery (chan_iax2.c:4374) ==30640== by 0x1EF8F497: socket_process_helper (chan_iax2.c:12010) ==30640== by 0x1EF99C37: iax2_process_thread (chan_iax2.c:12030) ==30640== by 0x56C458: dummy_start (utils.c:1192) ==30640== by 0x5E359C9: start_thread (pthread_create.c:300) ==30640== by 0x271A66FF: ???
-- _____________________________________________________________________ -- Bandwidth and Colocation Provided by http://www.api-digital.com -- asterisk-dev mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-dev
