Re: [Freeswitch-users] SQL ERR: library routine called out of sequence
You probably had it right in the last email. Polling the core's db from lua is not recommended. We use sla extensively and never once see your issue. On Feb 28, 2010 8:55 AM, "Yehavi Bourvine" <yehavi.bourvine@gmail.com> wrote: After a few more tests I *think* it is related to SLA. Since it cannot be reproduced consistenty I say "think". What I have is: - 80635 which is an SLA extension between two Polycoms. - 80636 which is a private extension on Polycom - 80632 which is a Cisco extension. - 86111 which is connected behind a Cisco SIP<->PSTN gateway. Since this gateway doesn't hav any login information it is not defined as a gateway but accepted via ACL. The tests that passed ok: - From/to 80636/80632 to 80635 (i.e. - all internal). - From/to 80635 to 80636/80632 (i.e. - all internal) - From/to 80636/80632 to 86111 and vice versa (i.e. via the gateway, single extensions). The one that fails after a few attempts: - From 86111 to 80635 (i.e. from the gateway to a shared extension). I hope that this gives some more clues. Thanks, __Yehavi: 2010/2/28 Yehavi Bourvine <yehavi.bourvine@gmail.com>
Hello,
I've installed a machine with CentOS-5.4 and the latest FreeSwitch
(16841M). The... _______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
Since I am aware that I am the only one who experience this problem, I am trying to ivestigate it, although I do not know the source code well... I removed all the datbase calls from our scripts and the problem still persists. It happens usually when two calls arrive to the same extension (usually to an SLA one) and one of them disconnects during the ringing phase. I've inserted some "printfs" in switch_core_sqldb.c and found that when this happens the variable dbh->io_mutex has a null value, thus no mutex is locked (as far as I understand it). I hope this might give some clues. Thanks! __Yehavi: 2010/3/1 Anthony Minessale <anthony.minessale@gmail.com>
You probably had it right in the last email. Polling the core's db from lua is not recommended. We use sla extensively and never once see your issue.
On Feb 28, 2010 8:55 AM, "Yehavi Bourvine" <yehavi.bourvine@gmail.com> wrote:
After a few more tests I *think* it is related to SLA. Since it cannot be reproduced consistenty I say "think".
What I have is: - 80635 which is an SLA extension between two Polycoms. - 80636 which is a private extension on Polycom - 80632 which is a Cisco extension. - 86111 which is connected behind a Cisco SIP<->PSTN gateway. Since this gateway doesn't hav any login information it is not defined as a gateway but accepted via ACL.
The tests that passed ok: - From/to 80636/80632 to 80635 (i.e. - all internal). - From/to 80635 to 80636/80632 (i.e. - all internal) - From/to 80636/80632 to 86111 and vice versa (i.e. via the gateway, single extensions).
The one that fails after a few attempts: - From 86111 to 80635 (i.e. from the gateway to a shared extension).
I hope that this gives some more clues.
Thanks, __Yehavi:
2010/2/28 Yehavi Bourvine <yehavi.bourvine@gmail.com>
Hello,
I've installed a machine with CentOS-5.4 and the latest FreeSwitch
(16841M). The...
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
_______________________________________________ FreeSWITCH-users mailing list FreeSWITCH-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
Yes accessing the db in the method you're doing is unsupport and not recommended. /b On Mar 2, 2010, at 6:56 AM, Yehavi Bourvine wrote:
Since I am aware that I am the only one who experience this problem, I am trying to ivestigate it, although I do not know the source code well...
I removed all the datbase calls from our scripts and the problem still persists. It happens usually when two calls arrive to the same extension (usually to an SLA one) and one of them disconnects during the ringing phase.
I've inserted some "printfs" in switch_core_sqldb.c and found that when this happens the variable dbh->io_mutex has a null value, thus no mutex is locked (as far as I understand it).
I hope this might give some clues.
Thanks! __Yehavi:
participants (3)
-
Anthony Minessale
-
Brian West
-
Yehavi Bourvine