My main asterisk server sometimes suddenly stop responding. When I'm at work, I always have a ssh console opened with "asterisk -r" in it. I have 23 sip hard and soft phones. Normaly, when I do a "sip show peers", I see all 23 sip users and the command finishes with 23 sip peers [Monitored: 8 online, 15 offline Unmonitored: 0 online, 0 offline] for exemple. Now, when my server begin not to respond, when I do a sip show peers, I only see a subset of the users, and not the summary line. If I do other commands like "core show channels", it simply shows nothing at all. Not even a "0 active channels". Reload, or any other similar commands goes unansered. The "core" works as it still answers calls on the PRI, but sip extensions can't be reached, etc. When it does this, I have to completly stop asterisk, often having to kill the processes, and then start it back again. We have a PRI with MTS/Allstream on a sangoma quad-port card, one other port is used to hookup a dial-up RAS, another one is used to do a point-to-point link with our office. The server is CentOS 5, and I am up to date with Asterisk 1.4.20. On a local support forum, I have been suggested that I compile asterisk with DEBUG_THREADS and do a "core show locks" when it happens. core show locks simply shows nothing...
On Tue, 27 May 2008 13:11:22 -0400, Nicolas Ross wrote> Now, when my server begin not to respond, when I do a sip show peers, I only > see a subset of the users, and not the summary line. If I do other commands > like "core show channels", it simply shows nothing at all. Not even a "0 > active channels". Reload, or any other similar commands goes unansered. The > "core" works as it still answers calls on the PRI, but sip extensions can't > be reached, etc.I have had asterisk do this if its hanging on a DNS request.. you got any peers in there using DNS?
Nicolas, Please contribute. There's a bug filed. http://bugs.digium.com/view.php?id=12709 Thanks, Mark. -----Original Message----- From: asterisk-users-bounces at lists.digium.com [mailto:asterisk-users-bounces at lists.digium.com] On Behalf Of Nicolas Ross Sent: May 27, 2008 1:11 PM To: asterisk-users at lists.digium.com Subject: [asterisk-users] Asterisk just stops working... My main asterisk server sometimes suddenly stop responding. When I'm at work, I always have a ssh console opened with "asterisk -r" in it. I have 23 sip hard and soft phones. Normaly, when I do a "sip show peers", I see all 23 sip users and the command finishes with 23 sip peers [Monitored: 8 online, 15 offline Unmonitored: 0 online, 0 offline] for exemple. Now, when my server begin not to respond, when I do a sip show peers, I only see a subset of the users, and not the summary line. If I do other commands like "core show channels", it simply shows nothing at all. Not even a "0 active channels". Reload, or any other similar commands goes unansered. The "core" works as it still answers calls on the PRI, but sip extensions can't be reached, etc. When it does this, I have to completly stop asterisk, often having to kill the processes, and then start it back again. We have a PRI with MTS/Allstream on a sangoma quad-port card, one other port is used to hookup a dial-up RAS, another one is used to do a point-to-point link with our office. The server is CentOS 5, and I am up to date with Asterisk 1.4.20. On a local support forum, I have been suggested that I compile asterisk with DEBUG_THREADS and do a "core show locks" when it happens. core show locks simply shows nothing... _______________________________________________ -- Bandwidth and Colocation Provided by http://www.api-digital.com -- asterisk-users mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
Maybe you should play around with FreeSwitch. It was written because of deadlocks and the issues that have been around since the the beginning.... http://www.freeswitch.org/node/117 "That's when I first heard it. That word. The one I can never forget, deadlock. That was the first time, but it wasn't the last. I learned all about the GNU debugger that day and it was just the first of many incidents. Deadlock in the queue app. Deadlock in the manager, Avoiding Deadlock on my console. It was starting to get to me a little but I kept going. By this time I was also quite familiar with the term Segmentation Fault another foe to the computer developer. After about a year's time wrestling with bugs I found myself a lot more well-versed in the C programming language than I even imagined and near Jedi caliber debugging skills. ...We've all seen the infamous "Avoiding initial deadlock" message which essentially is an attempt to lock a channel 10 times and if still won't lock, just go ahead and forget about the lock." Thanks, Steve Totaro On Thu, May 29, 2008 at 9:19 AM, Mark Hamilton <mark.h at cage151.com> wrote:> You're right Steve. > It's always been around.. and I think it's being ignored or simply can't be > fixed, or I have no idea what it is. The bug is receiving trickles of > response from the developers. > > It is so critical, I crash left and right every other day right now and I > don't know where to turn to. There's no response, and my hands are getting > sweaty. > > -----Original Message----- > From: asterisk-users-bounces at lists.digium.com > [mailto:asterisk-users-bounces at lists.digium.com] On Behalf Of Steve Totaro > Sent: May 28, 2008 11:46 AM > To: Asterisk Users Mailing List - Non-Commercial Discussion > Subject: Re: [asterisk-users] Asterisk just stops working... > > On Wed, May 28, 2008 at 10:44 AM, Nicolas Ross > <rossnick-lists at cybercat.ca> wrote: >>> The best approach is compile it with debug options, (make menuselect), >>> set the compile to DEBUG on, then make clean, make, make install. >>> >>> THen, fire up under the debugger. When you're all locked up, use ^C to >>> halt and leave the debugger in command, and do the "thread apply all bt" >>> thing. That should be revealing. >> >> Thanks, >> >> I'm not that much a developper. So I'll recompile with DEBUG_CHANNEL_LOCKS >> and DEBUG_THREADS. >> >> How do I start it up ? I curently start asterisk with redhat-style script > in >> /etc/rc.d, with safe_asterisk I think... >> >> Thanks, >> >> Nicolas >> > > BTW, not sure what it is worth but this has been an issue in Asterisk > since early 1.2. I always just worked around it since it was a minor > nuisance. > > Just an observation, but it seemed that if I did not issue reloads or > keep a connection to the Asterisk CLI, the issue would not occur. I > don't know if that is fact or just imagination... > > Thanks, > Steve Totaro > > _______________________________________________ > -- Bandwidth and Colocation Provided by http://www.api-digital.com -- > > asterisk-users mailing list > To UNSUBSCRIBE or update options visit: > http://lists.digium.com/mailman/listinfo/asterisk-users > > > _______________________________________________ > -- Bandwidth and Colocation Provided by http://www.api-digital.com -- > > asterisk-users mailing list > To UNSUBSCRIBE or update options visit: > http://lists.digium.com/mailman/listinfo/asterisk-users >
That is because they are busy with commas and pipes as delimiters and deprecating things that "just work" the way they are.... You know, the things that really matter. The dealbreakers. The things that make "Asterisk Ready". Thanks, Steve Totaro On Fri, May 30, 2008 at 10:43 AM, Mark Hamilton <mark.h at cage151.com> wrote:> Al, > > Either way, it doesn't matter. Because when the gdb is submitted afterall, > there is still nothing going on Mantis to try and fix it. > > For example, look at my deadlock issue I've reported - definitely critical. > And as you say it should be included in the "Next best thread that asks is > asterisk ready?" because this critical bug that I've filed has hardly > garnered any attention from the devels. > > Mark. > > -----Original Message----- > From: asterisk-users-bounces at lists.digium.com > [mailto:asterisk-users-bounces at lists.digium.com] On Behalf Of Al Baker > Sent: May 29, 2008 4:25 AM > To: Asterisk Users Mailing List - Non-Commercial Discussion > Subject: Re: [asterisk-users] Asterisk just stops working... > > Quote > > THen, fire up under the debugger. When you're all locked up, use ^C to >> halt and leave the debugger in command, and do the "thread apply all bt" >> thing. That should be revealing. > > If I may suggest , what would REALLY be 'Revealing' is if you could be > just a bit more clear in your explanation and about 900% LESS in the > techno babble. > While the thought is in the Right Place do you REALLY expect anybody to > know what the hell you mean by : > > When you're all locked up, use ^C to >> halt and leave the debugger in command, and do the "thread apply all bt" >> thing. That should be revealing > > > *Just a thought* > > Sherwood McGowan wrote: >> Steve Murphy wrote: >> >>> On Tue, 2008-05-27 at 14:31 -0400, Nicolas Ross wrote: >>> >>> >>>>> On Tue, 27 May 2008 13:11:22 -0400, Nicolas Ross wrote >>>>> >>>>> >>>>>> Now, when my server begin not to respond, when I do a sip show peers, > I >>>>>> only >>>>>> see a subset of the users, and not the summary line. If I do other >>>>>> commands >>>>>> like "core show channels", it simply shows nothing at all. Not even a > "0 >>>>>> active channels". Reload, or any other similar commands goes > unansered. >>>>>> The >>>>>> "core" works as it still answers calls on the PRI, but sip extensions >>>>>> can't >>>>>> be reached, etc. >>>>>> >>>>>> >>>>> I have had asterisk do this if its hanging on a DNS request.. you got > any >>>>> peers in there >>>>> using DNS? >>>>> >>>>> >>>> I don,t beleive it's the case. I beleive asterisk would still answer to > the >>>> "core show channels" command if it were the case... >>>> >>>> >>> Not necessarily! If the channels are locked while the holdup occurred, >>> then, you won't be able to even look at them. >>> >>> >>> >>>> The clients (phones) are configured to register with a fqdn domain. The >>>> peers are type=friend, host-dynamic. >>>> >>>> I did change some changes to my dialplan so that I don't use queues and >>>> agents anymore... We'll see if it happens again. >>>> >>>> By the way, my asterisk doesn't crash, I don't have a core dump. And > since >>>> core show locks doesn't provide anything when I'm in that situation, how > >>>> could I contribute ? >>>> >>>> >>>> >>> The best approach is compile it with debug options, (make menuselect), >>> set the compile to DEBUG on, then make clean, make, make install. >>> >>> THen, fire up under the debugger. When you're all locked up, use ^C to >>> halt and leave the debugger in command, and do the "thread apply all bt" >>> thing. That should be revealing. >>> >>> murf >>> >>> >>> >>>> Thanks, >>>>