Summary: | ASTERISK-00553: Public IP * forgets to check on registered dynamic IP *: Unreachable! | ||
Reporter: | philipp2 (philipp2) | Labels: | |
Date Opened: | 2003-11-21 07:13:57.000-0600 | Date Closed: | 2011-06-07 14:04:58 |
Priority: | Major | Regression? | No |
Status: | Closed/Complete | Components: | Core/General |
Versions: | Frequency of Occurrence | ||
Related Issues: | |||
Environment: | Attachments: | ||
Description: | SETUP server A: static IP (public) server B: dynamic IP (cable modem) none of the machines are behind NAT server B registers with server A as peer PROBLEM The cable modem link of server B regularly drops below 2000 ms but comes back up quickly again - fine so far. However, with "iax2 show peers" on server A I permanently see a status of UNKOWN for server B. A "reload" on server A fixes the problem and I see a status of "OK (76) ms", and now I can place calls again. As a workaround I could schedule a "reload" on server A once per hour, but I don't think this is the way it should be. | ||
Comments: | By: philipp2 (philipp2) 2003-11-21 07:15:34.000-0600 server A: Asterisk CVS-11/10/03-17:38:20 server B: Asterisk CVS-11/10/03-17:38:11 By: philipp2 (philipp2) 2003-11-21 07:21:02.000-0600 iax.conf on server A: [serverBp] context=default type=peer host=dynamic defaultip=111.222.333.444 secret=usurewouldlike2know auth=md5 ;qualify=yes qualify=2500 notransfer=yes disallow=all allow=gsm allow=ilbc By: zoa (zoa) 2003-11-21 08:44:28.000-0600 i think i noticed something similar with qualify = yes. Maybe there should be an option to have monitor=yes as an extra option to monitor connections. I think that if you have a timeout once, it gets unreachable and stays that way for a too long time. or maybe a qualify=yes,retry=60000ms or so to try again after 1 minute By: Brian West (bkw918) 2003-11-21 10:31:13.000-0600 qualify=yes won't recheck peers without a restart first. Dont know why.. but restart * and see what it does. By: zoa (zoa) 2003-11-21 10:41:30.000-0600 it should :) i also have sudden packetloss burst for some seconds a day on some links :) By: philipp2 (philipp2) 2003-11-21 10:50:58.000-0600 not sure what you mean here bkw: of course a restart will do what a reload does, so in that sense it'll work - but only until the next drop-out occurs! I have been sitting on this issue for a couple of weeks now, and I am pretty sure that restart doesn't make * suddenly check up regularly on a "lost child" to see if it returned home. :-> By: philipp2 (philipp2) 2003-12-16 10:06:27.000-0600 Looks like this is the same issue? http://bugs.digium.com/bug_view_page.php?bug_id=0000673 By: Brian West (bkw918) 2004-01-07 00:10:54.000-0600 no in some cases I see restart do stuff that a reload won't can you see if one or the other is true in this case? By: philipp2 (philipp2) 2004-01-07 04:06:46.000-0600 Neither restart nor reload helps, at least not on the long run. I now turned to register the servers using SIP as fallback if IAX fails, and now most if not all connections use SIP (which means that "SIP is better than IAX" - hehe!). First I tried the regular "reload", but you read aobut the "reload" bug as well... ;-) By: Brian West (bkw918) 2004-01-09 21:07:57.000-0600 NO this is a config issue host=dynamic defaultip=111.222.333.444 Don't use those together.. use one or the other. By: Brian West (bkw918) 2004-01-09 21:15:31.000-0600 ASTERISK-667 fixed this or so it should have. |