Summary: | ASTERISK-06324: Initial burst of calls is causing asterisk to have problems with the management connections | ||
Reporter: | Andrey S Pankov (casper) | Labels: | |
Date Opened: | 2006-02-15 08:56:57.000-0600 | Date Closed: | 2011-06-07 14:09:59 |
Priority: | Minor | Regression? | No |
Status: | Closed/Complete | Components: | Core/ManagerInterface |
Versions: | Frequency of Occurrence | ||
Related Issues: | |||
Environment: | Attachments: | ||
Description: | If I start 150 simultaneous calls at once, the telnet connection appears dead. Normally, I have the call manager starting the calls at a maximum rate of 10/sec. When we start, it takes 15 sec for the 150 calls to originate, but then the call manager doesn't get anything back from asterisk over the management connection. I tried turning the rate down to as low as 1 call/sec, but it still gets stuck. What is interesting is that if I start at 25 simultaneous calls, even at 10 calls/sec, the management connection is fine. Then I increase the simultaneous calls 25 at a time, up to 150 simultaneous calls it is fine. So it appears that the initial burst of calls, even at a slow rate is causing asterisk to have problems with the management connections. | ||
Comments: | By: Jason Parker (jparker) 2006-02-15 16:06:45.000-0600 Summary was funky...fixed By: Andrey S Pankov (casper) 2006-02-16 08:09:52.000-0600 We use SIP channel to originate calls. There were no problems with mid-summer CVS HEAD. By: Andrey S Pankov (casper) 2006-03-01 19:23:46.000-0600 Please close this bug... unless there are volunteers to re-write manager.c not being able to deal differently with external (not in "action callbacks") manager_event calls on "slow connections" (read: when there is problems writing to a session socket). By: Matt O'Gorman (mogorman) 2006-03-02 17:10:03.000-0600 have you tried the astmanproxy as a front end for your manager connection? By: Andrey S Pankov (casper) 2006-03-03 14:05:31.000-0600 No... Thanks for a suggestion... do you mean astmanproxy from http://svncommunity.digium.com? Or perl script included somewhere in the tar? Anyway this is not the way to resolve the bugs in the core AMI code... ;) By: Matt O'Gorman (mogorman) 2006-03-07 16:32:51.000-0600 i would think frontending ami locally with asteriskmangmentproxy would solve your problem as it deals much better with connection issues. By: Andrey S Pankov (casper) 2006-04-03 07:43:32 I hope recent changes in trunk will resolve the issue. Let's close this one for now! By: Andrey S Pankov (casper) 2006-04-12 09:46:02 PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE Close this one! Thanks! PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE PLEASE |