Summary:	ASTERISK-05028: "Presence" subscription causes Internal Server Error on Polycom 600
Reporter:	wjchan (wjchan)	Labels:
Date Opened:	2005-09-08 14:27:57	Date Closed:	2005-09-14 18:10:22
Priority:	Minor	Regression?	No
Status:	Closed/Complete	Components:	Channels/chan_sip/Subscriptions
Versions:		Frequency of Occurrence
Related Issues:
Environment:		Attachments:	( 0) debug-9-7-p1.txt ( 1) debug-9-7-p2.txt ( 2) polycom1.txt ( 3) polycom-500.txt ( 4) polycom-sub-debug.txt
Description:	I'm subscribed to a "buddy" on my Polycom phone.  Periodically I'd get this error: 'Incoming call: Got SIP response 500 "Internal Server Error" back from 10.1.1.200'.  SIP debug attached.
Comments:	By: Olle Johansson (oej) 2005-09-08 14:43:37 Guess that polycom has no bug tracker ;-) What firmware is this? By: Olle Johansson (oej) 2005-09-08 14:45:29 1.5.2 I see from the logs. That is weird. Anyone else that knows why this happens on this polycom? By: wjchan (wjchan) 2005-09-08 14:55:32 firmware is 1.4.1.0040 By: wjchan (wjchan) 2005-09-08 14:56:30 i mean 1.5.2.  sorry.  you got it already By: wjchan (wjchan) 2005-09-08 23:12:03 I unsubscribed and resubscribed a few hours ago and haven't seen the 500 error since then.  I'll let it stew overnight and report back. By: Olle Johansson (oej) 2005-09-09 10:59:36 I guess we need to check the tag-handling. We are changing to/from headers and keep the tags from the subscribe... Back to the RFCs. By: Olle Johansson (oej) 2005-09-09 11:13:53 Checking RFC3856, that is correct behaviour. Must be something else. By: Olle Johansson (oej) 2005-09-09 11:18:16 Ahh. This is a re-subscribe. You need to show me the full transaction, from the FIRST subscribe to a re-subscribe and the error. I have an idea. By: Olle Johansson (oej) 2005-09-09 11:20:00 And when you capture, turn on debug logging to 4, verbose to 4 and turn on SIP debugging for this connection only. THanks! By: wjchan (wjchan) 2005-09-09 11:26:21 I haven't received a 500 error since my re-sub last night.  When I was getting the errors, that subscription was started many chan_sip.c versions before.  Not sure if the subsequent Asterisk upgrades affected the subscription. By: wjchan (wjchan) 2005-09-09 11:40:16 When I say resubscribe, I mean I go to my phone's directory, select a contact, disable the "Watch Buddy", wait a few seconds, then enable "Watch Buddy" again.  I initially subscribed when the presence/notify stuff went into CVS-HEAD a few days ago.  Since that I've been updating Asterisk daily.  The 500 Error popped up a day or 2 ago.  So last night I "resub'ed" from my phone and the error has gone away. Have said that, what SIP trace would you like me to do? By: wjchan (wjchan) 2005-09-09 12:32:31 I'll try to get you something.  I think the phone refreshes presence info every hour so it'll take a few hours to get something interesting. By: wjchan (wjchan) 2005-09-09 12:51:43 polycom1.txt The log started with no subscription and me going to the phone and enabled a buddy watch on a contact.  After that the phone chats with Asterisk periodically. By: wjchan (wjchan) 2005-09-10 11:54:27 polycom-sub-debug.txt is the debug log mirroring polycom1.txt By: Olle Johansson (oej) 2005-09-12 10:48:49 I need a larger debug example with the 500 Internal Server Error happening. I don't see that in the latest file you did send me. By: wjchan (wjchan) 2005-09-12 10:53:33 The 500 error hasn't happened since I disabled then enabled buddy watch on the phone on 9/9/05. By: Olle Johansson (oej) 2005-09-12 10:56:59 Talked with Polycom engineers about this. The only reason for a 500 they can think of is that we are sending packets with a too low CSEQ. I need more packets from you to find out if we do, since we can not repeat this. Start with logging from the first SUBSCRIBE and keep logging until the phone re-subscribes and see what happens immediately after the update. Thank you. By: wjchan (wjchan) 2005-09-12 11:16:29 These 2 new logs probably don't have enough detail for you.  I went back to the old logs and found where the 500 error first occurred (see debug-9-7-p1.txt) and when it started repeatedly happening (see debug-9-7-p2.txt). By: Olle Johansson (oej) 2005-09-13 10:19:10 These logs does not help me, I need a full SIP debug with all the packets. By: Kevin P. Fleming (kpfleming) 2005-09-14 18:10:08 Since the poster has not been able to reproduce the problem after upgrading chan_sip, it's likely that the source of this problem has already been corrected. I'm closing this bug for now; if theproblem becomes reproducable, please request a re-open.