From: <mar...@be...> - 2008-07-16 12:27:16
|
Hi all, I'm using the trunk version extracted Jul 11 2008 on a linux 2.4.26 (slackware). Before I used version 3.0, upgrading to version 3.1 I gave up after some days because it wouldn't accept any of the scripts so I'm converting to the "latest" trunk version. Now I noticed that in the trunk version: 1. The -stdin option is needed when using -bg When starting in bg (-bg), sipp just crashes/stops without any message unless you also specify -stdin as option. This used to work without it, but OK, now I know;-) Just don't know whether it was meant like this, there is little documentation on it and a small error message might be helpful. 2. SIPp consuming all CPU When starting in bg (-bg -stdin) as client (didn't convert the server code yet), sipp starts rather quickly to use 100% of the CPU. A screen log shows "1 Running, 3 Paused, 189 Woken up" or even much larger numbers for the woken up, while there is only one session "1 calls (limit 1)" "Pause"-ing. The script continues to be executed correctly however, and the sipp process can be killed. Running in foreground, I get comparable behaviour (CPU 100%) after the screen switches using key 1, 2, 3, 4 -> empty screen + CPU eating + no reaction on USR2 or any key, even a normal kill can't kill the process any more (repeatable behaviour). So, probably, this is not exactly the same problem. I saw that comparable problems from the past should have been solved, but is this new or just slipped in again ?? 3. SIPp stopping on first authentication problem in message from SUT When the SUT fails to send an authentication "Proxy-Authenticate: " in the 407 message, SIPp just stops. Some months ago I sent a patch to have sipp log the error and continue testing (like for e.g. any reception of an unexpected message), is it gone or rejected or ?? 4. [local_port] value 0 when reusing UDP port When reusing a UDP port (max_sockets < # of sessions/calls), the value for [local_port] in substitutions is 0 instead of the correct value. This bug was also available in version 3.0, but working fine in version 2.0. I silently hoped it would be gone again now, but ... I tried to follow the code long ago (version 3.0), and the biggest problem is that when local_port should be set on reuse of a port, the value of the port number is nowhere in the available variables/structures. Does somebody have an idea in what structure (related to the port that will be reused) it should be added). I think this one I could handle myself, but my view on the data structures and global program flow is rather limited. Anybody knowing more on the state of these topics? Best regards (-: from Marc VAN DIEST (BELGACOM) ;-) |