JIRA goes.
Any ideas on how we might tickle this?
Andy
- Original Message -
From: "Joegen Baclor"
To: andrewpit...@comcast.net
Cc: "Discussion list for users of sipXecs software"
Sent: Tuesday, September 25, 2012 1:44:22 AM
Subject: Re: [sipx-users] new pat
Thanks Gerald,
Good to hear you are getting better mileage with patch #19. I sent a
patch to Andrew that is not included yet with patch #19. It has
something to do with TCP packets not getting reported to NAT traversal
plugin.
Joegen
On 09/25/2012 08:44 PM, Gerald Drouillard wrote:
On 9/2
On 9/25/2012 1:44 AM, Joegen Baclor wrote:
Andrew,
any update on this?
Update #19 fixed many issues we were having with offsite registration.
With #18 we had turned off 5061 and that seemed to fix most clients. We
have not turned on 5061 with #19 yet. I would like a day or two of
stability
comcast.net
*Cc: *"Discussion list for users of sipXecs software"
*Sent: *Wednesday, September 19, 2012 7:42:48 AM
*Subject: *Re: [sipx-users] new patch for XX-10177
Andrewe any news on the patch i sent you offlist?
On 09/13/2012 03:19 AM, andrewpit...@comcast.net wrote:
Joegen,
uot;
Cc: andrewpit...@comcast.net
Sent: Tuesday, August 21, 2012 10:16:20 PM
Subject: Re: [sipx-users] new patch for XX-10177
I've taken a quick look at the NAT traversal plugin and it seems to be not
accounting PUBLISH as a potential stream modifier. But it is not clear to me
how
, August 22, 2012 5:32:34 PM
*Subject: *Re: [sipx-users] new patch for XX-10177
Ooops, typo. Its not PUBLISH but UPDATE.
On 08/23/2012 01:58 AM, andrewpit...@comcast.net wrote:
Joegen,
Actually, it looks like we have the keepalive.sessionTimers option
turned off for all our phones
From: "Joegen Baclor"
To: andrewpit...@comcast.net
Cc: "Discussion list for users of sipXecs software"
Sent: Wednesday, August 22, 2012 5:32:34 PM
Subject: Re: [sipx-users] new patch for XX-10177
Ooops, typo. Its not PUBLISH but UPDATE.
On 08/23/2012 01:58 AM, andre
et
Sent: Tuesday, August 21, 2012 10:16:20 PM
Subject: Re: [sipx-users] new patch for XX-10177
I've taken a quick look at the NAT traversal plugin and it seems to be not
accounting PUBLISH as a potential stream modifier. But it is not clear to me
how it would cause a hang. Can you d
..@comcast.net
*Sent: *Tuesday, August 21, 2012 10:16:20 PM
*Subject: *Re: [sipx-users] new patch for XX-10177
I've taken a quick look at the NAT traversal plugin and it seems to be
not accounting PUBLISH as a potential stream modifier. But it is not
clear to me how it would cause a hang. Can
e.
-Andy
- Original Message -
From: "Joegen Baclor"
To: "Discussion list for users of sipXecs software"
Cc: andrewpit...@comcast.net
Sent: Tuesday, August 21, 2012 10:16:20 PM
Subject: Re: [sipx-users] new patch for XX-10177
I've taken a quick look at the NAT
...@list.sipfoundry.org
[mailto:sipx-users-boun...@list.sipfoundry.org] On Behalf Of Joegen Baclor
Sent: Tuesday, August 21, 2012 7:16 PM
To: Discussion list for users of sipXecs software
Subject: Re: [sipx-users] new patch for XX-10177
I've taken a quick look at the NAT traversal plugin a
On Tue, Aug 21, 2012 at 12:59 PM, andrewpitman wrote:
>
>
> Joegen, George,
>
> I noticed some messages in my sipXproxy log from a hang on a
> customer system just today, which may or may not be
> pertinent. Besides the parsing errors, I also see messages
> such as this immediately before the han
om: *"Tony Graziano"
*To: *"Discussion list for users of sipXecs software"
*Sent: *Tuesday, August 21, 2012 1:01:03 PM
*Subject: *Re: [sipx-users] new patch for XX-10177
can you elaborate what the UA's are that are involved in that transaction?
On Tue, Aug 21, 2012 at 5:59
, August 21, 2012 1:01:03 PM
Subject: Re: [sipx-users] new patch for XX-10177
can you elaborate what the UA's are that are involved in that transaction?
On Tue, Aug 21, 2012 at 5:59 AM, andrewpitman < andrewpit...@comcast.net >
wrote:
Joegen, George,
I noticed some messages
can you elaborate what the UA's are that are involved in that transaction?
On Tue, Aug 21, 2012 at 5:59 AM, andrewpitman wrote:
>
>
> Joegen, George,
>
> I noticed some messages in my sipXproxy log from a hang on a
> customer system just today, which may or may not be
> pertinent. Besides the pa
Joegen, George,
I noticed some messages in my sipXproxy log from a hang on a
customer system just today, which may or may not be
pertinent. Besides the parsing errors, I also see messages
such as this immediately before the hang:
"2012-08-21T15:40:56.031862Z"
:269589:NAT:WARNING:pbx1.sipdomain
I'll discuss this with GEroge later. Thanks for pointing it out.
On 08/20/2012 09:33 PM, andrewpitman wrote:
>
> Joegen, it looks like the keepalive processing in
> SipClient.cpp in the release-4.4 sipXtackLib is still very
> specific in what it matches. I thought it was supposed to
> consider a
Joegen, it looks like the keepalive processing in
SipClient.cpp in the release-4.4 sipXtackLib is still very
specific in what it matches. I thought it was supposed to
consider any combination of CR's and LF's (as long as the
buffer contains only these) as a keepalive?
Joegen Baclor wrote on Thu
noticed.
I've updated the JIRA to that effect.
Thanks!
Andy
- Original Message -
From: "Joegen Baclor"
To: "Discussion list for users of sipXecs software"
Cc: andrewpit...@comcast.net
Sent: Thursday, August 2, 2012 10:28:09 PM
Subject: Re: [sipx-users] new p
ned 16 in task 2966346640"
Hope this helps,
Andy
*From: *"Michael Picher"
*To: *"Discussion list for users of sipXecs software"
*Sent: *Thursday, August 2, 2012 12:42:12 PM
*Subject: *Re: [sipx-use
. The log entry says : "OsBSemLinux::~OsBSemLinux
pt_sem_destroy returned 16 in task 2966346640"
Hope this helps,
Andy
- Original Message -
From: "Michael Picher"
To: "Discussion list for users of sipXecs software"
Sent: Thursday, August 2, 2012 12:42:12 PM
Sub
lls on the remote end vary.
- Original Message -
From: "Michael Picher"
To: "Discussion list for users of sipXecs software"
Sent: Thursday, August 2, 2012 12:42:12 PM
Subject: Re: [sipx-users] new patch for XX-10177
do you lick down a lot of ports? :-P
i thin
Yeah, we've ruled that out. If it is a UA sending malformed
SIP traffic, sipXproxy still should drop that or respond
with an error (and maybe write a log entry to that effect)
but it should not cause it to lose its socket connection to
sipregistrar and have to be restarted.
Tony Graziano wrote
do you lick down a lot of ports? :-P
i think these are all Polycom phones with 3.2.6 firmware.
the only other thing a bit odd is they are coming through a Cisco ASA which
is known to work but could be a question mark. I think they were going to
try to route around this and then in through a pfS
The malformed crap could easily come from a misconfigured or badly designed
UA by the way. Also realize I have never seen it even with remote user
traversal WHEN I lick down pps to port 5060 in the firewall to a sane
functional number. One assumes you inspected the logs to verify there was
no outsi
Hi Joegen!
I dug around a bit in the code, and I might have a starting
point for where to look for this...
When this bug has manifested itself, we've been able to
recover by restarting just sipXproxy, and not both proxy and
registrar, so the issue doesn't seem to be with registrar.
When the se
No, but it does keep sipXproxy from logging when it receives
these keepalives (or similar), heading off a potential
avenue for a DOS attack. If you get enough of them, it
could fill the filesystem sipxecs logs to.
Andy
Domenico Chierico wrote on Mon, 25 June 2012 10:41
> so this doesn't fix th
users of sipXecs software" <
> sipx-users@list.sipfoundry.org>
> *Sent: *Monday, June 25, 2012 6:56:19 AM
>
> *Subject: *Re: [sipx-users] new patch for XX-10177
>
> or maybe
>
> if (readBuffer(i) == '\r' || readBuffer(i) == '\n')
>
> On Mon, Jun 25, 20
Yeah, I had corrected that in the patch I posted to the JIRA.
- Original Message -
From: "Domenico Chierico"
To: "Discussion list for users of sipXecs software"
Sent: Monday, June 25, 2012 6:56:19 AM
Subject: Re: [sipx-users] new patch for XX-10177
or mayb
On Mon, Jun 25, 2012 at 1:56 PM, Domenico Chierico
wrote:
> or maybe
>
> if (readBuffer(i) == '\r' || readBuffer(i) == '\n')
>
> On Mon, Jun 25, 2012 at 11:30 AM, Domenico Chierico
> wrote:
>>
>> see something strange in the code, can this be a leftover:
>>
>> if (readBuffer(i) == '\r' && readBu
or maybe
if (readBuffer(i) == '\r' || readBuffer(i) == '\n')
On Mon, Jun 25, 2012 at 11:30 AM, Domenico Chierico <
domenico.chier...@sip2ser.it> wrote:
> see something strange in the code, can this be a leftover:
>
> if (readBuffer(i) == '\r' && readBuffer(i) == '\n')
>
> shouldn't be something
see something strange in the code, can this be a leftover:
if (readBuffer(i) == '\r' && readBuffer(i) == '\n')
shouldn't be something like:
if (readBuffer(i) == '\r' && readBuffer(i+1) == '\n')
On Fri, Jun 22, 2012 at 8:43 PM, Hay, Nathan wrote:
> I believe we are experiencing this problem.
I believe we are experiencing this problem.
When I turn up the logging, I get these lines over and over again in the
sipXproxy log:
"2012-06-22T17:33:13.804499Z":26:SIP:WARNING:sip1.cedarville.edu:SipClientTcp-18:4159E940:SipXProxy:"SipMessage::getResponseSendAddress
No VIA address, using
>From a
Nah, I can just pull the source down with git, patch,
compile and build my own rpms. Thanks!
I'll also post the patch to the JIRA.
--
___
sipx-users mailing list
sipx-users@list.sipfoundry.org
List Archive: http://list.sipfoundry.org/archive/sipx-us
On 06/21/2012 10:34 PM, andrewpitman wrote:
Okay, I ran two versions of the Perl script against my test
server for a few hours last night, both of which loop and
send the crlf keepalive messages as fast as they could be
sent. One was sending UDP and the other TCP. I didn't
manage to get proxy
Joegen,
Actually, the patch seems to be very specific in what it
matches. For example, if I reformat the message I'm sending
so that it consists of line feeds, then carriage returns
instead of CRLFs, I can still generate tons of logging. As
a fix for a potential DOS attack, I don't think it's
Joegen, I don't mind at all. I've got a couple of servers with the patch
already applied. I'll point my new "DOS generator" at those and watch for a
few days. ;)
--
Sent from my iPhone appendage
On Jun 21, 2012, at 12:50, Joegen Baclor wrote:
> On 06/21/2012 10:34 PM, andrewpitman wrote:
>>
On 06/21/2012 10:34 PM, andrewpitman wrote:
>
> Okay, I ran two versions of the Perl script against my test
> server for a few hours last night, both of which loop and
> send the crlf keepalive messages as fast as they could be
> sent. One was sending UDP and the other TCP. I didn't
> manage to g
Okay, I ran two versions of the Perl script against my test
server for a few hours last night, both of which loop and
send the crlf keepalive messages as fast as they could be
sent. One was sending UDP and the other TCP. I didn't
manage to get proxy to stop responding to registration
requests,
Time to break out the call generator again...
--
Sent from my iPhone appendage
On Jun 19, 2012, at 21:29, Joegen Baclor wrote:
> Stressing sipx to its limits is a good litmus test so if you could, do it.
> Take a statistical snapshot of CPU and mem and see if this causes a gradual
> rise.
Sure, I'll give that a try.
--
Sent from my iPhone appendage
On Jun 19, 2012, at 22:00, Joegen Baclor wrote:
> Andrew,
>
> Another possibility is TCP. We are sending it UDP currently. Can you also
> try sending through TCP?
>
> On 06/20/2012 09:17 AM, Andrew Pitman wrote:
>> Hi Joegen,
>>
Andrew,
Another possibility is TCP. We are sending it UDP currently. Can you
also try sending through TCP?
On 06/20/2012 09:17 AM, Andrew Pitman wrote:
> Hi Joegen,
>
> Actually, after a couple of days running it with the delay, I took the 2
> second sleep out of the Perl script and ran it co
Stressing sipx to its limits is a good litmus test so if you could, do
it. Take a statistical snapshot of CPU and mem and see if this causes
a gradual rise. For whatever it is worth, the patch may save us a few
mb worth of logs in production or it could be a real issue altogether.
If you a
Hi Joegen,
Actually, after a couple of days running it with the delay, I took the 2 second
sleep out of the Perl script and ran it continuously from a while loop in the
shell. Still no dice.
I could make it tighter still by looping in the Perl script with no sleeps. ;)
Andy
--
Sent from my iP
Yes that is it. From the logs that was previously posted, this was the
most evident thing that was going crazy before the server was hung. So
two things,
1. We are totally mistaken that this was the cause.
2. 1 packet every 5 seconds is not enough to cause havoc in the system.
I am hoping
Joegen, like this?
Quote:
> "2012-06-19T21:49:58.561697Z"
> :14693872:SIP:ERR:pbx1.ncpbxtechs.com:SipClientUdp-12:B75B9B
> 90:SipXProxy: "Url::parseString no valid host found at
> char 0 in ';tag=R5DYzi', uriForm = name-addr"
> "2012-06-19T21:49:58.561790Z"
> :14693873:SIP:WARNING:pbx1.ncpbxtech
Hi Andrew,
Did you check the logs if we are able to trigger the message corruption
in your unpatched server?
On 06/18/2012 11:20 PM, andrewpitman wrote:
>
> George, Joegen,
>
> I've been running this Perl script against an unpatched test
> server for over 3 days now, and I still haven't been ab
George, Joegen,
I've been running this Perl script against an unpatched test
server for over 3 days now, and I still haven't been able to
tickle the bug. Any ideas?
Andy
--
___
sipx-users mailing list
sipx-users@list.sipfoundry.org
List Archive: ht
Joegen, I tried downloading that Perl script and all I get
is an empty file.
--
___
sipx-users mailing list
sipx-users@list.sipfoundry.org
List Archive: http://list.sipfoundry.org/archive/sipx-users/
Thanks, George!
- Original Message -
From: "George Niculae"
To: "Discussion list for users of sipXecs software"
Sent: Thursday, June 14, 2012 11:38:56 AM
Subject: Re: [sipx-users] new patch for XX-10177
On Thu, Jun 14, 2012 at 6:43 PM, andrewpitman wrote
On Thu, Jun 14, 2012 at 6:43 PM, andrewpitman wrote:
>
>
> Joegen, I tried downloading that Perl script and all I get
> is an empty file.
> --
Hm, I see the content, check this:
#!/usr/bin/perl
#udpclient.pl
use IO::Socket::INET;
# flush after every write
$| = 1;
my ($socket,$data);
# We ca
Change PeerAddr => '127.0.0.1:5060' to the address of sipx and cron
this perl script every 5 seconds.
On 06/13/2012 06:34 AM, andrewpitman wrote:
We've been running the patched version on our test servers
since yesterday, and so far so good. Once everything checks
out, we can apply this to
We've been running the patched version on our test servers
since yesterday, and so far so good. Once everything checks
out, we can apply this to customers who have experienced the
registration drops and make sure this resolves their issue.
Not sure if there's a way to fully test the fix solely
if (res == crlfCount)
I thought it looked as if there was some sort of conditional
that was missing. ;)
Fortunately, I've been so swamped with other things I
haven't had the chance to really try it out. I will apply
today and start testing.
Thanks!
--
___
Hey Andy,
I somehow uploaded a patch that is not complete to XX-10177, could you
please get the latest one attached to JIRA and build
(http://track.sipfoundry.org/secure/attachment/27824/0001-XX-10177-Server-stops-responding-to-registration-req.patch)?
(thanks Levend for pointing this out)
Thanks
55 matches
Mail list logo
