Hi,
I send 50 messages with a 2000B/s shaper, goes through OK.
Send 200 Messages with the same shaper, 78 gets lost.
Get two of these in the exception log:
"Failure: <type 'int'>: 201"
Using Twisted 1.3; Will try 2.2 still.
Tha'ts all. And the sender gets a "undeliverable message".
Maybe twisted 1.3 can't handle it?
Cheers,
David
On 7/20/06, James Bunton <[EMAIL PROTECTED]> wrote:
> Yeah, it seems to have fixed the problem (as far as I can tell).
> I used to be able to type 10 (1 char) messages quickly enough that
> only about 5 of them would get through. Now if I send them all at the
> same time (paste into the XML console), they still all get through,
> every time. I even tried it with 50.
>
>
> Maybe it'd be a good idea for somebody to talk to the ejabberd people
> about this and get their perspective.
>
> ---
>
> James
>
>
>
> On 20/07/2006, at 6:56 AM, David E Freitas wrote:
>
> >> It does seem to be a little better now.
> >
> > Better after increasing the shaper's limit?
> >
> > This lost messages is a real problem.
> >
> > Most of our users lose messages, mostly messages coming IN to them on
> > the gateway.
> >
> > Specially when above 1500 concurrent.
> > I don't trust typing notifications for some reason.
> >
> > Twisted 1.3
> > PyMSNt 0.11.1
> > Python 2.4
> > ejabberd 0.9.8
> >
> >
> > Sincerely,
> >
> >
> > David
> >
> >
> > P.S. Is Twisted 2.4 not supported by PyMSNt yet?
> >
> >
> > On 7/12/06, James Bunton <[EMAIL PROTECTED]> wrote:
> >> On 12/07/2006, at 3:18 PM, Trejkaz wrote:
> >>
> >> >
> >> >> I'm kind of at a loss as to what's going on here. I would really
> >> >> appreciate anybody's insight.
> >> >
> >> > FWIW, I've tried bumping up the shaper's limit, just to see if that
> >> > really
> >> > was the issue. I won't really know until the load hits the point
> >> > again.
> >> >
> >> > TX
> >>
> >> It does seem to be a little better now.
> >>
> >> Btw, I found the source of the duplicate debug log entries. Silly
> >> error in debug.py that is only triggered if you send the transport a
> >> SIGHUP.
> >>
> >> Its fixed in svn trunk.
> >>
> >> ---
> >>
> >> James
> >>
> >> _______________________________________________
From [EMAIL PROTECTED] Thu Jul 20 08:45:06 2006
From: [EMAIL PROTECTED] (Yves Goergen)
Date: Thu Jul 20 08:45:01 2006
Subject: [py-transports] Cannot start PyMSNt and PyICQt
In-Reply-To: <[EMAIL PROTECTED]>
References: <[EMAIL PROTECTED]> <[EMAIL PROTECTED]>
<[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
On 20.07.2006 00:59 CE(S)T, Lars T. Mikkelsen wrote:
> * Manually connect to the Jabber server (e.g. echo "<stream:stream>" |
> nc -vvv localhost 5347) and see if you get a response.
Works. The reply is:
</stream:stream>
then the connection is closed.
> * Check if the loopback devices is properly configured (ifconfig lo).
Here's what it said. No idea what it means.
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:6760 errors:0 dropped:0 overruns:0 frame:0
TX packets:6760 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:1647571 (1.5 MiB) TX bytes:1647571 (1.5 MiB)
> * Check if the issue is still present with the latest kernel prepatch
> (currently 2.6.18-rc2).
Okay, I retried the previous kernel versions I still had on my machine
now and none of them works. The Py transports simply died after the
recent reboot, it doesn't seem to depend on the kernel.
--
Yves Goergen "LonelyPixel" <[EMAIL PROTECTED]>
http://beta.unclassified.de ? My web laboratory.
From [EMAIL PROTECTED] Thu Jul 20 09:02:53 2006
From: [EMAIL PROTECTED] (Yves Goergen)
Date: Thu Jul 20 09:02:49 2006
Subject: [py-transports] Cannot start PyMSNt and PyICQt
In-Reply-To: <[EMAIL PROTECTED]>
References: <[EMAIL PROTECTED]> <[EMAIL PROTECTED]>
<[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
So there's some generic bug anywhere that prevents the Py transports and
JRU-PHP from running. I have tried to update the Twisted framework to
2.4.0 but that doesn't change anything. Since it's OpenSSL causing the
other log message, I wanted to reinstall the whole SSL crap but that
doesn't seem to be possible without uninstalling half the system with
it. Any chance I can repair it and use the transports again?
There's a problem with that computer I have since it's there: I cannot
compile ProFTPd on the same machine because when I start it, it
immediately segfaults with no action taken until then. Also when I
compile Exim on the same machine, I cannot use TLS SMTP connections,
they simply fail. Compiling both apps on another machine with exactly
the same software and even the same hardware architecture (Athlon64) but
only inside VMware, then copy the binaries to the web server, it works
fine. When I got that machine, it messed up everything because of broken
RAM. I had them replace that RAM and since then it seemed to work. But
it's still not 100% though... :( Today's computers are far too complex
to manage them reliably... Or is it the evil neutrinos?
--
Yves Goergen "LonelyPixel" <[EMAIL PROTECTED]>
http://beta.unclassified.de ? My web laboratory.
From [EMAIL PROTECTED] Thu Jul 20 10:05:09 2006
From: [EMAIL PROTECTED] (James Bunton)
Date: Thu Jul 20 10:05:23 2006
Subject: [py-transports] Cannot start PyMSNt and PyICQt
In-Reply-To: <[EMAIL PROTECTED]>
References: <[EMAIL PROTECTED]> <[EMAIL PROTECTED]>
<[EMAIL PROTECTED]>
<[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
On 20/07/2006, at 7:02 PM, Yves Goergen wrote:
> So there's some generic bug anywhere that prevents the Py
> transports and
> JRU-PHP from running. I have tried to update the Twisted framework to
> 2.4.0 but that doesn't change anything. Since it's OpenSSL causing the
> other log message, I wanted to reinstall the whole SSL crap but that
> doesn't seem to be possible without uninstalling half the system with
> it. Any chance I can repair it and use the transports again?
>
> There's a problem with that computer I have since it's there: I cannot
> compile ProFTPd on the same machine because when I start it, it
> immediately segfaults with no action taken until then. Also when I
> compile Exim on the same machine, I cannot use TLS SMTP connections,
> they simply fail. Compiling both apps on another machine with exactly
> the same software and even the same hardware architecture
> (Athlon64) but
> only inside VMware, then copy the binaries to the web server, it works
> fine. When I got that machine, it messed up everything because of
> broken
> RAM. I had them replace that RAM and since then it seemed to work. But
> it's still not 100% though... :( Today's computers are far too complex
> to manage them reliably... Or is it the evil neutrinos?
>
> --
> Yves Goergen "LonelyPixel" <[EMAIL PROTECTED]>
> http://beta.unclassified.de ? My web laboratory.
Gotta hate them neutrinos :)
Good luck figuring that out :\
---
James
