Problems running Tux w/ 2.4.19 generic

Hello,

I am currently trying to run Red Hat 8.0 with Tux compiled into a 2.4.19 generic kernel.  I have applied the
tux2-full-2.4.18-final-A3 patch.  Under what I consider is very minor work loads Tux will, for
apparently no reason, drop all of its threads and stop responding to requests.    I realize that the patch was
not intended for this kernel version, but it is necessary in order for my SMP machine to work properly.

My question is this.
   Has anyone attempted to do this, and if so did you see the same problems?  If you didn't see the same problems,
am I missing something?  If you did see the same problems, did you ever find a work around for this issue?

Thanks. Joshua  

Re: Problems running Tux w/ 2.4.19 generic

On Wed, Jan 08, 2003 at 05:57:48PM -0600, joshua.schnee <at> amd.com wrote:
> Hello,
> 
> I am currently trying to run Red Hat 8.0 with Tux compiled into a 2.4.19 generic kernel.  I have applied the
tux2-full-2.4.18-final-A3 patch.  Under what I consider is very minor work loads Tux will, for
apparently no reason, drop all of its threads and stop responding to requests.    I realize that the patch was
not intended for this kernel version, but it is necessary in order for my SMP machine to work properly.
> 
> My question is this.
>    Has anyone attempted to do this, and if so did you see the same problems?  If you didn't see the same problems,
am I missing something?  If you did see the same problems, did you ever find a work around for this issue?
> 
> Thanks. Joshua  

Did you use a patch that had been ported to 2.4.19?

When you patched, were there any error messages saying "Hunk #N Failed"
where N is an integer?

I've patched 2.4.21pre2 with tux, but I had to modify the patch in order
to make the patching successful.

Also, there have been numerous threads on-list regarding tux hanging
with logging enabled. Check the archives for that info, and try
disabling logging if you have it enabled.

--

-- 
Shane Allen <shane <at> sell.com>

RE: Problems running Tux w/ 2.4.19 generic

Thank you for responding,

I am unaware of any patch that has been ported to 2.4.19.  Is there a ported patch available somewhere?

Yes I had several hunks fail, but as this is not my area of expertise, I am unable to create a work around.

Only Logging, and redirect_logging are enabled, from reading past archive messages I had thought that
referer_logging was causing the errors.  Should I disable all logging?

Thanks

-----Original Message-----
From: Shane Allen [mailto:shane <at> sell.com]
Sent: Wednesday, January 08, 2003 9:10 PM
To: tux-list <at> redhat.com
Subject: Re: Problems running Tux w/ 2.4.19 generic

On Wed, Jan 08, 2003 at 05:57:48PM -0600, joshua.schnee <at> amd.com wrote:
> Hello,
> 
> I am currently trying to run Red Hat 8.0 with Tux compiled into a 2.4.19 generic kernel.  I have applied the
tux2-full-2.4.18-final-A3 patch.  Under what I consider is very minor work loads Tux will, for
apparently no reason, drop all of its threads and stop responding to requests.    I realize that the patch was
not intended for this kernel version, but it is necessary in order for my SMP machine to work properly.
> 
> My question is this.
>    Has anyone attempted to do this, and if so did you see the same problems?  If you didn't see the same problems,
am I missing something?  If you did see the same problems, did you ever find a work around for this issue?
> 
> Thanks. Joshua  

Re: Problems running Tux w/ 2.4.19 generic

SuSE Linux 8.1 ships tux with it's 2.4.19 kernel -

Which come to think of it is basically 2.4.19-aa -

You may want to grab the -aa patches for your
2.4.19 base, that's probably the easiest way to
get 2.4.19 + tux -

Joe

joshua.schnee <at> amd.com wrote:

>Thank you for responding,
>
>I am unaware of any patch that has been ported to 2.4.19.  Is there a ported patch available somewhere?
>
>Yes I had several hunks fail, but as this is not my area of expertise, I am unable to create a work around.
>
>Only Logging, and redirect_logging are enabled, from reading past archive messages I had thought that
referer_logging was causing the errors.  Should I disable all logging?
>
>Thanks
>
>-----Original Message-----
>From: Shane Allen [mailto:shane <at> sell.com]
>Sent: Wednesday, January 08, 2003 9:10 PM
>To: tux-list <at> redhat.com
>Subject: Re: Problems running Tux w/ 2.4.19 generic
>
>

Linux TCP performance

Hi,
I have a generic question that I hope is appropriate on this list.

Are there some sort of performance stats for Linux TCP  that measure throughput, calls/sec, simultaneous
calls supported etc available in public domain? I'm trying to figure out how well a Linux host can do in a
mission-critical environment handling thousands of connections and lots of data traffic, and I thought
it will be nice to seewhat is the best out there.

Regards,
Nagesh

Re: Linux TCP performance

On Thu, Jan 09, 2003 at 12:35:03PM -0500, Nagesh Chintada wrote:

> I have a generic question that I hope is appropriate on this list.
> 
> Are there some sort of performance stats for Linux TCP  that measure throughput, calls/sec, simultaneous
calls supported etc available in public domain?

Search google for ttcp. Small, easy to use, and (by all accounts) gives
accurate data. What else could you ask for? =)

--n

--

-- 
"Look, Clark! I made you an ashtray!"

Re: Problems running Tux w/ 2.4.19 generic

On Thu, Jan 09, 2003 at 08:49:54AM -0600, joshua.schnee <at> amd.com wrote:
> Thank you for responding,
> 
> I am unaware of any patch that has been ported to 2.4.19.  Is there a ported patch available somewhere?

I thought someone had made one...

> Yes I had several hunks fail, but as this is not my area of expertise, I am unable to create a work around.

Then that is likely the source of your problem...

> Only Logging, and redirect_logging are enabled, from reading past archive messages I had thought that
referer_logging was causing the errors.  Should I disable all logging?

I wouldn't worry about it at this point, hunks failing is in my best
guess the source of the problem. Check the archives for the link to my
patch against 2.4.21-pre2. Download and unpack it, and change the following:
-at line 1674, delete the three lines reading:
 EXPORT_SYMBOL(yield);
 EXPORT_SYMBOL(__cond_resched);
 EXPORT_SYMBOL(jiffies);
and add three lines reading:
 EXPORT_SYMBOL(sys_sched_yield);
 EXPORT_SYMBOL(jiffies);
 EXPORT_SYMBOL(xtime);

Then run the modified patch against vanilla 2.4.19. It should patch
without any hunks failing.

The previous suggestion of 2.4.19-aa may work, I have not tried it

Re: Linux TCP performance

Von:     Nagesh Chintada <nchintada <at> aol.com>

> Are there some sort of performance stats for Linux TCP  that measure
> throughput, calls/sec, simultaneous calls supported etc available in public
> domain? I'm trying to figure out how well a Linux host can do in a
> mission-critical environment handling thousands of connections and lots of
> data traffic, and I thought it will be nice to seewhat is the best out
> there.

Check out http://www.specbench.org and especially http://www.specbench.org/web99/results/, a lot
of the results use Tux as the web server.

Greetings,

Johann
----------------------------------------------------------------------------
Arcor-Video on Demand, Ihre Online-Videothek, bietet Ihnen vom 23. Dezember
2002 bis zum 2. Januar 2003 100 festliche Filme zu funkensprühenden Preisen
Festtags-Movies und Silvesterknaller gibt es SCHON AB 99 CENT!
http://www.arcor.de/home/redir.php/FooterMD
-----------------------------------------------------------------------

Re: Linux TCP performance

RE: Problems running Tux w/ 2.4.19 generic

Thanks for the help,

This did in fact stop the hunks from failing when patching the generic 2.4.19 kernel.  However, I am still
seeing Tux drop threads for no reason that I can see.

Has anyone had a problem with Tux dropping threads, and such, no longer responding to requests?

-----Original Message-----
From: Shane Allen [mailto:shane <at> sell.com]
Sent: Thursday, January 09, 2003 1:37 PM
To: tux-list <at> redhat.com
Subject: Re: Problems running Tux w/ 2.4.19 generic

On Thu, Jan 09, 2003 at 08:49:54AM -0600, joshua.schnee <at> amd.com wrote:
> Thank you for responding,
> 
> I am unaware of any patch that has been ported to 2.4.19.  Is there a ported patch available somewhere?

I thought someone had made one...

> Yes I had several hunks fail, but as this is not my area of expertise, I am unable to create a work around.

Then that is likely the source of your problem...

> Only Logging, and redirect_logging are enabled, from reading past archive messages I had thought that
referer_logging was causing the errors.  Should I disable all logging?

I wouldn't worry about it at this point, hunks failing is in my best
guess the source of the problem. Check the archives for the link to my
patch against 2.4.21-pre2. Download and unpack it, and change the following:
-at line 1674, delete the three lines reading:
 EXPORT_SYMBOL(yield);
 EXPORT_SYMBOL(__cond_resched);
 EXPORT_SYMBOL(jiffies);
and add three lines reading:
 EXPORT_SYMBOL(sys_sched_yield);
 EXPORT_SYMBOL(jiffies);
 EXPORT_SYMBOL(xtime);

Then run the modified patch against vanilla 2.4.19. It should patch
without any hunks failing.

The previous suggestion of 2.4.19-aa may work, I have not tried it
myself, and do not run SuSE, so I cannot say.

--

-- 
Shane Allen <shane <at> sell.com>

_______________________________________________
tux-list mailing list
tux-list <at> redhat.com
https://listman.redhat.com/mailman/listinfo/tux-list