I'm just ringing in, although I don't have a vote. I was working with Rainer on an issue with 1.2.21 on some Red Hat machines. I started with mod_jk because those were the only error messages I was getting. It turns out that those messages were not related to my ultimate problem. Thanks again to Rainer for helping me track that down and sort it out.

My ultimate problem was that apache was returning bad data in some requests, but logging 200 error code and the proper file sizes in the access_log. The files themselves were of variable size and often contained other "junk" that was obviously from apache memory space (chunks of log entries and other things).

The theories I went though included targeting sendfile, mmap, and mod_cache (mod_cache_mem) . I disabled all of these and still was experiencing the problem.

Eventually I stopped it from happening by switching from mod_jk to mod_proxy_ajp. I still have the other things enabled.

Since I'm not a developer I can't say exactly where that problem might be. It does seem as though there is some problem there. I couldn't reproduce it with ab... apache thinks everything is normal. It didn't show up until I put the servers in production, and then it was only a small percentage of overall requests. If I hadn't seen the b0rked images and files through my browser I would not have known the problem existed.

I'd be happy to help troubleshoot this if anyone wants to investigate further.

Thanks;

Eric

Henri Gomez wrote:
Well it shouldn't stop the release since it may be related to iSeries beast.

I should check about compile flags (we don't use configure on iSeries)
and may be I miss something but it seems related to logging area ;(



2007/3/1, Rainer Jung <[EMAIL PROTECTED]>:
Henri,

since this alredy happened in 1.2.20 and it runs well on most platforms,
  think it should not stop the release. But I'm still very interested in
fixing his.

In your last problem report, I think I saw a stack trace from which we
could see the method, were the crash was happening. Maybe we can then
build a debug version, to find out, which things didn't get initialized
properly.

Regards,

Rainer

Henri Gomez wrote:
> Stable on Linux PPC but still problem on iSeries V5R3 :
>
>      0000000018, instruction X'0000'.
>     Pointeur non défini pour position mémoire référencée.
>     Application error.  MCH3601 unmonitored by MOD_JK1221 at statement
>       0000000037, instruction X'0000'.
>     Value for call stack counter parameter not valid.
>     Application error.  CEE9901 unmonitored by QP0WPINT at statement
>       0000000018, instruction X'0000'.
>     Pointeur non défini pour position mémoire référencée.
>     Application error.  MCH3601 unmonitored by MOD_JK1221 at statement
>       0000000037, instruction X'0000'.
>     Value for call stack counter parameter not valid.
>     Application error.  CEE9901 unmonitored by QP0WPINT at statement
>       0000000018, instruction X'0000'.
>
> Apache is 2.0.58 on V5R3 :(
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>

--
--
kippdata
informationstechnologie GmbH   Tel: 0228 98549 -0
Bornheimer Str. 33a            Fax: 0228 98549 -50
53111 Bonn                     www.kippdata.de

HRB 8018 Amtsgericht Bonn / USt.-IdNr. DE 196 457 417
Geschäftsführer: Dr. Thomas Höfer, Rainer Jung, Sven Maurmann
===============================
kippdata
informationstechnologie GmbH   Tel: +49 228 98549 -0
Bornheimer Str. 33a            Fax: +49 228 98549 -50
D-53111 Bonn                   www.kippdata.de

HRB 8018 Amtsgericht Bonn / USt.-IdNr. DE 196 457 417
Geschäftsführer: Dr. Thomas Höfer, Rainer Jung, Sven Maurmann

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to