Re: [EMAIL PROTECTED] Apache2.0.48 Illegal instruction Segmentation fault issues

2005-09-27 Thread Joe Orton
On Mon, Sep 26, 2005 at 07:36:34PM -0700, Jignesh Badani wrote:
...
 [Mon Sep 26 18:25:45 2005] [notice] child pid 326 exit signal Illegal 
 instruction (4)
 
 The management is wanting to know the cause and so do I. What is the best 
 way to debug this ? I know we could upgrade to the latest 2.0.54 and we 

Follow the instructions on obtaining a backtrace at this URL, and post 
the backtrace here:

  http://httpd.apache.org/dev/debugging.html#crashes

joe

-
The official User-To-User support forum of the Apache HTTP Server Project.
See URL:http://httpd.apache.org/userslist.html for more info.
To unsubscribe, e-mail: [EMAIL PROTECTED]
  from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Re: [EMAIL PROTECTED] Apache2.0.48 Illegal instruction Segmentation fault issues

2005-09-27 Thread Nick Kew
On Tuesday 27 September 2005 03:36, Jignesh Badani wrote:
 Greetings Folks,

 We have been running pretty smoothly 3 instances each of Apache2 (2 of
 them on 2.0.48 and the 3rd one is running 2.0.51) for over an year now on
 2 load balanced Solaris 8 (SunFire 480R 2 - Ultra Sparc 3 900Mhz CPU and 4
 GB RAM) boxes.

If they started misbehaving, what has changed?  Any system libraries?
Kernel?  Hardware and/or drivers?  And if so, have you rebuilt apache
from clean source?

 2 of the instances (2.0.48 - Extranet and 2.0.51 - Intranet) have
 Netegrity Siteminder Agents 

... whatever that is ...

 running on them and nothing else (no perl, no 
 php, nothing. They are straight forward Reverse Proxies). I do have  SSL
 enabled on each of the instances and use a Rainbow CSwift accelerator
 hardware.

Do you have anything in Apache to take advantage of that hardware?

 The management is wanting to know the cause and so do I. What is the best
 way to debug this ?

Apart from running gdb on a core dump?

There are some debugging aids at http://people.apache.org/~trawick/


-- 
Nick Kew

-
The official User-To-User support forum of the Apache HTTP Server Project.
See URL:http://httpd.apache.org/userslist.html for more info.
To unsubscribe, e-mail: [EMAIL PROTECTED]
  from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Re: [EMAIL PROTECTED] Apache2.0.48 Illegal instruction Segmentation fault issues

2005-09-27 Thread Jignesh Badani
Thanks Nick,

If they started misbehaving, what has changed?  Any system libraries?
Kernel?  Hardware and/or drivers?  And if so, have you rebuilt apache
from clean source?

ANSWER: Nothing has changed apart from the OS patches from time to time. 
And nothing recently. Also this problem has occured 3-4 times this year 
and generally once every 2-3 months. 

 2 of the instances (2.0.48 - Extranet and 2.0.51 - Intranet) have
 Netegrity Siteminder Agents 

... whatever that is ...

ANSWER: It is Apache module which is responsible for Access control 
(Auth/Az) and SSO and communicates with a backend Policy Server for 
decision making.

 running on them and nothing else (no perl, no 
 php, nothing. They are straight forward Reverse Proxies). I do have  SSL
 enabled on each of the instances and use a Rainbow CSwift accelerator
 hardware.

Do you have anything in Apache to take advantage of that hardware?

ANSWER: If you mean, any directive in httpd.conf/ssl.conf to make use of 
the Accelerator card, yes, I have compiled Apache with a CFLAGS 
-DSSL_EXPERIMENTAL -DSSL_ENGINE which allows me to use the directive 
SSLCryptoDevice cswift which instructs Apache to offload SSL processing 
to the hardware.

 The management is wanting to know the cause and so do I. What is the 
best
 way to debug this ?

Apart from running gdb on a core dump?

There are some debugging aids at http://people.apache.org/~trawick/

ANSWER:  Thanks. But I believe you need to compile Apache with debug 
option, and second, the Apache instance does not crash, the parent and one 
or 2 child processes are often up. But it is not responding to web 
request.  So I guess there will be no core dump !?


regards
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 
- - 
Jignesh Badani
Intranet/Extranet Technical Services
x13563




Nick Kew [EMAIL PROTECTED] 
09/27/2005 10:38 AM
Please respond to
users@httpd.apache.org


To
users@httpd.apache.org
cc

Subject
Re: [EMAIL PROTECTED] Apache2.0.48  Illegal instruction  Segmentation fault 
issues






On Tuesday 27 September 2005 03:36, Jignesh Badani wrote:
 Greetings Folks,

 We have been running pretty smoothly 3 instances each of Apache2 (2 of
 them on 2.0.48 and the 3rd one is running 2.0.51) for over an year now 
on
 2 load balanced Solaris 8 (SunFire 480R 2 - Ultra Sparc 3 900Mhz CPU and 
4
 GB RAM) boxes.

If they started misbehaving, what has changed?  Any system libraries?
Kernel?  Hardware and/or drivers?  And if so, have you rebuilt apache
from clean source?

 2 of the instances (2.0.48 - Extranet and 2.0.51 - Intranet) have
 Netegrity Siteminder Agents 

... whatever that is ...

 running on them and nothing else (no perl, no 
 php, nothing. They are straight forward Reverse Proxies). I do have  SSL
 enabled on each of the instances and use a Rainbow CSwift accelerator
 hardware.

Do you have anything in Apache to take advantage of that hardware?

 The management is wanting to know the cause and so do I. What is the 
best
 way to debug this ?

Apart from running gdb on a core dump?

There are some debugging aids at http://people.apache.org/~trawick/


-- 
Nick Kew

-
The official User-To-User support forum of the Apache HTTP Server Project.
See URL:http://httpd.apache.org/userslist.html for more info.
To unsubscribe, e-mail: [EMAIL PROTECTED]
  from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]




-
The official User-To-User support forum of the Apache HTTP Server Project.
See URL:http://httpd.apache.org/userslist.html for more info.
To unsubscribe, e-mail: [EMAIL PROTECTED]
  from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Re: [EMAIL PROTECTED] Apache2.0.48 Illegal instruction Segmentation fault issues

2005-09-27 Thread Jignesh Badani
Thanks Geoff,

 Out of interest, what SSL session cache type are you using on these 
hosts? 
 shmcb?

 Cheers,
 Geoff

We are using the default dbm on all 3 Apache instances.

We have had some shared memory errors when using shm* in the past and 
reverted back to the default DBM.

regards
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 
- - 
Jignesh Badani
Intranet/Extranet Technical Services





Geoff Thorpe [EMAIL PROTECTED] 
09/26/2005 08:50 PM

To
users@httpd.apache.org
cc
Jignesh Badani [EMAIL PROTECTED]
Subject
Re: [EMAIL PROTECTED] Apache2.0.48  Illegal instruction  Segmentation fault 
issues






Hi there,

On September 26, 2005 10:36 pm, Jignesh Badani wrote:
 [Mon Sep 26 18:01:09 2005] [notice] child pid 11892 exit signal
 Segmentation fault (11)

 [Mon Sep 26 18:25:03 2005] [notice] child pid 27766 exit signal Illegal
 instruction  (4)
 [Mon Sep 26 18:25:42 2005] [notice] child pid 27765 exit signal Illegal
 instruction (4)
 [Mon Sep 26 18:25:45 2005] [notice] child pid 326 exit signal Illegal
 instruction (4)

Out of interest, what SSL session cache type are you using on these hosts? 

shmcb?

Cheers,
Geoff

-- 
Geoff Thorpe
[EMAIL PROTECTED]
http://www.geoffthorpe.net/

Même ceux qui se sentent pas des nôtres, ne nous voyant plus à genoux,
seront, plus que jamais, chez eux chez nous.
  -- Loco Locass



-
The official User-To-User support forum of the Apache HTTP Server Project.
See URL:http://httpd.apache.org/userslist.html for more info.
To unsubscribe, e-mail: [EMAIL PROTECTED]
  from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



[EMAIL PROTECTED] Apache2.0.48 Illegal instruction Segmentation fault issues

2005-09-26 Thread Jignesh Badani
Greetings Folks,

We have been running pretty smoothly 3 instances each of Apache2 (2 of 
them on 2.0.48 and the 3rd one is running 2.0.51) for over an year now on 
2 load balanced Solaris 8 (SunFire 480R 2 - Ultra Sparc 3 900Mhz CPU and 4 
GB RAM) boxes.

2 of the instances (2.0.48 - Extranet and 2.0.51 - Intranet) have 
Netegrity Siteminder Agents running on them and nothing else (no perl, no 
php, nothing. They are straight forward Reverse Proxies). I do have  SSL 
enabled on each of the instances and use a Rainbow CSwift accelerator 
hardware.

But since the begining of the year, we have had this issue 3-4 times when 
the Extranet instance (on either of the box) stops responding, the child 
processes reduces from 8-10 to 2-3 and the error log is filled with the 
errors below. Even the server-status page fails to load.

It happened again twice today, once on each server but on the same 
Extranet instance which has the web agent. Also fyi, the Extranet instance 
serves bulk of the load, say about 80%. And today it happened very late in 
the day (around 5:00pm and 6:00pm PST) which is normally very light for 
us. The server CPU and memory usages are pretty low (less then 30%).

[Mon Sep 26 18:01:09 2005] [notice] child pid 11892 exit signal 
Segmentation fault (11)

[Mon Sep 26 18:25:03 2005] [notice] child pid 27766 exit signal Illegal 
instruction  (4)
[Mon Sep 26 18:25:42 2005] [notice] child pid 27765 exit signal Illegal 
instruction (4)
[Mon Sep 26 18:25:45 2005] [notice] child pid 326 exit signal Illegal 
instruction (4)

The management is wanting to know the cause and so do I. What is the best 
way to debug this ? I know we could upgrade to the latest 2.0.54 and we 
are in the process to do so and move from Solaris to Suse SLES 9 in about 
2-3 weeks time. Keep in mind, it only happens sporadically and we cannot 
reproduce it. A restart of the instance resolves the issue. 

Here is how I have compiled Apache (config.nice)

CFLAGS=-DSSL_EXPERIMENTAL -DSSL_ENGINE; export CFLAGS
echo $CFLAGS
./configure \
--prefix=/opt/apache2e \
--with-mpm=worker \
--enable-ssl \
--with-ssl=/opt/openssl \
--enable-nonportable-atomics=yes \
--enable-proxy=shared \
--disable-ipv6 \
--cache-file=/export/home/x/install/apache_openssl/httpd-2.0.48e/config.cache
 
\
$@

regards
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 
- - 
Jignesh Badani
Intranet/Extranet Technical Services


-
The official User-To-User support forum of the Apache HTTP Server Project.
See URL:http://httpd.apache.org/userslist.html for more info.
To unsubscribe, e-mail: [EMAIL PROTECTED]
  from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Re: [EMAIL PROTECTED] Apache2.0.48 Illegal instruction Segmentation fault issues

2005-09-26 Thread Geoff Thorpe
Hi there,

On September 26, 2005 10:36 pm, Jignesh Badani wrote:
 [Mon Sep 26 18:01:09 2005] [notice] child pid 11892 exit signal
 Segmentation fault (11)

 [Mon Sep 26 18:25:03 2005] [notice] child pid 27766 exit signal Illegal
 instruction  (4)
 [Mon Sep 26 18:25:42 2005] [notice] child pid 27765 exit signal Illegal
 instruction (4)
 [Mon Sep 26 18:25:45 2005] [notice] child pid 326 exit signal Illegal
 instruction (4)

Out of interest, what SSL session cache type are you using on these hosts? 
shmcb?

Cheers,
Geoff

-- 
Geoff Thorpe
[EMAIL PROTECTED]
http://www.geoffthorpe.net/

Même ceux qui se sentent pas des nôtres, ne nous voyant plus à genoux,
seront, plus que jamais, chez eux chez nous.
  -- Loco Locass

-
The official User-To-User support forum of the Apache HTTP Server Project.
See URL:http://httpd.apache.org/userslist.html for more info.
To unsubscribe, e-mail: [EMAIL PROTECTED]
  from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]