Does SuSE 9.1 share the problems discussed here 
http://www.apsis.ch/pound/pound_list/archive/2004/2004-03/1080232463000#1080232463000
We are getting the same errors "MONITOR: worker exited nurmally 127, 
restarting..."
Is SuSE 9.x NPTL-based?

I understand that using LD_ASSUME_KERNEL is a viable workaround but are 
there any drawbacks?
Should we be using another Linux distribution?

Best Regards
Joakim Thornberg

On Thursday 04 November 2004 11.24, Joakim Thornberg wrote:
> Does SuSE 9.1 share the problems discussed here
> http://www.apsis.ch/pound/pound_list/archive/2004/2004-03/1080232463000#108
>0232463000 We are getting the same errors "MONITOR: worker exited nurmally
> 127, restarting..."
> Is SuSE 9.x NPTL-based?
>
> I understand that using LD_ASSUME_KERNEL is a viable workaround but are
> there any drawbacks?
> Should we be using another Linux distribution?
>
> Best Regards
> Joakim Thornberg

SuSE 9.1 is NPTL based, but latest -current detects, autoconfigures and runs 
just fine on it. No workaround is required.

The drawback of the LD_ASSUME_KERNEL thingy is that it reverts to the old 
threads model, which is inefficient by comparison. Our tests show NPTL to run 
about 20-40% faster on identical hardware and config.
-- 
Robert Segall
Apsis GmbH
Postfach, Uetikon am See, CH-8707
Tel: +41-1-920 4904

Robert Segall wrote:

>On Thursday 04 November 2004 11.24, Joakim Thornberg wrote:
>  
>
>>Does SuSE 9.1 share the problems discussed here
>>http://www.apsis.ch/pound/pound_list/archive/2004/2004-03/1080232463000#108
>>0232463000 We are getting the same errors "MONITOR: worker exited nurmally
>>127, restarting..."
>>Is SuSE 9.x NPTL-based?
>>
>>I understand that using LD_ASSUME_KERNEL is a viable workaround but are
>>there any drawbacks?
>>Should we be using another Linux distribution?
>>
>>Best Regards
>>Joakim Thornberg
>>    
>>
>
>SuSE 9.1 is NPTL based, but latest -current detects, autoconfigures and runs 
>just fine on it. No workaround is required.
>
>The drawback of the LD_ASSUME_KERNEL thingy is that it reverts to the old 
>threads model, which is inefficient by comparison. Our tests show NPTL to run 
>about 20-40% faster on identical hardware and config.
>  
>
Great news Robert.
Looking forward to the 1.8 release wich i assume will incorporate the 
above change.

Best Regards
Joakim Thornberg

Joakim Thornberg wrote:

> Robert Segall wrote:
>
>> On Thursday 04 November 2004 11.24, Joakim Thornberg wrote:
>>  
>>
>>> Does SuSE 9.1 share the problems discussed here
>>> http://www.apsis.ch/pound/pound_list/archive/2004/2004-03/1080232463000#108

>>>
>>> 0232463000 We are getting the same errors "MONITOR: worker exited 
>>> nurmally
>>> 127, restarting..."
>>> Is SuSE 9.x NPTL-based?
>>>
>>> I understand that using LD_ASSUME_KERNEL is a viable workaround but are
>>> there any drawbacks?
>>> Should we be using another Linux distribution?
>>>
>>> Best Regards
>>> Joakim Thornberg
>>>   
>>
>>
>> SuSE 9.1 is NPTL based, but latest -current detects, autoconfigures 
>> and runs just fine on it. No workaround is required.
>>
>> The drawback of the LD_ASSUME_KERNEL thingy is that it reverts to the 
>> old threads model, which is inefficient by comparison. Our tests show 
>> NPTL to run about 20-40% faster on identical hardware and config.
>>  
>>
> Great news Robert.
> Looking forward to the 1.8 release wich i assume will incorporate the 
> above change.
>
> Best Regards
> Joakim Thornberg
>
Still getting these errors after installing Pound 1.8 on SuSe 9.1.
Happens every third request or so.

Nov  8 01:41:45 fantomen pound: MONITOR: worker exited normally 127, 
restarting...
Nov  8 01:42:49 fantomen pound: MONITOR: worker exited on signal 11, 
restarting...

Best Regards
Joakim Thornberg

On Monday 08 November 2004 09.27, Joakim Thornberg wrote:
> Still getting these errors after installing Pound 1.8 on SuSe 9.1.
> Happens every third request or so.
>
> Nov  8 01:41:45 fantomen pound: MONITOR: worker exited normally 127,
> restarting...
> Nov  8 01:42:49 fantomen pound: MONITOR: worker exited on signal 11,
> restarting...

You'll have to provide more info (when exactly does it happen, can you 
reproduce it or is it random, etc) for us to help.

Do you run it as root? Have you defined sufficient resources (CPU time, 
memory) for it?

Are you sure this is Pound? I am not aware of any way in which the program 
could exit: look at the code and you'll see the main loop runs forever, thus 
a "normal exit" is just not possible. A SIGSEGV might happen (a bug to be 
sure) but a normal exit is curious.
-- 
Robert Segall
Apsis GmbH
Postfach, Uetikon am See, CH-8707
Tel: +41-1-920 4904

Robert Segall wrote:

>On Monday 08 November 2004 09.27, Joakim Thornberg wrote:
>  
>
>>Still getting these errors after installing Pound 1.8 on SuSe 9.1.
>>Happens every third request or so.
>>
>>Nov  8 01:41:45 fantomen pound: MONITOR: worker exited normally 127,
>>restarting...
>>Nov  8 01:42:49 fantomen pound: MONITOR: worker exited on signal 11,
>>restarting...
>>    
>>
>
>You'll have to provide more info (when exactly does it happen, can you 
>reproduce it or is it random, etc) for us to help.
>
>Do you run it as root? Have you defined sufficient resources (CPU time, 
>memory) for it?
>
>Are you sure this is Pound? I am not aware of any way in which the program 
>could exit: look at the code and you'll see the main loop runs forever, thus 
>a "normal exit" is just not possible. A SIGSEGV might happen (a bug to be 
>sure) but a normal exit is curious.
>  
>
It happens when we start Pound 1.8 with /usr/sbin/pound -f 
/etc/pound/pound.cfg
If we start it with LD_ASSUME_KERNEL=2.4.19 /usr/sbin/pound -f 
/etc/pound/pound.cfg everything works fine.
Exactly the same problem as in 1.7.

Perhaps the thread detection isn't working as intended?

I'm pretty sure its Pound. Started without LD_ASSUME_KERNEL it restarts 
every third request or so.

Nov  8 03:54:06 fantomen pound: 192.168.123.20 192.168.123.6 - - 
[08/Nov/2004:03
:54:06 -0800] "GET /hillstreet/hs.nsf/DumpLUL2.gif HTTP/1.1" 200 4492 
"http://19
2.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U; 
Windows NT 5.1
; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
Nov  8 03:54:06 fantomen pound: MONITOR: worker exited normally 127, 
restarting.
..
Nov  8 03:54:08 fantomen pound: 192.168.123.20 192.168.123.6 - - 
[08/Nov/2004:03
:54:08 -0800] "GET /hillstreet/hs.nsf/DumpFilmdelta2.gif HTTP/1.1" 200 
4329 "htt
p://192.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U; 
Windows
NT 5.1; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
Nov  8 03:54:18 fantomen pound: MONITOR: worker exited normally 127, 
restarting...


Best regards
Joakim Thornberg

On Monday 08 November 2004 11.47, Joakim Thornberg wrote:
> It happens when we start Pound 1.8 with /usr/sbin/pound -f
> /etc/pound/pound.cfg
> If we start it with LD_ASSUME_KERNEL=2.4.19 /usr/sbin/pound -f
> /etc/pound/pound.cfg everything works fine.
> Exactly the same problem as in 1.7.
>
> Perhaps the thread detection isn't working as intended?
>
> I'm pretty sure its Pound. Started without LD_ASSUME_KERNEL it restarts
> every third request or so.
>
> Nov  8 03:54:06 fantomen pound: 192.168.123.20 192.168.123.6 - -
> [08/Nov/2004:03
>
> :54:06 -0800] "GET /hillstreet/hs.nsf/DumpLUL2.gif HTTP/1.1" 200 4492
>
> "http://19
> 2.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U;
> Windows NT 5.1
> ; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
> Nov  8 03:54:06 fantomen pound: MONITOR: worker exited normally 127,
> restarting.
> ..
> Nov  8 03:54:08 fantomen pound: 192.168.123.20 192.168.123.6 - -
> [08/Nov/2004:03
>
> :54:08 -0800] "GET /hillstreet/hs.nsf/DumpFilmdelta2.gif HTTP/1.1" 200
>
> 4329 "htt
> p://192.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U;
> Windows
> NT 5.1; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
> Nov  8 03:54:18 fantomen pound: MONITOR: worker exited normally 127,
> restarting...

We use Pound on SuSE 9.1 and have seen no problems at all, even under serious 
load. Could it be you have some strange installation? Bad RAM? Have you 
updated your system regularly (I show gcc/libgcc 3.3.3-41, glibc 2.3.3-97 and 
openssl 0.9.7d-15)? Could you possibly try on a fresh install? If it sounds 
like I'm grasping at straws it is because I am.
-- 
Robert Segall
Apsis GmbH
Postfach, Uetikon am See, CH-8707
Tel: +41-1-920 4904

Robert Segall wrote:

>On Monday 08 November 2004 11.47, Joakim Thornberg wrote:
>  
>
>>It happens when we start Pound 1.8 with /usr/sbin/pound -f
>>/etc/pound/pound.cfg
>>If we start it with LD_ASSUME_KERNEL=2.4.19 /usr/sbin/pound -f
>>/etc/pound/pound.cfg everything works fine.
>>Exactly the same problem as in 1.7.
>>
>>Perhaps the thread detection isn't working as intended?
>>
>>I'm pretty sure its Pound. Started without LD_ASSUME_KERNEL it restarts
>>every third request or so.
>>
>>Nov  8 03:54:06 fantomen pound: 192.168.123.20 192.168.123.6 - -
>>[08/Nov/2004:03
>>
>>:54:06 -0800] "GET /hillstreet/hs.nsf/DumpLUL2.gif HTTP/1.1" 200 4492
>>
>>"http://19
>>2.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U;
>>Windows NT 5.1
>>; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
>>Nov  8 03:54:06 fantomen pound: MONITOR: worker exited normally 127,
>>restarting.
>>..
>>Nov  8 03:54:08 fantomen pound: 192.168.123.20 192.168.123.6 - -
>>[08/Nov/2004:03
>>
>>:54:08 -0800] "GET /hillstreet/hs.nsf/DumpFilmdelta2.gif HTTP/1.1" 200
>>
>>4329 "htt
>>p://192.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U;
>>Windows
>>NT 5.1; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
>>Nov  8 03:54:18 fantomen pound: MONITOR: worker exited normally 127,
>>restarting...
>>    
>>
>
>We use Pound on SuSE 9.1 and have seen no problems at all, even under serious 
>load. Could it be you have some strange installation? Bad RAM? Have you 
>updated your system regularly (I show gcc/libgcc 3.3.3-41, glibc 2.3.3-97 and 
>openssl 0.9.7d-15)? Could you possibly try on a fresh install? If it sounds 
>like I'm grasping at straws it is because I am.
>  
>
openssl 0.9.7d-15
gcc/libgcc 3.3.3-41
glibc 2.3.3-93

The system is a fresh install. SuSE 9.1 Minimal Config.
I'm only running Heartbeat and Pound on it with the necessary components 
for them to run.

Best Regards
Joakim Thornberg

Simon Matter wrote:

>>Robert Segall wrote:
>>
>>    
>>
>>>On Monday 08 November 2004 09.27, Joakim Thornberg wrote:
>>>
>>>
>>>      
>>>
>>>>Still getting these errors after installing Pound 1.8 on SuSe 9.1.
>>>>Happens every third request or so.
>>>>
>>>>Nov  8 01:41:45 fantomen pound: MONITOR: worker exited normally 127,
>>>>restarting...
>>>>Nov  8 01:42:49 fantomen pound: MONITOR: worker exited on signal 11,
>>>>restarting...
>>>>
>>>>
>>>>        
>>>>
>>>You'll have to provide more info (when exactly does it happen, can you
>>>reproduce it or is it random, etc) for us to help.
>>>
>>>Do you run it as root? Have you defined sufficient resources (CPU time,
>>>memory) for it?
>>>
>>>Are you sure this is Pound? I am not aware of any way in which the
>>>program
>>>could exit: look at the code and you'll see the main loop runs forever,
>>>thus
>>>a "normal exit" is just not possible. A SIGSEGV might happen (a bug to be
>>>sure) but a normal exit is curious.
>>>
>>>
>>>      
>>>
>>It happens when we start Pound 1.8 with /usr/sbin/pound -f
>>/etc/pound/pound.cfg
>>If we start it with LD_ASSUME_KERNEL=2.4.19 /usr/sbin/pound -f
>>/etc/pound/pound.cfg everything works fine.
>>Exactly the same problem as in 1.7.
>>
>>Perhaps the thread detection isn't working as intended?
>>
>>I'm pretty sure its Pound. Started without LD_ASSUME_KERNEL it restarts
>>every third request or so.
>>
>>Nov  8 03:54:06 fantomen pound: 192.168.123.20 192.168.123.6 - -
>>[08/Nov/2004:03
>>:54:06 -0800] "GET /hillstreet/hs.nsf/DumpLUL2.gif HTTP/1.1" 200 4492
>>"http://19
>>2.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U;
>>Windows NT 5.1
>>; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
>>Nov  8 03:54:06 fantomen pound: MONITOR: worker exited normally 127,
>>restarting.
>>..
>>Nov  8 03:54:08 fantomen pound: 192.168.123.20 192.168.123.6 - -
>>[08/Nov/2004:03
>>:54:08 -0800] "GET /hillstreet/hs.nsf/DumpFilmdelta2.gif HTTP/1.1" 200
>>4329 "htt
>>p://192.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U;
>>Windows
>>NT 5.1; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
>>Nov  8 03:54:18 fantomen pound: MONITOR: worker exited normally 127,
>>restarting...
>>    
>>
>
>Whats your CPU? At least on RedHat we had some problems with NPTL which
>only showed up on some CPU's like AMD K6. This has finally been fixed in
>Fedora Core 3. I don't know SuSE but maybe there is also a possible
>problem here.
>
>Simon
>
>  
>
The box is a P3 800Mhz with 256Mb RAM.

Best Regards
Joakim Thornberg

> Robert Segall wrote:
>
>>On Monday 08 November 2004 09.27, Joakim Thornberg wrote:
>>
>>
>>>Still getting these errors after installing Pound 1.8 on SuSe 9.1.
>>>Happens every third request or so.
>>>
>>>Nov  8 01:41:45 fantomen pound: MONITOR: worker exited normally 127,
>>>restarting...
>>>Nov  8 01:42:49 fantomen pound: MONITOR: worker exited on signal 11,
>>>restarting...
>>>
>>>
>>
>>You'll have to provide more info (when exactly does it happen, can you
>>reproduce it or is it random, etc) for us to help.
>>
>>Do you run it as root? Have you defined sufficient resources (CPU time,
>>memory) for it?
>>
>>Are you sure this is Pound? I am not aware of any way in which the
>> program
>>could exit: look at the code and you'll see the main loop runs forever,
>> thus
>>a "normal exit" is just not possible. A SIGSEGV might happen (a bug to be
>>sure) but a normal exit is curious.
>>
>>
> It happens when we start Pound 1.8 with /usr/sbin/pound -f
> /etc/pound/pound.cfg
> If we start it with LD_ASSUME_KERNEL=2.4.19 /usr/sbin/pound -f
> /etc/pound/pound.cfg everything works fine.
> Exactly the same problem as in 1.7.
>
> Perhaps the thread detection isn't working as intended?
>
> I'm pretty sure its Pound. Started without LD_ASSUME_KERNEL it restarts
> every third request or so.
>
> Nov  8 03:54:06 fantomen pound: 192.168.123.20 192.168.123.6 - -
> [08/Nov/2004:03
> :54:06 -0800] "GET /hillstreet/hs.nsf/DumpLUL2.gif HTTP/1.1" 200 4492
> "http://19
> 2.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U;
> Windows NT 5.1
> ; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
> Nov  8 03:54:06 fantomen pound: MONITOR: worker exited normally 127,
> restarting.
> ..
> Nov  8 03:54:08 fantomen pound: 192.168.123.20 192.168.123.6 - -
> [08/Nov/2004:03
> :54:08 -0800] "GET /hillstreet/hs.nsf/DumpFilmdelta2.gif HTTP/1.1" 200
> 4329 "htt
> p://192.168.123.20/hillstreet/hs.nsf/clients" "Mozilla/5.0 (Windows; U;
> Windows
> NT 5.1; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1"
> Nov  8 03:54:18 fantomen pound: MONITOR: worker exited normally 127,
> restarting...

Whats your CPU? At least on RedHat we had some problems with NPTL which
only showed up on some CPU's like AMD K6. This has finally been fixed in
Fedora Core 3. I don't know SuSE but maybe there is also a possible
problem here.

Simon

On Monday 08 November 2004 12.20, Joakim Thornberg wrote:
> openssl 0.9.7d-15
> gcc/libgcc 3.3.3-41
> glibc 2.3.3-93
>
> The system is a fresh install. SuSE 9.1 Minimal Config.
> I'm only running Heartbeat and Pound on it with the necessary components
> for them to run.

Try running a memory check and see if that shows anything, have a look at 
CPU/case temperatures, generally check your hardware. Update your glibc (just 
in case). Make sure your limits are OK.

We have Pound running just fine on an identical config (P III/800, 256M, stock 
SuSE 9.1 kernel and libs), so there's not much else I can think of, unless 
you want to try installing it on another box.
-- 
Robert Segall
Apsis GmbH
Postfach, Uetikon am See, CH-8707
Tel: +41-1-920 4904

Robert Segall wrote:

>On Monday 08 November 2004 12.20, Joakim Thornberg wrote:
>  
>
>>openssl 0.9.7d-15
>>gcc/libgcc 3.3.3-41
>>glibc 2.3.3-93
>>
>>The system is a fresh install. SuSE 9.1 Minimal Config.
>>I'm only running Heartbeat and Pound on it with the necessary components
>>for them to run.
>>    
>>
>
>Try running a memory check and see if that shows anything, have a look at 
>CPU/case temperatures, generally check your hardware. Update your glibc (just 
>in case). Make sure your limits are OK.
>
>We have Pound running just fine on an identical config (P III/800, 256M, stock

>SuSE 9.1 kernel and libs), so there's not much else I can think of, unless 
>you want to try installing it on another box.
>  
>
Reinstalled with all the latest patches. Installed Pound 1.8 and we 
still get the same errors.
This is with only one webbrowser sending requests..

Not sure what you mean by Make sure your limits are OK. Should we do 
something at install like increase threads?

Best Regards
Joakim Thornberg

On Wednesday 10 November 2004 10.13, Joakim Thornberg wrote:
> Reinstalled with all the latest patches. Installed Pound 1.8 and we
> still get the same errors.
> This is with only one webbrowser sending requests..

Sorry to hear that. As I have said in a previous post I have personal, direct 
experience of Pound running fine on SuSE 9.1. I have no idea where to go from 
here - anybody else?

BTW - did you reinstall on the same hardware as before or on another box?

After reinstalling (and before compiling Pound) did you run all the SuSE 
updates (kernel, libraries)?

Where did you get Pound from and did you verify the distribution (GPG 
signature)?

> Not sure what you mean by Make sure your limits are OK. Should we do
> something at install like increase threads?

Not at install, but at runtime: there usually are limits on what resources a 
process may use (sh/Bash/Ksh: ulimit -a, Csh/Tcsh: limit). Make sure you have 
allowed enough for Pound to run.

If you try to run as non-root you should also make sure you use unprivileged 
ports (> 1024).
-- 
Robert Segall
Apsis GmbH
Postfach, Uetikon am See, CH-8707
Tel: +41-1-920 4904

Robert Segall wrote:

>On Wednesday 10 November 2004 10.13, Joakim Thornberg wrote:
>  
>
>>Reinstalled with all the latest patches. Installed Pound 1.8 and we
>>still get the same errors.
>>This is with only one webbrowser sending requests..
>>    
>>
>
>Sorry to hear that. As I have said in a previous post I have personal, direct 
>experience of Pound running fine on SuSE 9.1. I have no idea where to go from 
>here - anybody else?
>
>BTW - did you reinstall on the same hardware as before or on another box?
>
>After reinstalling (and before compiling Pound) did you run all the SuSE 
>updates (kernel, libraries)?
>
>Where did you get Pound from and did you verify the distribution (GPG 
>signature)?
>
>  
>
>>Not sure what you mean by Make sure your limits are OK. Should we do
>>something at install like increase threads?
>>    
>>
>
>Not at install, but at runtime: there usually are limits on what resources a 
>process may use (sh/Bash/Ksh: ulimit -a, Csh/Tcsh: limit). Make sure you have 
>allowed enough for Pound to run.
>
>If you try to run as non-root you should also make sure you use unprivileged 
>ports (> 1024).
>  
>
I did reinstall on the same hardware. Will try another box today and see 
what happens.
Also did the online update and updated everything.
I downloaded Pound from www.apsis.ch.

Heres my ulimit -a as root (im running Pound as root atm)..
core file size        (blocks, -c) 0
data seg size         (kbytes, -d) unlimited
file size             (blocks, -f) unlimited
max locked memory     (kbytes, -l) unlimited
max memory size       (kbytes, -m) unlimited
open files                    (-n) 1024
pipe size          (512 bytes, -p) 8
stack size            (kbytes, -s) unlimited
cpu time             (seconds, -t) unlimited
max user processes            (-u) 2047
virtual memory        (kbytes, -v) unlimited

Best Regards
Joakim Thornberg

I doubt this will be of any use, but I also use Pound
on SuSE 9.1 and have never had any problems.

Our install process is very plain. Specifically, we
use all the defaults except under software we add
'C/C++ Compiler and Tools' and remove 'Office
Applications', then we use the online update, and then
compile OpenSSL and Pound using the defaults. We then
run Pound as root using a pretty simple conf file.

We've compiled everything from the original Pound 1.7
with OpenSSL 0.9.7c on a SuSE 9.1 without any updates
all the way through Pound 1.8 with OpenSSL 0.9.7e on a
SuSE 9.1 updated through this past weekend.

Sorry - hope you have better luck with your other box!



-----Original Message-----
From: Robert Segall [mailto:roseg(at)apsis.ch] 
Sent: Wednesday, November 10, 2004 3:52 AM
To: pound(at)apsis.ch
Subject: Re: SuSE 9.1

On Wednesday 10 November 2004 10.13, Joakim Thornberg
wrote:
> Reinstalled with all the latest patches. Installed
Pound 1.8 and we
> still get the same errors.
> This is with only one webbrowser sending requests..

Sorry to hear that. As I have said in a previous post
I have personal, direct 
experience of Pound running fine on SuSE 9.1. I have
no idea where to go from 
here - anybody else?

BTW - did you reinstall on the same hardware as before
or on another box?

After reinstalling (and before compiling Pound) did
you run all the SuSE 
updates (kernel, libraries)?

Where did you get Pound from and did you verify the
distribution (GPG 
signature)?

> Not sure what you mean by Make sure your limits are
OK. Should we do
> something at install like increase threads?

Not at install, but at runtime: there usually are
limits on what resources a 
process may use (sh/Bash/Ksh: ulimit -a, Csh/Tcsh:
limit). Make sure you have 
allowed enough for Pound to run.

If you try to run as non-root you should also make
sure you use unprivileged 
ports (> 1024).
-- 
Robert Segall
Apsis GmbH
Postfach, Uetikon am See, CH-8707
Tel: +41-1-920 4904



		
__________________________________ 
Do you Yahoo!? 
Check out the new Yahoo! Front Page. 
www.yahoo.com 
 

On Wednesday 10 November 2004 13.15, Joakim Thornberg wrote:
> I did reinstall on the same hardware. Will try another box today and see
> what happens.

Please do - hardware is (or at least I suspect it to be) the most likely 
culprit.

> Also did the online update and updated everything.
> I downloaded Pound from www.apsis.ch.

Did you check the signature? I expect it to be OK, but one never knows...

> Heres my ulimit -a as root (im running Pound as root atm)..
> core file size        (blocks, -c) 0
> data seg size         (kbytes, -d) unlimited
> file size             (blocks, -f) unlimited
> max locked memory     (kbytes, -l) unlimited
> max memory size       (kbytes, -m) unlimited
> open files                    (-n) 1024
> pipe size          (512 bytes, -p) 8
> stack size            (kbytes, -s) unlimited
> cpu time             (seconds, -t) unlimited
> max user processes            (-u) 2047
> virtual memory        (kbytes, -v) unlimited

That looks fine.
-- 
Robert Segall
Apsis GmbH
Postfach, Uetikon am See, CH-8707
Tel: +41-1-920 4904

Robert Segall wrote:

>On Wednesday 10 November 2004 13.15, Joakim Thornberg wrote:
>  
>
>>I did reinstall on the same hardware. Will try another box today and see
>>what happens.
>>    
>>
>
>Please do - hardware is (or at least I suspect it to be) the most likely 
>culprit.
>
>  
>
>>Also did the online update and updated everything.
>>I downloaded Pound from www.apsis.ch.
>>    
>>
>
>Did you check the signature? I expect it to be OK, but one never knows...
>
>  
>
>>Heres my ulimit -a as root (im running Pound as root atm)..
>>core file size        (blocks, -c) 0
>>data seg size         (kbytes, -d) unlimited
>>file size             (blocks, -f) unlimited
>>max locked memory     (kbytes, -l) unlimited
>>max memory size       (kbytes, -m) unlimited
>>open files                    (-n) 1024
>>pipe size          (512 bytes, -p) 8
>>stack size            (kbytes, -s) unlimited
>>cpu time             (seconds, -t) unlimited
>>max user processes            (-u) 2047
>>virtual memory        (kbytes, -v) unlimited
>>    
>>
>
>That looks fine.
>  
>
Reinstalled on another box, identical except for the harddrive.
Still the same problems,

I did however try commenting out RootJail this time wich also seems to 
fix it.
What are the drawbacks of doing this?

Best Regards
Joakim Thornberg

On Friday 12 November 2004 15.09, Joakim Thornberg wrote:
> I did however try commenting out RootJail this time wich also seems to
> fix it.

When running with RootJail did you create the device links as described in the 
README/FAQ?

> What are the drawbacks of doing this?

Not much, but some people want the extra insurance.
-- 
Robert Segall
Apsis GmbH
Postfach, Uetikon am See, CH-8707
Tel: +41-1-920 4904