Client Dies HP-UX 11.11 and 11.23 with no error
list Camp
Hello all I am currently using 4.1.2p1. I have about 40 clients mixed HP-UX servers. All are working except for two. The client dies and there is no core, no error message to indicate why. One client is HP-UX 11.11 and the other is 11.23. Both were compiled using gcc 4.1.1 and are 32 bit. Any hints are suggestions about how I can troubleshoot this would be greatly appreciated.
list Henrik Størner
▸
On Tue, Jun 13, 2006 at 12:47:43PM -0400, user-ec8e46bcfbb3@xymon.invalid wrote:
I am currently using 4.1.2p1. I have about 40 clients mixed HP-UX servers. All are working except for two. The client dies and there is no core, no error message to indicate why. One client is HP-UX 11.11 and the other is 11.23. Both were compiled using gcc 4.1.1 and are 32 bit. Any hints are suggestions about how I can troubleshoot this would be greatly appreciated.
First, check for strange stuff in the ~hobbit/client/logs/* files. Is hobbitlaunch running ? Is the file ~hobbit/client/tmp/msg.HOSTNAME.txt generated ? This is the first step, where data is collected and packed into a client message. Finally, check if the "bb" utility in ~hobbit/client/bin/bb works. Just try to run it - it should print out a usage message. Regards, Henrik
list Camp
Nothing strange in the logs dir. The clientlaunch.log file shows that it is starting and loading the tasklist. Hobbitlaunch is not running. Yes, msg.txt exists but it does not have the hostname as part of the filename. Running the "bb" command does produce a usage message. This is so strange. It will startup with no problems, but after 5-15 minutes (not sure) it just completely dies. :-(
▸
-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid]
Sent: Tuesday, June 13, 2006 5:03 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] Client Dies HP-UX 11.11 and 11.23 with no error
On Tue, Jun 13, 2006 at 12:47:43PM -0400, user-ec8e46bcfbb3@xymon.invalid
wrote:I am currently using 4.1.2p1. I have about 40 clients mixed HP-UX servers. All are working except for two. The client dies and there is no core, no error message to indicate why. One client is HP-UX 11.11 and the other is 11.23. Both were compiled using gcc 4.1.1 and are 32 bit. Any hints are suggestions about how I can troubleshoot this would be greatly appreciated.
First, check for strange stuff in the ~hobbit/client/logs/* files. Is hobbitlaunch running ? Is the file ~hobbit/client/tmp/msg.HOSTNAME.txt generated ? This is the first step, where data is collected and packed into a client message. Finally, check if the "bb" utility in ~hobbit/client/bin/bb works. Just try to run it - it should print out a usage message. Regards, Henrik
list Camp
It happened again. I tried adjusting the vmstat from 300 to 290 just as a guess, but no luck. I cannot find any cores, any logs anything that indicates anything going on. Is there some sort of debug mode I can turn on or tracing I can do to figure out why it is dying?
▸
-----Original Message-----
From: Camp Neil (QNA Information Technology External)
Sent: Wednesday, June 14, 2006 7:08 AM
To: user-ae9b8668bcde@xymon.invalid
Subject: RE: [hobbit] Client Dies HP-UX 11.11 and 11.23 with no error
Nothing strange in the logs dir. The clientlaunch.log file shows that it
is starting and loading the tasklist.
Hobbitlaunch is not running.
Yes, msg.txt exists but it does not have the hostname as part of the
filename.
Running the "bb" command does produce a usage message.
This is so strange. It will startup with no problems, but after 5-15
minutes (not sure) it just completely dies. :-(
-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid]
Sent: Tuesday, June 13, 2006 5:03 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] Client Dies HP-UX 11.11 and 11.23 with no error
On Tue, Jun 13, 2006 at 12:47:43PM -0400, user-ec8e46bcfbb3@xymon.invalid
wrote:I am currently using 4.1.2p1. I have about 40 clients mixed HP-UX servers. All are working except for two. The client dies and there is no core, no error message to indicate why. One client is HP-UX 11.11 and the other is 11.23. Both were compiled using gcc 4.1.1 and are 32 bit. Any hints are suggestions about how I can troubleshoot this would be greatly appreciated.
First, check for strange stuff in the ~hobbit/client/logs/* files. Is hobbitlaunch running ? Is the file ~hobbit/client/tmp/msg.HOSTNAME.txt generated ? This is the first step, where data is collected and packed into a client message. Finally, check if the "bb" utility in ~hobbit/client/bin/bb works. Just try to run it - it should print out a usage message. Regards, Henrik
list Henrik Størner
▸
On Thu, Jun 15, 2006 at 07:48:53AM -0400, user-ec8e46bcfbb3@xymon.invalid wrote:
It happened again. I tried adjusting the vmstat from 300 to 290 just as a guess, but no luck. I cannot find any cores, any logs anything that indicates anything going on. Is there some sort of debug mode I can turn on or tracing I can do to figure out why it is dying?
Login as the hobbit client user, cd to the the client/ directory, then cut-and-paste this to start the hobbitlaunch utility as a foreground task with debug output enabled: MACHINEDOTS="`uname -n`" \ MACHINE="`uname -n | sed -e's/\./,/g'`" \ BBOSTYPE="`uname -s | tr '[A-Z]' '[a-z]'`" \ BBOSSCRIPT="hobbitclient-`uname -s | tr '[A-Z]' '[a-z]'`.sh" \ HOBBITCLIENTHOME="`pwd`" \ bin/hobbitlaunch --config=etc/clientlaunch.cfg --no-daemon --debug Henrik
list Camp
Thanks this really helped! Come to find out the processes were being terminated. I asked the admins, and come to find out they had some scripts that killed "Orphaned" processes. They made modifications to the script, and now I am golden. Thanks for your help!
▸
-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid]
Sent: Thursday, June 15, 2006 9:08 AM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] Client Dies HP-UX 11.11 and 11.23 with no error
On Thu, Jun 15, 2006 at 07:48:53AM -0400, user-ec8e46bcfbb3@xymon.invalid
wrote:It happened again. I tried adjusting the vmstat from 300 to 290 just as a guess, but no luck. I cannot find any cores, any logs anything that indicates anything going on. Is there some sort of debug mode I can turn on or tracing I can do to figure out why it is dying?
Login as the hobbit client user, cd to the the client/ directory, then cut-and-paste this to start the hobbitlaunch utility as a foreground task with debug output enabled: MACHINEDOTS="`uname -n`" \ MACHINE="`uname -n | sed -e's/\./,/g'`" \ BBOSTYPE="`uname -s | tr '[A-Z]' '[a-z]'`" \ BBOSSCRIPT="hobbitclient-`uname -s | tr '[A-Z]' '[a-z]'`.sh" \ HOBBITCLIENTHOME="`pwd`" \ bin/hobbitlaunch --config=etc/clientlaunch.cfg --no-daemon --debug Henrik