Re: [hobbit] [unsolved]defunct problem on sun 5.7

[John Glowacki <johng (at) idttechnology.com>]

I am seeing defunct processes on a few solaris 7 clients too. Most 7 
clients work OK. I have not had time to look into why. It seems to run 
once which sends the status to the hobbit server OK. Then the processes 
go defunct and a second status never gets sent. Results are purple after 
30 minutes. My first guess is that it is a patch level issue. So far all 
working servers are at 106541-16 or higher.

# ps -ef | grep bb
      bb 22873     1  0 10:08:20 ?        0:00 
/opt/hobbit/client/bin/hobbitlaunch 
--config=/opt/hobbit/client/etc/clientlaunc
      bb 22911 22909  0 10:08:22 ?        0:00 vmstat 300 2
      bb   549     1  0   Aug 19 ?        0:16 
/opt/bbc1.9e-btf/bin/bbrun -a /opt/bbc1.9e-btf/bin/bb-local.sh
    root 22921 22195  0 10:08:33 pts/0    0:00 grep bb
      bb 22875 22873  1 10:08:20 ?        0:00 
/opt/RICHPse/bin/se.sparcv9.5.7 /opt/hobbit/client/etc/hobbit.se
      bb 22909     1  0 10:08:22 ?        0:00 sh -c vmstat 300 2 
1>/opt/hobbit/client/tmp/hobbit_vmstat.22878 2>&1; mv /opt/h
      bb 22876 22873  0                   0:00 <defunct>
      bb 22874 22873  0                   0:00 <defunct>

# cat clientlaunch.log
2006-09-06 10:08:20 hobbitlaunch starting
2006-09-06 10:08:20 Loading tasklist configuration from 
/opt/hobbit/client/etc/clientlaunch.cfg
# ps -ef | grep bb
      bb 22873     1  0 10:08:20 ?        0:00 
/opt/hobbit/client/bin/hobbitlaunch 
--config=/opt/hobbit/client/etc/clientlaunc
      bb   549     1  0   Aug 19 ?        0:16 
/opt/bbc1.9e-btf/bin/bbrun -a /opt/bbc1.9e-btf/bin/bb-local.sh
      bb 22875 22873  0 10:08:20 ?        0:00 
/opt/RICHPse/bin/se.sparcv9.5.7 /opt/hobbit/client/etc/hobbit.se
      bb 22876 22873  0                   0:00 <defunct>
      bb 22874 22873  0                   0:00 <defunct>

John

thomas.seglard.enata (at) cnp.fr wrote:
> Hello,
>
> I wrote some weeks ago about a 'defunct' problem on a solaris 7 client. 
> Today I got this problem on another server... This problem worries me a 
> bit as I can't find any clues on how to solve it. Is it a bug in hobbit 
> client ? A bug in the system ? Do you encounter this type of error under 
> hobbit or under solaris ? 
> All system commands normally respond when I separately launch them. All my 
> other applications on this server are normally working. I really don't 
> understand the problem...
> Thanks for your help.
> Best regards,
>
> Thomas
>
> __________________
>
> Hello,
>
> I got a strange behaviour on a server with solaris 7, I got this output 
> when I launch client :
>
> $ ./runclient.sh start
> Hobbit client for sunos started on edsvideo
> $ ps -efd|grep hobbit
>   hobbit 21806 21805  0 11:46:21 ?        0:00 /usr/lib/sa/sadc 300 2
>   hobbit 21898 21897  0 11:46:24 ?        0:00 vmstat 300 2
>   hobbit 21780     1  0 11:46:20 ?        0:00 
> /opt/hobbit/client/bin/hobbitlaunch 
> --config=/opt/hobbit/client/etc/clientlaunc
>   hobbit 21785 21780  0                   0:00 <defunct>
>   hobbit 21781 21780  0                   0:00 <defunct>
>   hobbit 21784 21780  0                   0:00 <defunct>
>   hobbit 21787 21780  0                   0:00 <defunct>
>   hobbit 21783 21780  0 11:46:20 ?        0:00 /usr/bin/sh 
> /opt/hobbit/client/ext/bb-sar.sh
>   hobbit 21897     1  0 11:46:24 ?        0:00 sh -c vmstat 300 2 
> 1>/opt/hobbit/client/tmp/hobbit_vmstat.edsvideo.21811 2>&1; 
>   hobbit 21904 21903  0 11:46:24 ?        0:00 iostat -c 300 2
>   hobbit 21903     1  0 11:46:24 ?        0:00 sh -c iostat -c 300 2 
> 1>/opt/hobbit/client/tmp/hobbit_iostatcpu.edsvideo.21811 
>   hobbit 15431 15421  0 10:51:57 pts/2    0:00 -ksh
>   hobbit 21805 21783  0 11:46:21 ?        0:00 /usr/bin/sar -A -o 
> /opt/hobbit/client/tmp/sar.21783 300 1
>   hobbit 21786 21780  0                   0:00 <defunct>
>   hobbit 21782 21780  0                   0:01 <defunct>
>
> Few minutes later, I got this :
>
> $ ps -efd|grep hobbit
>   hobbit 21780     1  0 11:46:20 ?        0:00 
> /opt/hobbit/client/bin/hobbitlaunch 
> --config=/opt/hobbit/client/etc/clientlaunc
>   hobbit 21785 21780  0                   0:00 <defunct>
>   hobbit 21781 21780  0                   0:00 <defunct>
>   hobbit 21784 21780  0                   0:00 <defunct>
>   hobbit 21787 21780  0                   0:00 <defunct>
>   hobbit 21783 21780  0                   0:00 <defunct>
>   hobbit 15431 15421  0 10:51:57 pts/2    0:00 -ksh
>   hobbit 21786 21780  0                   0:00 <defunct>
>   hobbit 21782 21780  0                   0:01 <defunct>
>
> Thus, nothing is graphed on hobbit server and all tests are purple. I got 
> several servers with solaris 7 and everything goes fine on these, so I'm 
> just posting here to know if someone got this problem too.
> Sincerly,
>
> Thomas