| Title: | DIGITAL UNIX (FORMERLY KNOWN AS DEC OSF/1) |
| Notice: | Welcome to the Digital UNIX Conference |
| Moderator: | SMURF::DENHAM |
| Created: | Thu Mar 16 1995 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 10068 |
| Total number of notes: | 35879 |
Hi, My customer has a rare problem, again : DUNIX 3.2F We has an application, when it's running about 40 hours all process are on state DEFUNCT. He saw it with the ps au command. He must change the application to another machine and the system process are O.K. He didn't need to reboot it. Can anybody say me why the system can to change the process all to defunct ?? Thanks for all, Arancha
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 9612.1 | CSC32::B_GRUBBS | Thu Apr 24 1997 12:47 | 17 | ||
I saw one just the other day like this. ps showed all the
process names a <defunct?, but things still seemed to be
working ok, albeit slowly.
vmstat -M showed less than 1.2Mb memory free and 166Mb used.
One single bucket accounted for a good portion of what was
used.
Polycenter Data Collector was the culprit. The MIPC type at
the bottom of the output showed a VERY large amount of memory
being used.
In their case there was a patch for PDC on DUNIX 3.2c to fix
the problem.
--Bert
| |||||
| 9612.2 | SMURF::DENHAM | Digital UNIX Kernel | Thu Apr 24 1997 14:19 | 2 | |
Yep, it's a memory leak/shortage. Starting with vmstat -M and
looking for huge hunks of allocated memory is a good idea.
| |||||
| 9612.3 | vmstat -M command??? How see the information | MDR01::ARANCHA | MCS Madrid | Wed May 07 1997 10:42 | 12 |
Hi,
Thanks for the reply. But why do I interprete or analize the output of
vmstat -M command?
Are there any utility to determinate which is the application or
process that it's doing the all defunt process?
You say: "Polycenter Data Collector was the culprit" .
How Must I do to detect the culprit process , if I can see nothing??
Thanks for all.
Arancha
| |||||