Unable to umount (all processes killed) 2

[aix4me2]

While trying to umount a file system keep getting message "...device busy" even after eliminating all the processes with open files.

fuser and lsof show nothing.

Ran slibclean. No change.

Used ipcrm to remove all share memory segments, messsage queues, semaphores. No change.

Killed all user processes and non essential root processes. No change.

Ran svmon -S. Found a number of ...lib.a files that were mapped to the file system. Tried modifying some and got the message files are open!? So what is keeping these files opened and how to clean up without rebooting? These are some 3rd party application files and are not owned by root. Nothing to do with AIX system files.

Alex

 

[d3funct]

Have you done "fuser -cuxV"? the "x" will report on executable and loadable objects in addition to the standard fuser output.

An infinite number of monkey typing at an infinite number of keyboards will eventually populate the internet.

 

[tekken97]

The following shows what I did before umount the filesystem.
1. # cd /
2. # fuser -xuck /app
3. # umount /app
umount: 0506-349 Cannot unmount /dev/fslv01: The requested resource is busy.
4. # fuser -cu /app
nothing processe is found.
The vg is online only on this machine.

>_<

 

[mrn]

I know it's really obvious but you weren't in the filesystem while trying to unmount it (My fav trick)

E.g

cd /fs1
...do something
...have lunch
umount fs1

I lost count of the number of times I've done this....

--
| Mike Nixon
| Unix Admin
|
----------------------------

 

[piemaster]

Did you ever discover the cause of this problem? I also have a similar problem witha file system which I cannot unmount which is also losing free space. I believe that in my case the problem of loss of free space is connected with the inability to unmount the file system.

 

[ap1605]

Had you solved the problem?

We have a problem same as you now.
On a HACMP 4.4.1 cluster , AIX oslevel 5.1 ML4
A filesystem cannot be umounted after i installed Ca7 agent.
A product from Computer Assosiates
I have a problem when i try to stop the cluster and Ca7 agent was running. When i don´t start the Ca7 agent and stop the cluster. I do not have a problem.

Did someone ever discover the cause of this problem?

Kind regards

Ps: I have tried fuser, lsof, ipcs -s, ipcrm

 

[Chapter11]

I've had the same core problem show up a few times myself, in my case the open files on the filesystem I was trying to umount were logfiles that syslogd were logging to.

Specifically, Legato Networker added some lines to syslog.conf that log to /nsr/logs so I could never umount /nsr and it took me weeks to figure out why.

 

[noideatoo]

We had the same issue (with a HACMP style script) and found a large number of defunct / zombie processes that IBM support found where somehow holding open files, but they where not showing in lsof etc...

The script being used was trying to do a nice kill (e.g. -15) to close any application processes still around then followed by -9 soon afterwards, causing the process to go defunct

No choice but to reboot in the end !!

The script has now been changed that after the application has tried it's own close to kill -9 before making any request to umount.

 

[baggetta]

Same thing here, HACMP 4.5 on AIX 5.1 running a Progress database. Users f/s was unable to unmount, tried everything to unmount but nothing worked. IBM says is a Progress problem, and Progress states its an IBM problem...so it looks like were in a circle pointing fingures and no one wants to fix the problem. When we first installed HA 4 years back, it seemed to work fine on failover and takeovers. Since we installed Java components and other Progress components that support java, i think in the emails above someone was talking about the .lib files being open .so/.a files....next time i need to shutdown i'll have another look...