Removing LV 1

[tristan308]

Need to unmount a file system to delete it. However when trying to do this get a resource busy on a LV that needs to go also. Tried to delete LV but get LV must be closed. If the LV contains a file systems the unmount command will close. Don't know where to go from here.

 

[Yegolev]

Are you saying the fs will not umount? Like so:

umount: 0506-349 Cannot unmount /dev/emcd09: The requested resource is busy.

You'll have to kill processes and cd to some other dir.

o14777@xp0:/utc/home/o14777>lsvg -l datavg
datavg:
LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
emcd09 jfs 457 457 1 open/syncd /utc
loglv01 jfslog 1 1 1 open/syncd N/A
btc jfs 192 192 1 open/syncd /btc

When umount succeeds LV STATE will be "closed/syncd".

hth

 

[bi]

You can't unmount a filesystem if it is in use. It is in use if you (or someone else) has cd'd into it or if there is a process running from that filesystem.

Make sure no one has cd'd into the filesystem. If everyone is out and you still cannot cd into it, try to determine what the process is that is running from the filesystem. You can run fuser -fu /<filesystem> to see who is running the process and what the pid is. Then kill the process if possible.

In addition, if you have a two-tier filesystem layout (for example, /main, /main/data, /main/index, /main/rbs), you aren't going to be able to unmount /main until /main/data, /main/index, and /main/rbs are unmounted.

 

[tristan308]

No one is in the filesystem, nor was I. Everything under that filesystem is now gone, the only thing left to get rid of is /whatever/temp. The templv is still in use whenever I try to delete it. The volume group is gone however that it was in Tempvg. Could that have something to do with it?

 

[Yegolev]

I'm not sure that I understand what you are saying. If you removed the vg then your problem is solved since the lv will also be gone. Perhaps your fs is still listed in /etc/filesystems? Perhaps your ODM is corrupted? Some output would probably help.

#lsvg
#lsvg -l $vgname
#lsvgfs $vgname

Also the command you are using to get this info and the output.

 

[bi]

I have had a similar problem once, when I was having massive hardware problems (caused by a poorly cooled computer room, not IBM equipment!). The VG wasn't known to the system, but the LVs and FSs were still known to the system.

I think, if I remember correctly, that I did an importvg and it seemed to get the ODM back in sync. I was astonished that it worked. I only tried it because at that point I had nothing to lose.

 

[MoshiachNow]

Things to try:

1.fuser -k /dev/<LV>
This will kill any process using the LV/FS

2.Export/importvg

3.syncvg -p <hdiskname>
This will sinc ODM with the Disk info.
&quot;Long live king Moshiach !&quot;