This content has been marked as final. Show 10 replies
i have seen this several times. I guess you dropped this disk from a disk group successfully and now want to delete the disk header. This happens from time to time
and can be resolved by dismounting the diskgroup the disk originates from. Itmight sound strange but ASM seems to keep the device open even after dropping the disk.
Dismounting the disk group seems to close the device handle and you can delete the asm header successfully.
My blog: http://ronnyegner.wordpress.com
Edited by: Ronny Egner on Oct 9, 2009 7:32 AM
actually, some investigation around metalink brought up bug 7225720 which is exactly the problem i have.
Just going through testing it out. However, it looks as though a workaround would be to stop all DB instances
and ASM instamces and then remove the disk - not so good on a 24x7 prod system though is it :)
for your information: At least in my environment the 11g Release 2 ASM shows this behavior as well. Not so good......
We are hitting this problem in 11.2 ASM; the patch was supposed to be in the base release. We were able to successfully drop the disk (here AMS_102) and cannot delete the disk unless we stop the ASM instance.
Anyone else hitting this problem in 11.2?
root(/etc)# oracleasm deletedisk AMS_102
Clearing disk header: failed
Unable to clear disk "AMS_102"
Linux 2.6.18-164.el5 #1 SMP Thu Sep 3 04:15:13 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux
I have the same 18.104.22.168 RAC.
/etc/init.d/oracleasm deletedisk SYS1
Removing ASM disk "SYS1": [FAILED]
All nodes see this DISK after scan
So I disable asm, stop it on all nodes, run deletedisk doesn't help.
and I can delete it only from node I created it on.
Same problem with 22.214.171.124 on x86-64 RH Linux 5.4.
Had to :
1. shutdown CRS on all nodes of the cluster (# ./crsctl stop crs)
2. disable CRS from starting on reboot (# ./crsctl disable crs)
3. Reboot all the servers (# shutdown -r now)
4. After reboot, stop ASMLIB (# service oracleasm stop)
5. Remove the disk using physical name rather than logical name (service oracleasm deletedisk /dev/mapper/mpath01p1)
6. Restart ASMLIB (# service oracleasm start)
7. Enable CRS on re-boot (# ./crsctl disable crs)
8. Reboot the servers
I ran into this issue with
oracleasm version 2.1.3
ASM Oracle Database 11g Enterprise Edition Release 126.96.36.199.0
This is what I did :-
$ fuser /dev/sdes1
$ dd if=/dev/zero of=/dev/sdes1 bs=1024 count=100
100+0 records in
100+0 records out
102400 bytes (102 kB) copied, 0.000324 seconds, 316 MB/s
$ /etc/init.d/oracleasm deletedisk RECO84
Removing ASM disk "RECO84": [ OK ]
And then on all nodes.
$ oracleasm scandisks
This works for me.
I take the dd approach as well as it seems to be the best option most of the time. Then, if I need to pull the disk back into ASM, I just run oracleasm createdisk <name> <path>.
ncalappu = awesome
That fixed it for me.