cannot remove volume from AdvancedDisk pool

[Will_Restore]

Creating a new v7.6.0.4 system and put wrong volume name in AdvancedDisk pool.  After realizing the error, successfully added the correct volume.  There are no images on the wrong volume.  Can't seem to delete the it however:

  DSM has encounted an invalid state for the following object: /wrongname

Tried removing the Disk Pool and start over, but there are now some images on the correct volume, and expiring them doesn't seem to help either.

Searching prior posts & only way out is to get custom script from support?   Seems like a bad bug to me!  Why have -deletedv switch if it doesn't work?

[mph999]

I'll try and  have a look at this tomorrow, will need a case really, and if you can log this ask whoever takes it that I am aware of the issue and to contact me (Martin Holt).  Not, I have a high priority case tomorrow, so it might be a colleague, unless you are happy to wait for me. 

Well need nbsu -c -t output, nbdb_unload output, nbdb_backup -online output and the netbackup/logs/nbdb log @ verbose 5 showing an example of trying to delete the diskpool. 

[INT_RND · Accepted solution]

According to a guy named "Will Restore" you should run

nbdelete -allvolumes -force

https://www-secure.symantec.com/connect/forums/cant-delete-disk-pool-0

J/K ;-D

Ensure that no policies are using the diskpool. Ensure that the pool is not part of a storage unit. You cannot delete a pool that is "in use".

If it's in an invalid state I would recommend putting the disk pool in a down state:

nbdevconfig -changestate -state DOWN

[mph999 · Accepted solution]

OK, not sure of the exact procedure you followed, but I did the following:

Created a diskpool (testdisk) with 1 disk volume (/disk3p3)
Created another diskpool (testdisk2) with one disk volume (/disk3p4)

Ran a backup to testdisk
Merged the two together so that I was left with only one disk pool, testdisk, containing two volumes.

/disk3p3
/disk3p4

Ran another backup (mainly to check it was working)

Expired all backups (so no fragments on either of the disks)

Deleted the storage unit

Set the diskpool DOWN

Attempted to delete /disk3p4

root@womble disk3p3 $ nbdevconfig -deletedv -stype AdvancedDisk -dp testdisk -dv /disk3p4
DSM has encountered an invalid state for the following object: /disk3p4

EMM log showed

12/18/14 13:03:29.642 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 5 [MachineObject::MachineObject]  - MachineName womble Type 3
12/18/14 13:03:29.642 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 3 [MachineObject::Init] MachineLearning = <1>
12/18/14 13:03:29.642 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 4 [MachineObject::Exists] Looking for machine record with Name <womble> Type <3> checkCluster <TRUE>
12/18/14 13:03:29.642 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 4 [MachineObject::getRecordByNameType] MachineName womble Type 3
12/18/14 13:03:29.651 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 5 [MachineObject::~MachineObject]
12/18/14 13:03:29.651 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 5 [MachineObject::MachineObject]  - MachineName womble Type 3
12/18/14 13:03:29.651 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 3 [MachineObject::Init] MachineLearning = <1>
12/18/14 13:03:29.651 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 4 [MachineObject::Exists] Looking for machine record with Name <womble> Type <3> checkCluster <TRUE>
12/18/14 13:03:29.652 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 4 [MachineObject::getRecordByNameType] MachineName womble Type 3
12/18/14 13:03:29.659 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 2 [MachineConfigImpl::helperGetConfigValue] machineKey = < 1000002 >, configName = < AUDIT >
12/18/14 13:03:29.661 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 4 [MachineConfigImpl::helperGetConfigValue] retval = < 0 >
12/18/14 13:03:29.662 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 5 [MachineObject::~MachineObject]
12/18/14 13:03:29.662 [Debug] NB 51216 nbemm 111 PID:7316 TID:54 File ID:111 [No context] 1 [MachineConfigImpl::~MachineConfigImpl()]  Exiting
12/18/14 13:03:29.678 [Application] NB 51216 dsm 178 PID:7316 TID:54 File ID:111 [No context] [Error] V-178-46 DSM has encountered an invalid state for the following object: /disk3p4
12/18/14 13:03:29.678 [Application] NB 51216 dsm 178 PID:7316 TID:54 File ID:111 [No context] [Error] V-178-2 DSM has encountered an illegal argument exception: deleteDiskVolume():3410 Argument: emm_dv->getState()

I was able to delete the complete diskvolume, and thus both disk pools.

I then tried again:

I created diskpool ad-disk_pool containing two volumes (disk3p3 and diskk3p4) in one go via the wizard.  No backups were run or STUs created.
I then set the diskpool as DOWN, and tried to delete one of the volumes.

root@womble  $ nbdevconfig -deletedv -stype AdvancedDisk -dp ad_disk_pool -dv /disk3p4
DSM has encountered an invalid state for the following object: /disk3p4

Seems to be same error in EMM

12/18/14 13:10:19.237 [Debug] NB 51216 nbemm 111 PID:7316 TID:61 File ID:111 [No context] 5 [MachineObject::~MachineObject]
12/18/14 13:10:19.237 [Debug] NB 51216 nbemm 111 PID:7316 TID:61 File ID:111 [No context] 1 [MachineConfigImpl::~MachineConfigImpl()]  Exiting
12/18/14 13:10:19.266 [Application] NB 51216 dsm 178 PID:7316 TID:61 File ID:111 [No context] [Error] V-178-46 DSM has encountered an invalid state for the following object: /disk3p4
12/18/14 13:10:19.266 [Application] NB 51216 dsm 178 PID:7316 TID:61 File ID:111 [No context] [Error] V-178-2 DSM has encountered an illegal argument exception: deleteDiskVolume():3410 Argument: emm_dv->getState()

Checked the diskpool :

root@womble disk3p3 $ nbdevquery -listdp -stype AdvancedDisk -dp ad_disk_pool -U
Disk Pool Name   : ad_disk_pool
Disk Pool Id     : ad_disk_pool
Disk Type        : AdvancedDisk
Status           : DOWN
<snip>

Checked the disk volume

root@womble disk3p3 $ nbdevquery -listdv -stype AdvancedDisk -U

Disk Pool Name      : ad_disk_pool
Disk Type           : AdvancedDisk
Disk Volume Name    : /disk3p3
Disk Media ID       : @aaaap
Total Capacity (GB) : 33.49
Free Space (GB)     : 33.18
Use%                : 0
Status              : UP
Flag                : ReadOnWrite
Flag                : AdminUp
Flag                : InternalUp
Num Read Mounts     : 0
Num Write Mounts    : 1
Cur Read Streams    : 0
Cur Write Streams   : 0
Num Repl Sources    : 0
Num Repl Targets    : 0

Disk Pool Name      : ad_disk_pool
Disk Type           : AdvancedDisk
Disk Volume Name    : /disk3p4
Disk Media ID       : @aaaaq
Total Capacity (GB) : 33.49
Free Space (GB)     : 33.46
Use%                : 0
Status              : UP
Flag                : ReadOnWrite
Flag                : AdminUp
Flag                : InternalUp
Num Read Mounts     : 0
Num Write Mounts    : 1
Cur Read Streams    : 0
Cur Write Streams   : 0
Num Repl Sources    : 0
Num Repl Targets    : 0

I then set the disk volumes dDOWN

root@womble disk3p3 $ nbdevconfig -changestate -stype AdvancedDisk -dp ad_disk_pool -dv /disk3p4 -state DOWN
successfully changed the state of disk volume

root@womble disk3p3 $ nbdevconfig -changestate -stype AdvancedDisk -dp ad_disk_pool -dv /disk3p3 -state DOWN
successfully changed the state of disk volume

I then tried again :

nbdevconfig -deletedv -stype AdvancedDisk -dp ad_disk_pool -dv /disk3p4

Disk volume was deleted successfully.

So, both DiskPool and DiskVolume have to be set down for the delete to work.

[Will_Restore]

Thanks.  I changed state to DOWN, deleted the associated SLP, storage unit, and disk pool, then recreated everything.  Seems an awful long way around. 

Yeah, I tried nbdelete but there was nothing more to delete. :)

[Will_Restore]

Just saw your post.  That must be the trick.

So, both DiskPool and DiskVolume have to be set down for the delete to work.

[mph999]

YEs, it doesn't work with only the disk pool down, you still get an error, though a slightly different one to what you experienced - thoigh in this state you can in fact delete 'everything', that is the disk pool and all volumes.

But yes, to delete just one volume from the pool, both the pool and volume have to be down.

[Will_Restore]

Created a dummy diskpool added and deleted volume to confirm I was missing the volume state before.

  -changestate  ... -dv ... -state DOWN

And the command syntax is very exact; doesn't like parameters out of order.