Active IQ Unified Manager Discussions

Problem with vfiler volumes on OnCommand 5.0 interface

f_duranti
7,100 Views

Hi, I've a strange problem with a vfiler. Some time ago I had to move some vfiler volume from one aggregate to another one. I copied the volumes (vfiler root volume and one data volume), destroyed vfiler, renamed old volume to volumename_old and new volume to the original volume name then I recreated the vfiler with vfiler create -r. At this point after being sure the vfiler was available and running correctly i deleted the "_old" volumes.

After some day I got to oncommand server and now I see that the vfiler on OnCommand server seems to be configured with the old volumes "_old" and the new volume are configured under the normal netapp storage (so seems to be on vfiler0).

At storage level all is correct and the correct volume are assigned to the vfiler but OnCommand seems to have the old configuration inside and it seems that there's nothing I can do so that it will get the new data for the vfiler.

Anyone has an idea of what I can do?

Thanks

Francesco

1 ACCEPTED SOLUTION

columbus_admin
7,100 Views

Francesco,

     If you have the object ID of the volumes, usually column 1 of a dfm volume list output, you can purge them.

Stop all the DFM services except the database service

Run dfm volume delete volume_id -f

Then restart your DFM services.

This will force DFM to remove it from the database.

- Scott

View solution in original post

9 REPLIES 9

f_duranti
7,100 Views

It seems that after some dfm vfiler delete/dfm vfiler add/restart of dfm/refresh of vfiler/storage from dfm interface all is now working correctly.

arunchak
7,100 Views

>>dfm vfiler delete/dfm vfiler add

   You need not delete/add in DFM. Whenever you change any object from ONTAP you need to refresh the particulat storage or vfiler from DFM and wait until the refresh monitors complete for the changes to get reflected in DFM.

To refresh the storage/vfiler use "dfm host discover <host ip / host id>"

To check the monitoring timestamp "dfm host diag <host ip / host id>"

-Arun

f_duranti
7,100 Views

Hi Arhun, playing around with dfm xxxxxx commands I've saw that for example with -a I see also old volumes deleted much time ago from older storage that don't exist anymore. There's some way to purge those older data?

Francesco

Inviato da iPad

Il giorno 14/dic/2011, alle ore 03:13, "arunchak" <xdl-communities@communities.netapp.com> ha scritto:

arunchak created the discussion

"Re: Problem with vfiler volumes on OnCommand 5.0 interface"

To view the discussion, visit: https://communities.netapp.com/message/69742#69742

>

arunchak
7,100 Views

HI,

Yes you can see the deleted items using –a option. I am not sure whether we have any internal option to remove the deleted item from database. However you have options to purge old jobs and events.

-Arun

From: c-xdl-communities

Sent: Wednesday, December 14, 2011 12:56 PM

To: Chakravarthy, Arun

Subject: You have been mentioned by Francesco Duranti in Re: Problem with vfiler volumes on OnCommand 5.0 interface in NetApp Community

NetApp Online Community <https://communities.netapp.com/index.jspa>

<http://media.netapp.com/images/divider-600x3.jpg>

You have been mentioned

by Francesco Duranti <https://communities.netapp.com/people/f.duranti> in Re: Problem with vfiler volumes on OnCommand 5.0 interface in NetApp Community - View Francesco Duranti's reference to you <https://communities.netapp.com/message/69767#69767>

pradeepl
7,100 Views

Hi Francesco,

You can see the deleted objects using the -a option of the CLI. In dfm detail <obj-id> ,you can see when the object got deleted.

After marking the object as deleted , in sometime the object details will be  purged from the database.

agireesh
7,100 Views

Hi Francesco,

If you delete a volume from dfm using "dfm volume delete <vol-name-or-id> " command then dfm stop monitoring that deleted object (volume) and dfm doesn't delete that

object details from database. If user delete the objects (volume) from storage system then dfm retain that object detail in it's database for some period of time (For volume it's 1 hour), After that period of time if monitor trigger then dfm will delete that object detail from it's database. If monitor run before that period (1hour in volume case) then that object will not be deleted from dfm database.

In your case, you have to wait for an hour and then run the run the monitor using "dfm host discover <host-name-or-id>".command.

Regards,

Gireesh

f_duranti
7,100 Views

The strange things is that I see object that was deleted months ago in dfm vfiler list -a and volume that was connected to a storage that don't exist anymore (from 1 year) in the dfm volume list -a.

columbus_admin
7,101 Views

Francesco,

     If you have the object ID of the volumes, usually column 1 of a dfm volume list output, you can purge them.

Stop all the DFM services except the database service

Run dfm volume delete volume_id -f

Then restart your DFM services.

This will force DFM to remove it from the database.

- Scott

f_duranti
7,100 Views

Thanks it worked and deleting the old host that was inside the db it also deleted all of his volume in cascade. In my case it worked with dfm host delete -f hostid (the -f before the id).

Public