1 Rookie
•
3 Posts
0
1586
May 6th, 2022 04:00
Attempting to restore data from tape clones
I have a backup policy for VMware virtual machines which performs backup to primary backup storage (Data Domain) and a clone action which clones the data to physical tape library with longer retention. This is running fine, but when trying to restore a VM from a tape I cannot see any progress being made.
The recovery log contains only the following:
155752:nsrvproxy_recover: Unable to find a usable device for recovery of saveset 407484030: Unable to find enabled Data Domain device for volume '462APSL6'. An attempt will be made to perform automatic clone resurrection
165713:nsrvproxy_recover: Attempting to automatically resurrect clone 407484030 with 7 days retention to DD volume in pool Data Domain Default Clone
Also in alerts tab I can see the following warning:
Waiting for 1 writable volume to backup clone pool 'Data Domain Clone' disk(s) or tape(s) on networker for client(s) vcenter
I am not sure what is wrong, I assume that it tries to use the Data Domain as a staging area for the restore but something seems off. I am using "Data Domain Default" for primary backup job and a custom "Clone to Tape" media pool for the clone job.
Networker is installed on Windows Server 2019 - NetWorker 19.4.0.2.Build.127
bingo.1
2.4K Posts
1
May 9th, 2022 07:00
Of course you can use a DD for multiple backup devices.I do this all the time.
I do not remember if each will create its own M-Tree. My personal guess is that there is just one M-Tree per 'DD-Client', means NW server/snode.
bingo.1
2.4K Posts
0
May 6th, 2022 05:00
The NVP solution has to work with a DD as backup device. Consequently, if the save set is not available there any longer, you must first clone it to such a volume. As it states, NW will automatically clone the save set to a media labeled for the DD Default Clone pool.
You have the choice:
- you can manually clone the save set to the DD Clone pool before you start the recovery (yes, it is possible to clone to a backup type pool) or
- you just add a new DD device and label it for the DD Default Clone Pool. In this case NW will automatically delete this instance after 7 days.
mijak017
1 Rookie
•
3 Posts
0
May 9th, 2022 01:00
Hello, thank you for the response.
Just to clarify, when you say "add a new DD device and label it for the DD Default Clone Pool" do you mean to add a separate DataDomain array or can I use the same DataDomain array that I am already using as a backup pool? Can I create another MTree and use it for the DD Default Clone Pool?
Thanks!
mijak017
1 Rookie
•
3 Posts
0
May 12th, 2022 01:00
I managed to solve this by creating another device and assigning it to the Data Domain Default Clone Pool. I also had to create a separate M-Tree in the DD filesystem.
Thank you for your help.
barry_beckers
393 Posts
0
May 15th, 2022 08:00
You can create a separate mtree (or better it is called storageunit when using ddboost) but the default when using the device wizard would be that for each pool you'd create a subfolder underneath the default mtree/storageunit which is named as the shortname of the backup server.
This is a change from the initial approach from the lower nw7 versions with initial support for ddboost that was using an mtree per device. With nw7.6 (if I remember correctly) the big change came that each backup server would create devices underneath the mtree and not as separate mtree's. This as a dd was limited to 99 mtree's and when having more than 20 simulataneously used mtree's or so, it would experience a performance hit. This was stated as being the case for years.
So at the time we made nw7.6 mandatory, so not to have to migrate to the new approach at the time.
Nw even supports sharing one and the same boost volume with multiple devices (from nw8.2 or so onwards), even when using the same nw storagenode. This to be able to go beyond the max number of 60 sessions for a ddboost device, so for certain large environments we have one and the same volume shared with 4 devices on the same nw sn. So with the nw device wizard you'd choose the same subfolder of a pool and give it another device name.
So no need to use a separate mtree, but rather use the nw device wizard and enter uset boost user name and password and the create a new folder in the next step. When sharing a device, choose an existing folder...
bingo.1
2.4K Posts
0
June 5th, 2022 15:00
Today I had some time to dig into this issue and I found the following:
- A vproxy backup & restore must transfer the data to/from a DD.
- If there is only a cloned copy on a non-DD volume, NW might support you in recovering the save set to a DD media at the begin of the recover process.
- By default, NW will use a media from the 'Data Domain Default Clone Pool'. You may define another pool if it has the attributes 'Media type required = Data Domain' and 'Pool Type= Backup Clone'.
- As mentioned in the message, NW will apply a retention date of 7 days for this save set.
The strange thing is that NW uses a separate pool at all - it could simply recover/stage to the original backup pool and apply the 7 days retention to this instance ... as it is the case if you recover from a cloud media (ECS). No idea why it uses a new method vor vproxy recoveries.
However, if you run a vproxy 'File Level Recovery', NW will not even display the option to recover from a non-DD medial. At this point you are stuck.
Hope this helps.