Cancel sessions/staging

[peter0201]

Hello,

I'm searching a solution (command) to stop savegroup running on Networker 7.2 for Windows. I had twice the problem that a FULL backup never ends (savegrp is already running) and it's not possible to stop in nwadmin. I do not like to restart the service or kill the savegrp command.

Also I like to cancel a staging session e.g. from Disk to Tape, when I need to restore urgently from the staged device, and I couldn't wait until staging will be finished.

Thank you for your help.

Peter

 

[1018DADA]

To my knowlede if you cannot identify the individual Networker process in task manger to stop, the only other way might be restarting Networker server services and if other backups are running, that could be detrimental.
Perhaps if you have large enviroment isolate problem clients in different groups\storage node and stagger start times to achive what you want.
Certain Disk type device create a RO Device along with RW and restore should be able to read while staging is ongoing, if it does not you may need to look into what type of disk device have you configured.

 

[peter0201]

Thank you for your help. We are using an advanged file system (Windows volume) and for this kind of device we can only read one time for each pool. So staging and reading at the same time is unfortunately not possible. I just thought about an networker command to view and edit (e.g. cancel) all networker tasks.

 

[605]

You can definitively start multiple read processes from a NetWorker device, especially from (Advanced) File Type Devices. However, the automatic processes will not allow this - you have to start multiple recover/clone/stage commands from the command line.

For an AFTD you can even can read from both instances ...
- the physical save set (R/W device) and from
- the virtual save set (R/O device)
at the same time to different pools.

 

[peter0201]

I tried using the recover command at the command line during the staging process of an Advanced File System, and it is not possible. The recover process will wait until staging will be finished. Can you explain how to recover a single file or directory with command line that it will work?

Thanx,
Peter

 

[605]

O.K. - i should have been a bit more precise. Do the following:

- Create a backup of several 100 MB on an AFTD
- Create clone pool A, for example on a 'normal' FTD
- Create clone pool B (FTD)
- Label media to both pools
- Run mminfo to get the save set info as follows:
C:\nsr\bin>mminfo -av -r "name, ssid, cloneid"
name ssid clone id
H:\TESTFILE_RANDOM 4165211728 1128535632
H:\TESTFILE_RANDOM 4165211728 1128535631

C:\nsr\bin>
- Open two Command windows and run these commands, one per
window:

nsrclone -b A -S 4165211728/1128535631

nsrclone -b A -S 4165211728/1128535632

NW will actually copy both streams at the same time, proving that you can create multiple reads at the same time, even create simultanous reads on the same save set.


However, i have to admit that this obviously does not work with staging commands (nsrstage) as this command locks the save set. This makes sense as the source will be deleted at the and of the stage process.

 

[emcaixlegato]

My experience shows that Peter0201 is correct. Only one process can read from an Advanced File System device. For example, I have a backup disk pool "disk_backup_1". We backup multiple clients to this backup pool and 24x7, we have an nsrclone process running to clone the data to tape. If someone needs to recover something from the "disk_backup_1" pool, we have to stop the nsrclone process to allow the recover. Also, only one nsrclone process can be run from the disk backup pool at a time. The second will just hang waiting for the first to complete. Additionally, killing the cloning process can produce aborted clones, which can only be found using the clflags option for the mminfo command. These appear as copies, but they are aborted copies, which aren't very useful for restores.

 

[605]

In the meantime i learnt that for whatever reason the recover process does not behave exactly the same as the staging process - while you can run multiple recoveries, you can not run simulteanous staging processes. To my knowledge there is currently no solution to this problem.