Scheduled Cloning / Volume Cloninng Saveset limitation

[tom303r]

Does anybody knows something about a saveset limitation by the NMC GUI Cloning?

If I start a Scheduled Cloning job where I want to clone all Backups of one pool (10 years of Backup Data), the job will not come to an end and nothing will be cloned to tape. The same result if I start nsrclone from the cli.

If I start NMC Volume Cloning and select 6 volumes (a lot of savesets) the job fails directly. If I select 2 of the 6 Tapes to clone, the job starts normal.

 

[605]

Excuse me, but a bit more beef around the bone would be helpful.

" ...the job will not come to an end and nothing will be cloned to tape."
This sounds to me like the server is just waiting for an appropriate media. Are there any alerts?

What's the command and the response/error message(s)?

 

[tom303r]

I have started following nsrclone job on CLI:

nsrclone -v -S -D 0 -s BackupServer -J BackupServer -b targetpool -d BackupServer -F -c Client1 -c Client2 -B pool1 -B pool2 -t "Mon Aug 24 10:49:56 1992" -e "Fri Aug 24 10:00:00 2012"

After about 80 hours I stopped the job because nothing was cloned and there was no mount request and no error/alert messages.

the only output is like:
...
79360:nsrclone: Failed to find a clone for ssid 4082911462 in source pool specified, skipping
32358:nsrclone: skipping aborted save set 1109359204 cloneid 1296408252
Device location of volume 000041 is unknown. Checking recover storage node of Client host client1.
Device location of volume 000041 is being set to host BackupServer

but there is no error messege like waiting of a volume or something else.

there are round about 40.000 savesets to clone.

 

[605]

The issue is the parameter sequence. With your command NW will never find a save set to clone because ...
"-S Causes nsrclone to treat subsequent command line parameters as save set identifiers, not volume names. Save set identifiers are unsigned numbers. ..."

"-e <retention_date" unfortunately is also invalid. Use "-y <retention_date>" instead.
And what does "-D 0" stand for? - according to the NW 762 manpages, this is unknown parameter.


May i suggest the following procedure:

1. Prepare the save set list for example using
mminfo -q "client=client1,pool=pool1,savetime>=start_time,savetime<end_time" -r "ssid" [-ot] >> ssids_to_be_cloned
Of course you must run the appropriate query (multiple times).
For verification, you can also report more parameters and prepare the list in an Excel sheet.

2. Delete the headlines from the output file.

3. Run the clone command as follows:
nsrclone -v [-s BackupServer -J BackupServer -d BackupServer] -F -b dest_pool -y mm/dd/yyyy -S -f ssids_to_be_cloned [> logfile 2>&1]

Fortunately, these old save sets will not be so large. What did you use to create them - NW 3.0?
Good luck.

 

[tom303r]

We use NW 7.6.4

I copied the parameter sequence from the task manager output when I start the Scheduled Cloning GUI task.
Should the parameter sequence not running at CLI too? Networker don't complain about wrong command options.

-e "end time" should be ok

-D 0 I think it is the GUI Option don't stop cloning on error

Your workaround suggestion should be run. I have done it for the daily cloning before Scheduled Cloning came out.

But my target is to do the job at the NMC GUI.

So my workaround was GUI Volume Cloning. But If I select more than 2 - 3 Volmes to clone the Cloning Task terminates immediately.
If I select 2 Volumes to clone the cloning runs normal.

So my conjecture was that there is a quantity restriction of the Savesets to clone.
And that is my question. Is there a quantity restriction of the Savesets to clone?

 

[605]

We use NW 7.6.4
You created NW backups 20 years ago with NW 7.6.4? - You are a liar ;-))
But i love to hear that you have been using NW for such a long time.

I copied the parameter sequence from the task manager output when I start the Scheduled Cloning GUI task.
Should the parameter sequence not running at CLI too? Networker don't complain about wrong command options.
NW will not always complain about wrong options - i have just seen it today again.
Of course you can still test/verify other options/sequences.

-e "end time" should be ok
Probably - it is just not listed as parameter list for nsrclone which is the performing command.

-D 0 I think it is the GUI Option don't stop cloning on error
"-F" is the parameter in charge of this option. Have a look at the manpage.

Your workaround suggestion should be run. I have done it for the daily cloning before Scheduled Cloning came out.

But my target is to do the job at the NMC GUI.
Seriously - forget it. I have never used it so far.
Of course you can argue a lot about the pros and cons but always remember that the GUI is not as flexible as the CLI.
You can add verbosity and you can redirect the output to a logfile for verification.

So my workaround was GUI Volume Cloning. But If I select more than 2 - 3 Volmes to clone the Cloning Task terminates immediately.
If I select 2 Volumes to clone the cloning runs normal.
Sorry, i cannot comment this as i have never used it. But i do not see a reason.
Just run "nsrclone -v -F -b pool volume1 volume2 volume3" and watch the output.
However, there must be a message somewhere - at least the daemon.raw file should tell you more.

So my conjecture was that there is a quantity restriction of the Savesets to clone.
And that is my question. Is there a quantity restriction of the Savesets to clone?
Not any more to my knowledge. However, there was a restriction of 2000 save sets up to NW 7.6.1. We suffered a lot from this issue.
But when we run our monthly clones today (NW 7.6.3.5), we clone several thousand save sets each day. And 3 volumes may just contain some hundred save sets.

 

[tom303r]

Thank you very much for your help.

 

[tom303r]

C:\>nsrclone -v -F -b targetpool 000004
Obtaining media database information on server server1
Device location of volume 000004 is unknown. Checking recover storage node of Client host server2.
Device location of volume 000004 is being set to host null
Not enough space
5867:nsrclone: add_to_snodes: FALSE from loc_volid
39077:nsrclone: error, Cannot add saveset 3054255 in volume 3888158052 to the clone list

 

[605]

Hmm,

I suppose that with such very old save sets, you are cloning from a tape. It seems that for whatever reason the volume cannot be mounted automatically.

What happens, if you manually mount the media before you start the clone process?
This would also verify whether you even can mount the media in the device. This might be impossible if you have selected specific devices for the pool. However, in this case i would also expect better error messages.