Unsolved
This post is more than 5 years old
33 Posts
0
1098
Cloning Optimisation
Hi,
This is sort of like the other thread about nsrclone recently, but I thought I'd start a new one rather than hijacking someone else thread.
Our Networker 7.2.2 setup has one Networker server, and a Networker Storage Node. Each server has 4 Virtual CDL based Tape drives which receive the backup sessions from the backup clients. Each server also has 2 real LTO-3 Tape drives from an autochanger.
We have about 6 groups which run each day. The reasons for so many groups is partly because some backups need to occur at different times, and partly because i've tried to group the slow backups into one group and the faster ones into another so auto-cloning can start sooner (for the faster group).
There are a few problems (well inefficiencies) with this.
1) I can't seem to get the Storage Node to ever do any cloning. Even though I fill in the 'Clone Storage Node' property of the networker server client object.
2) A lot of the time one group (say the slow group) will be using the tape which the fast group is trying to clone. So the cloning is haulted. I realize this is by design, but I am wondering how everyone else handles this? I don't really want to create pools for each group, and I think we need to do saveset cloning rather than volume cloning for many reasons (virtual backup tape/clone tape sizes are different, virtual tape could be appended to the next day, etc.)
3) Sometimes (once every couple of weeks) two groups stop progressing (backing up or cloning), seemingly because they each need a tape the other has (a deadlock of sorts).
I realize I may be asking the impossible, but any help would be appreciated.
This is sort of like the other thread about nsrclone recently, but I thought I'd start a new one rather than hijacking someone else thread.
Our Networker 7.2.2 setup has one Networker server, and a Networker Storage Node. Each server has 4 Virtual CDL based Tape drives which receive the backup sessions from the backup clients. Each server also has 2 real LTO-3 Tape drives from an autochanger.
We have about 6 groups which run each day. The reasons for so many groups is partly because some backups need to occur at different times, and partly because i've tried to group the slow backups into one group and the faster ones into another so auto-cloning can start sooner (for the faster group).
There are a few problems (well inefficiencies) with this.
1) I can't seem to get the Storage Node to ever do any cloning. Even though I fill in the 'Clone Storage Node' property of the networker server client object.
2) A lot of the time one group (say the slow group) will be using the tape which the fast group is trying to clone. So the cloning is haulted. I realize this is by design, but I am wondering how everyone else handles this? I don't really want to create pools for each group, and I think we need to do saveset cloning rather than volume cloning for many reasons (virtual backup tape/clone tape sizes are different, virtual tape could be appended to the next day, etc.)
3) Sometimes (once every couple of weeks) two groups stop progressing (backing up or cloning), seemingly because they each need a tape the other has (a deadlock of sorts).
I realize I may be asking the impossible, but any help would be appreciated.
mllegato
20 Posts
0
August 7th, 2007 12:00
I am not sure how you set up your cloning, but for auto clone all you need to do is enable cloning for a group in question and specify the cloning pool. I run all my clones from this option, except for one group which is unreliable due to some clients holding it running forever. It's done by saveset cloning later in the day
2,3
That was exactly what I did when separating slow and fast clients. I has two jobs competing for the same tape, and I did create another pool for a slow group. Since then - no deadlocks. I suppose you will have to go this route if you want to avoid it
Hope it helped
ble1
14.3K Posts
0
August 8th, 2007 09:00
bc9am
33 Posts
0
August 12th, 2007 16:00
ble1
14.3K Posts
0
August 13th, 2007 00:00
danag2
20 Posts
0
December 27th, 2011 11:00
You need to segment the backup jobs and the clone jobs by Pool and assign them devices so that they never interfere with each other and you can then assign the Storage Node associated with the device you want to use for the Clone job. If the Storage Node doesn't have access the the device you wanted to use then obviously it can use it for cloning.
coganb
736 Posts
0
December 29th, 2011 06:00
Hi Charlie,
Just in case you don't know it, there is a NetWorker Cloning Integration document which details all of the available options for cloning and might help you find the best solution.
-Bobby