Unsolved
This post is more than 5 years old
202 Posts
0
835
Networker 9 cloning optimalization
Dear NW experts!
We upgraded to NW9 from 8.2.3. We like it, and we migrated old clone jobs to the new networker.
The clone group isnt modified we leaved the default settings.
The environment is the following:
- DD vtl with many drives (20) (target session=1)
- one physical tape lib with lto5 drives (2) (target session=12,max sessions=32)
- multiple clone jobs with default settings (paralellism=0)
When we start the cloning, many sub clone tasks generated, but we have only one stream which is writing the drive.
Do you have any recommendations or suggestions? Thanks in advance.
BR:
Pa
bingo.1
2.4K Posts
1
November 15th, 2016 00:00
Your statement does not really reflect completely what you see. Is there any problem with the throughput?
If these streams are writing with full speed then this is all you can achieve.
Multiplexing streams to tape is of course possible but with respect to recoveries, you want to avoid that.
Why? - because you most likely will never recover all these streams at the same time. However it could be that such single stream save set might have a caveat during recovery ... if the client cannot move the read data fast enough. But this is something which you have to test/optimize in your specific environment.
Jobs could be started and still be queued - the jobs db should be able to tell you that.
If you really want to speed up cloning from DD to tape, may I suggest that you install more tape drives, if possible.