Splitting it into two separate jobs is certainly the easiest way to achieve this otherwise it will be a mixture of savegrp parallelism and also trying to use the client priority setting - but not seen anyone do this within their environments (while I was a backup admin or working in support) not to say it does not work, but even this could have both clients backing up concurrently if parallelism is greater than 1. The other easy way it to send them to different pools, but this would also necessitate 2 groups.
If you want one client to backup prior to the other client you need to look at adding these clients to seperate groups and staggering these so the first client/group should finish backing up prior to the other one starting, possibly enforcing this using parallelism settings.
However, it seems to me like this is not an issue; you are saying that you are not happy that you are getting say two tapes which are 55% full instead of one tape full and one tape which is 10% full. That doesn't strike me as necessarily a problem. Also the saveset data may get split across two tapes but only when the tape being written too gets full - Networker will not write the same saveset data across two appendable tapes simultaneously.
create seperate "pools". you could have "Full_a" and "nonfull_a", along with "Full_b" , and "nonfull_b". They can both run at the same time, if you have it setup right, as each client can be forced to use a specific "pool" of tapes.
We do this with some of our big archives. We only want them using specific tapes, and the only data going on those tapes, are these archives.
masonb
445 Posts
0
December 10th, 2010 15:00
Davek,
Splitting it into two separate jobs is certainly the easiest way to achieve this otherwise it will be a mixture of savegrp parallelism and also trying to use the client priority setting - but not seen anyone do this within their environments (while I was a backup admin or working in support) not to say it does not work, but even this could have both clients backing up concurrently if parallelism is greater than 1. The other easy way it to send them to different pools, but this would also necessitate 2 groups.
Regards,
Bill Mason
DavidHampson-rY
294 Posts
0
December 10th, 2010 16:00
If you want one client to backup prior to the other client you need to look at adding these clients to seperate groups and staggering these so the first client/group should finish backing up prior to the other one starting, possibly enforcing this using parallelism settings.
However, it seems to me like this is not an issue; you are saying that you are not happy that you are getting say two tapes which are 55% full instead of one tape full and one tape which is 10% full. That doesn't strike me as necessarily a problem. Also the saveset data may get split across two tapes but only when the tape being written too gets full - Networker will not write the same saveset data across two appendable tapes simultaneously.
noladave
1 Rookie
•
92 Posts
0
December 15th, 2010 14:00
create seperate "pools". you could have "Full_a" and "nonfull_a", along with "Full_b" , and "nonfull_b". They can both run at the same time, if you have it setup right, as each client can be forced to use a specific "pool" of tapes.
We do this with some of our big archives. We only want them using specific tapes, and the only data going on those tapes, are these archives.