Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Chriss Miller on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Media Pool Confusion.

Status
Not open for further replies.

L4ts

Technical User
Aug 17, 2004
29
GB
I have a question regarding Medial Pool. We use Storagetek L20 library with LTO drives and Arcserve 11 to backup 100 or so servers. I am quite new to Arcserve so please forgive any misunderstanding.
When I click on Job Status there is around 10 backup jobs and each of these jobs are named as rack1 - rack10 as each one contains around 10 servers. Is this normal?
In devices I have 2 groups. GROUP0 and GROUP1. GROUP0 contains all the slots and GROUP1 contains 3 different tapes for each day! I currently change this manually everyday without using Media Pool so I am missing out big time.
If I place 15 tapes into the library and want to use Rotation Scheme to backup fully Monday to Friday into Media Pool, do I need to create any more groups? How do I assign each media job to a group when I goto Modify and select Destination or is this done when I rotation scheme.
Thanks in advance.
Lats
 
Lats,

when you use rotation jobs you should not configure groups at all. Put all slots in one group and let the rotation scheme(s) do the tape handling. The rules for tape handling by the rotation schemes can be setup by yourself.

regards

 
Cyklops,

Thanks for your reply. Thats one thing out of the way now. My next question is, when modifying a backup job, do I select anything in the destination tab or again is this all controlled from what I configure in Schedule when doing media pooling.
Also, when I tried a test, it used all the tapes in the group when it should have used only about 3!
Thanks again.
 
You just select the Group in the Destination and let the job select the media that is needed for the job...
 
Yes, but if all the media I am using is new (<blank>), why does it use 4-5 tapes when it should be using just 3!?
 
Yes, or atleast trying to.
 
Ooops. Sorry. I was was thinking of multistreaming, multiplexing. I hate these terms! Anyways, I was going to say that multistreaming can cause you to use more tapes that usual since multiple drives are backing up at the same time. Multiplexing shouldn't cause more tape usage. This doesn't apply.

If you have 10 jobs, are they all running sequentially or at the same time? Do you have more than one drive in your library?

Running two jobs at the same time to two different drives will mean they will use a tape each. Let them both span once each and now four tapes are used!

If you only have one drive then the jobs will run seqeuentially. If they all append to the tape then you will only use as many tapes as it takes to store all the data.

You keep mentioning three tapes. Why three? How did you come up with that count?

What is the size of each of the tapes? How much data gets backed up to the 4-5 tapes?
 
Ok, start time is 6pm and finish time is 6am. We use both drives in the library for backing up around ten jobs.
Total data size of all jobs is around 550GB and LTO tapes we use can hold 200GB compressed.
What we like to do is get all these servers backed up onto tapes daily using media pooling so all daily backups are automated.

ACiDelic, are you saying that becuase we have two drives in the library, we should also have two groups?
 
There may be two possibilities:

1. 200GB compressed means 100GB native. In a perfect world all the data you backup would compress 2:1, but we don't live in a perfect world. So, some data will not compress to a smaller size. Jpeg and zip files are good examples. These file formats are already compressed an won't compress much if any. So your tape usage would lean more towards 100GB (or less once you count in the meta information for each file). If you're backing up text files you would get at least 200GB on each tape.

I suspect that the data you are backing up is not compressing very well. Check to see how much data is really on each tape. See if all the tapes are indeed full.

2. I'm not sure if ARCserve is smart enough to know that one of the drives are free. If it is then the jobs may be running simultenously, each to a different drive. If all the jobs append to the media in the drive then there is a good chance that each drive will span tapes. If each spans once then you are no longer using 2 tapes, but 4. Let one of the drives span again and now you are up to 5.

This is possible, but since I only have a single drive library I can't say whether or not it is true.
 
According to 2:
ARCserve uses all available tape drives. If a tape library holds two drives and you schedule three jobs then two jobs will kick off immediately and the ARCserve job queue will hold the third job until one of the two jobs is finished and a tape drive becomes available.
If you only schedule one job then only one tape drive is used. ARCserve v11.1 has an enterprise option for submitting just one 'master' job that will be split into a number of 'slave' jobs and use all available tape drives for this 'master job'. This is called 'multistreaming'

regards
 
Sorry ACiDelic,

didn't read your prior posts to this thread so my comments regarding multistreaming are a bit overdone

regards
 
Upgrading v11 options to v11.1 is free of charge, but the new Enterprise Option has to be purchased seperately.

 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top