01-21-2014 12:25 AM
Hi all,
i don't understand why NBU use diffrent number of tapes for 2 copies. We have two Volume Pools with the same retention. We want to vault a copy to each pool. We now see that the images on one pool are written to one tape (Pool 2) and on the other pool NBU use 2 tape (Pool 1), see the following table:
B60 - Pool 1 | |||||
ROHKGPEX01_1356703223 | 15 | Mon Dec 26 22:00:23 CST 2022 | 2 | 1 | 681433 |
ROHKGPEX01_1356703222 | 8 | Mon Dec 26 22:00:22 CST 2022 | 2 | 1 | 25 |
ROHKGPEX01_1356703221 | 27 | Mon Dec 26 22:00:21 CST 2022 | 2 | 1 | 1206446 |
rohkgpfp01_1356717627 | 42726 | Tue Dec 27 02:00:27 CST 2022 | 2 | 1 | 5265685 |
rohkgpfp01_1356717626 | 7 | Tue Dec 27 02:00:26 CST 2022 | 2 | 1 | 632026 |
ROHKGPFP02_1356703235 | 7 | Mon Dec 26 22:00:35 CST 2022 | 2 | 1 | 643586 |
ROHKGPFP02_1356703234 | 39793 | Mon Dec 26 22:00:34 CST 2022 | 2 | 1 | 5500920 |
B59 - Pool 1 | |||||
ROHKGPFP02_1356703233 | 787007 | Mon Dec 26 22:00:33 CST 2022 | 2 | 1 | 1048576000 |
ROHKGPFP02_1356703233 | 787007 | 2 | 2 | 51215102 | |
ROHKGPFP02_1356703232 | 2692559 | Mon Dec 26 22:00:32 CST 2022 | 2 | 1 | 702023900 |
ROHKGPFP02_1354520062 | 2643355 | Thu Dec 01 15:34:22 CST 2022 | 2 | 1 | 673707097 |
B74 - Pool 2 | |||||
ROHKGPEX01_1356703223 | 15 | Mon Dec 26 22:00:23 CST 2022 | 3 | 1 | 681433 |
ROHKGPEX01_1356703222 | 8 | Mon Dec 26 22:00:22 CST 2022 | 3 | 1 | 25 |
ROHKGPEX01_1356703221 | 27 | Mon Dec 26 22:00:21 CST 2022 | 3 | 1 | 1206446 |
rohkgpfp01_1356717627 | 42726 | Tue Dec 27 02:00:27 CST 2022 | 3 | 1 | 5265685 |
rohkgpfp01_1356717626 | 7 | Tue Dec 27 02:00:26 CST 2022 | 3 | 1 | 632026 |
ROHKGPFP02_1356703235 | 7 | Mon Dec 26 22:00:35 CST 2022 | 3 | 1 | 643586 |
ROHKGPFP02_1356703234 | 39793 | Mon Dec 26 22:00:34 CST 2022 | 3 | 1 | 5500920 |
ROHKGPFP02_1356703233 | 787007 | Mon Dec 26 22:00:33 CST 2022 | 3 | 1 | 1048576000 |
ROHKGPFP02_1356703233 | 787007 | 3 | 2 | 51215102 | |
ROHKGPFP02_1356703232 | 2692559 | Mon Dec 26 22:00:32 CST 2022 | 3 | 1 | 702023900 |
ROHKGPFP02_1354520062 | 2643355 | Thu Dec 01 15:34:22 CST 2022 | 3 | 1 | 673707097 |
Any idea?
Regards, Martin
Solved! Go to Solution.
01-21-2014 04:38 AM
Martin
Setting the partially full media to 1 means that it will only be able to use the one tape during duplications so should help
To work out exactly what is going on in your environment would be tricky without having an indepth look at it - unless nothing else writes to tape apart from this vault policy?
My thinking is that you have more than the two tape drives and other jobs also using the drives so that they queue up waiting for a resource
When the vault duplication finishes one image it requests the drive and tape again to write the next image - but something else that was queued up grabs the drive allocation - as that job has the drive and the tape is being unmounted the vault job grabs another drive .. but has to also grab another tape as the original is still being unmounted
Hope that makes some sort of sense - somewhere in the job details or vault log you would probably see this happening
How you could prevent it ahppening would be down to how everything in your environment is setup and runs
01-21-2014 02:28 AM
Hi Martin,
this is acataully depends on how Vault profile is configured and also the Number of Tape Drives that are being used for duplication.
please show us the screenshot of Vault profile and attach to this post to see if we can find some clue on it.
also read the how EMM server allocates the tapes to the jobs from Netbackup admin guide II for better understand...
01-21-2014 02:38 AM
Hi Nagalla,
i have attached the vault configuration. We're using one drive for each duplication.
Thanks for your help ...
01-21-2014 02:43 AM
Hi,
It possible that your B60 media might still have had unexpired images on it when the duplication started. It therefore had to continue on the next (B59) media?
01-21-2014 02:51 AM
Martin
Not totally sure i am with you here - could you explain a little more please
The vault you show goes from disk only to tape and makes 2 copies - one to ENCR_Restore_HKG and one to ENCR_Archive_HKG
But in the opening of the thread you show three pools - B60, B59 and B74 - but only as images and not tapes or volume pools that relate to the vault profile
The images of the two pool 1's equate to the pool 2 - but we dont see tapes here
Please clarify exactly what is what and what you are seeing that you feel is wrong
Thanks
01-21-2014 03:49 AM
Mark, I think those numbers reflect media, not pools (B59,B60, B74).
Otherwise I'm just as lost as you.
01-21-2014 03:59 AM
AH! - in that case it would be down to how many drives there are in total available and exactly how the job actually run - i am guessing that the job that used 2 tapes may have had a pause (due to other jobs running / tape drive cleaning etc) and mounted a second tape into a different drive as it saw that as a free resource
It may be possible to control that behavoir by using the maximum number of partially full media on the pools in question to 1
01-21-2014 04:10 AM
yes, the B59,B60 and B74 are the tapes. i will try it with the option maximum number of partialy full media. but does a value of 1 not mean that we can have 1 partially full media?
maybe it's better to prevent those jobs from interrupting, is that possible?
01-21-2014 04:37 AM
does Allow MultipleRetentions per Media media is enabled? may be the first media got occumpied with other images?
01-21-2014 04:38 AM
Martin
Setting the partially full media to 1 means that it will only be able to use the one tape during duplications so should help
To work out exactly what is going on in your environment would be tricky without having an indepth look at it - unless nothing else writes to tape apart from this vault policy?
My thinking is that you have more than the two tape drives and other jobs also using the drives so that they queue up waiting for a resource
When the vault duplication finishes one image it requests the drive and tape again to write the next image - but something else that was queued up grabs the drive allocation - as that job has the drive and the tape is being unmounted the vault job grabs another drive .. but has to also grab another tape as the original is still being unmounted
Hope that makes some sort of sense - somewhere in the job details or vault log you would probably see this happening
How you could prevent it ahppening would be down to how everything in your environment is setup and runs