I am having alot of issues with my DSS units. majority of the jobs are failing because the disk space gets full and it takes a long time to go onto tape then to clear the space on disk.
Three Disk Staging Storage Units were created:
One for Files
One for Databases
One for Windows
The Disk Staging Storage units space is taken from the NetApp Appliance. (2TB) drive letter (T)
I've created folders inside the T driver namely (Files)(Databases)(Windows) and pointed each DSSU to those folders.
Low watermark = 75%
High Watermark = 98%
Frequency= 6 hrs
Start Window= 24/7
Destination= Multiple Copies (On site Pool and Offsite Pool)
Is there anyway i can sent the data faster to tape and also allow for the image to be cleaned up on the disk for other disk jobs to run?
Lower your water marks. You're currently set to wait until its 98% full until netbackup will run a clean up. I'd put that figure (HIGH) to 75% which means it would clean up when there is still 500GB left free. I'd put to LOW to 50% or less. This all depends on what you write per night (or backup run).
Also, having 3 folders/stu's isn't of much use. Only thing it might do i scew your view of what is free on them (as they're all on the same drive they'll all show X free, which might look like a lot but its global)
The Disk Staging Unit is only a temporary place for your backups. They must be relocated to a 'final' storage unit (Where the data stays until it expires)
The frequency dictates how often data is 'destaged' to its final storage unit (Tape for instance)
When data is 'destaged' from the staging unit, it remains on the staging disk as long as possible, but is considered purgable (As it is also on tape).
When Netbackup needs space (As in your case), it will purge the files on the staging disk it has succesfully destaged to tape.
The staging disk then has more room again...
As others have mentioned: The watermarks are important as well!
If you decrease the frequency (To for instance 1 hour), Netbackup will try to destage any backups that are present on the staging disk every hour...
Data that is destaged already, won't be destaged again.
I have a setting of 1 hour, but your setting depends on how often you want to destage...
The lower the frequency you set, the more often NBU tries to destage (And free up space, which you can use for backups)
Just give it a try and find out which setting works best for you. You can ALWAYS change back to 6 hours as well
Ok Will try that also. I would like data to go to Tape as soon as possible. The reason for doing disk staging was because alot of the jobs were failing before. I had them set to go to tape only.
Hi, each DSSU should be big enough to hold the total amount of data sent during a backup session.
When the backup session is finished (all your jobs are done), you should have enough time to destage data to final destination, that means : relocation must be done and finished before next backup session.
As HWM and LWM look the % full of the file system, and don't know about different folders, using 3 folders for 3 DSSU on the same file system make NBU find a 98% reached because of one folder (1 DSSU) but can not delete data, because they have not been destaged from another folder !
Rule of the thumb : 1 DSSU = 1 file system !
Now, the size of the file system is to be calculated regarding the time you want data to remain in disk, and for how long you want to have a fast access for restoring these data.
If your backup session amount of data is 500gb each day, and you want to keep a history of 3 days for fast restoration, you need at least 1500gb of disk (3 x 500gb).
When the 4th day, the backup session starts ... assuming staged data was relocated to final destination, NBU will be able to remove from disk enough data to reach the low water mark. (in fact, as soon as the HWM was reached an image cleanup job would have done this deletion).
So, if you want a HWM at 98%, for these 3 days of 500gb, that means your 1500gb are the 98% ... so your file system should be 1531gb. With a HWM of 80%, you'll need 1500/80*100=1875gb !
Now, your LWM defines how much data must be deleted when HWM is reached after a backup, or when 100% is touched during a backup.
If you want a max of 500gb (the older backup session) to be deleted from your DSSU of 1531gb, that means you want to keep 1531-500=1031gb which is 67% of 1531 .... a LWM of 67% can be used.
To delete 500gb from the 1875gb DSSU, thats 1875-500=1375gb ... 73% of 1875 for your LWM.
Ok, I did the steps mentioned above and the jobs are working ok. One things is we are really short on disk space. What I would like to know is, are there any settings that can be altered to limit the number of days of backup being sent to tape? i.e. keeping just one backup on Disk instead of having a number of days.
If you only want one day of data on the disk, you need to calculate how big that is.
if your one day is 500G, then 2 days would be 1000G - depending on whether you have one large or many small backups, you should set your storage size and watermarks so that you have enough space to do two days of backups, but set your watermarks to trigger cleanup before you get there, and clean up back to one day.
You can also set priorities on your policies, so your gold backups are cleaned up last - I think.
I would set my size to 3 days worth of data, clean up after the second day (>66%), back to one day( 33%). That way you never clean up too soon. You can cheat a bit on the percents say 70 and 40...
I keep one week on disk. My tape are for DR, or archive restores, 90% of my restores are from disk.
I use SLP to copy to tape from disk. You can set the expiration of the 1st copy in the SLP to a fixed time or expire after duplication validated. Either way, the SLP guarentee it will not remove copy 1 until copy 2 is complete.