I am using LTO6 tapes with the HPE MSL2024 tape library.
I am sure the Hardware Compression has been enabled at the tape library level, but the data writing to the tapes does not seem to be compressed at all. Each tape shows Full when the usage reachs to 2.8TB which is just raw capacity.
BTW, my NetBackup system is a NetBackup Appliance 5230. It looks like the hardware compression is disabled from the Appliance somewhere.
Any idea would appreciated.
Ross
Hello
Well all depends on the data you do back up - some data does not compress well, ie. movies, pictures etc as it is already compressed. In regards of LTO6 it is 2.5 TB in native as per this roadmap LTO6 capacity is 6.25 TB and assumed compresion ratio is 2.5 so when we will divide the 6.25/2.5=2.5 TB - so looks like some tiny compression is there... as you tapes do show 2.8 TB, right?
Thank you, quebek.
Yes, the tapes in Full show about 2.8TB. Most of the backup data are Hyper-V VMs, Oracle RMAN backups, and some regular document files.
I have two almost identical NetBackup sites/systems. The other one shows about 5.0TB in Full. Comparing with the other one, I thought this site was not compressed. Maybe it is because this site has more VMs backups and Oracle RMAN data which can't get high compression rates?
Thanks again,
Ross
Please ensure the compression option in the policy in unchecked.
Compression in the policy is Netbackup built-in software compression, it is very slow and will greatly reduce the number of data you can have on tape (compressed data can't compressed a second time).
LTO tape drive will per default compress data, you don't have to do somthing "extra".
Thank you, Nicolai.
Yes, the compression option in the policy in unchecked.
I was wondering if the compression was disabled at the Appliance/Linux OS level. But as quebek counted, it does have a tiny compression. It's kind of interesting.
Thanks again,
Ross
rossxu - Please see this tech note: http://www.veritas.com/docs/000091135
Since a Veritas appliance is running Linux, you can use the GEN_DATA directive to generate random data. My idea here is to generate a 50% compressible data stream and then fill up a tape. If you still get native capacity, for sure somthing is not right, but else you should see a tape with a lot of data on it.
Writing a entire tape may take some time ...
