I really don't see why you can't use multiple drive - caveat is on Celerra side where you can have only 4 steams per data mover.
The speed you get is not good. While I didn't do network based Celerra backups, I did with directly attached drives using tar/dump and VBB. Usually I would get 50-70MB/s (I had one drive per data mover). So, I'm quite sure Celerra is able to perform. Since you have network in between I would check first that network card settings are in sync between both sides (and ports on switch).
You will probably need EMC support (Celerra) on this as VBB requires a file (actually a file system) on each DM which is used by restore operation of that kind. Once Celerra is configured for VBB you can start enjoying block based backups which should make difference in performance and speed you see.
I did speak with a Celerra engineer and he swayed me away from using VBB as it is doing volume backups, and the majority of our restore requests are file level restores. So this probably wouldnt be the best solution for us.
While doing some testing, I found that each backup thread from the NAS backs up at around 7Mb/s. So if I have a backup session going from the NAS, it is backing up at around 24Gb/hour. However, if I start another backup from the NAS, it too will run at the same speed. I almost expected to see both jobs running at 12Gb/hour to total 24Gb/hour, but it didnt. It seemed as though each job was backing up at the same speed.
That just seems to confirm there is some type of configuration somewhere that is limiting each backup thread to only go a certain speed.
ble1
4 Operator
•
14.4K Posts
0
April 23rd, 2007 08:00
The speed you get is not good. While I didn't do network based Celerra backups, I did with directly attached drives using tar/dump and VBB. Usually I would get 50-70MB/s (I had one drive per data mover). So, I'm quite sure Celerra is able to perform. Since you have network in between I would check first that network card settings are in sync between both sides (and ports on switch).
briantrotter
15 Posts
0
April 23rd, 2007 12:00
I did a "server_pax server_2 -stats -v" and this is one of the things I see.
Not sure if this is helpful to anyone.
** nasa thid 3 is running backup with dump format **
Backup root directory: /root_vdm_11/DENFS1
Total bytes processed: 1737035041893
Total file processed: 819081
throughput: 7 MB/sec
average file size: 2071KB
Total nasa wait nass count: 13491
Total nasa wait nass time: 34885 msec
Total time since last reset: 233437 sec
1177355630: NDMP: 5: PaxDataModule get NULL tape device
Tape device name: unknown
dir or 0 size file processed: 70170
1 -- 8KB size file processed: 276639
8KB+1 -- 16KB size file processed: 62804
16KB+1 -- 32KB size file processed: 95295
32KB+1 -- 64KB size file processed: 68521
64KB+1 -- 1MB size file processed: 170842
1MB+1 -- 32MB size file processed: 68148
32MB+1 -- 1GB size file processed: 6387
1G more size file processed: 275
fs /root_vdm_11/DENFS1 size is: 1891772768256 Bytes
Estimated time remain is 21081 sec
ble1
4 Operator
•
14.4K Posts
0
April 23rd, 2007 13:00
briantrotter
15 Posts
0
April 25th, 2007 09:00
ble1
4 Operator
•
14.4K Posts
1
April 25th, 2007 11:00
briantrotter
15 Posts
0
April 26th, 2007 08:00
While doing some testing, I found that each backup thread from the NAS backs up at around 7Mb/s. So if I have a backup session going from the NAS, it is backing up at around 24Gb/hour. However, if I start another backup from the NAS, it too will run at the same speed. I almost expected to see both jobs running at 12Gb/hour to total 24Gb/hour, but it didnt. It seemed as though each job was backing up at the same speed.
That just seems to confirm there is some type of configuration somewhere that is limiting each backup thread to only go a certain speed.
ble1
4 Operator
•
14.4K Posts
0
April 26th, 2007 13:00
ajw2
1 Rookie
•
88 Posts
0
April 27th, 2007 08:00
ble1
4 Operator
•
14.4K Posts
0
April 29th, 2007 05:00