2017-06-10 07:27 AM
Recently we bout new all flash netapp AFF8040 model.
We are trying to have better structure for our Oracle Database in term of performance for backuppool filesystem in solaris.
We try few structure to test the time fo the backup as below
Day 1 : 1 Volume 1 Lun at controller B. Backup time is 30 minutes
Day 2 : Multiple volume with single lun at controller B. Backup time is 22 minutes
Day 3: 3 volumes with single lun at Controller A, 3 volumes with single lun at Controller B. Backup time is still 22 minutes.
Please refer the attachment.
So the question is why went we setup volume in different controller but the performance look like same as setup multiple volume in same controller?
Is that the load not distributed evenly to the both controller?
2017-06-11 07:51 PM - edited 2017-06-11 07:51 PM
Do I understand correctly that in scenario 2 and 3 you are concatenating the LUNs in the host, using Sun Volume Manager/ VXVM or similar?
How is the host connected to the array? From your data, I see you have about 420GB of data. At 22 minutes, that works out to ~325MByte/sec