We have several HP DL385 boxes with onboard RAID and SFF SAS drives (146GB, some 10K, some 15K).
When doing Advanced disk benchmarking, we consistently get sequential write numbers of 2-3MB on a 6-disk array when configured with RAID 5.
800MB test file
16K block size
Async with depth 20
This 3MB number does not reflect the actual disk performance when ballparked with simple file copies. It appears to be a testing anomaly. The sequential read numbers appear to be fine (120-140MB).
When the same drives are configured for RAID10, we get what appear to be reliable numbers (130-150MB sequential read, 100-110MB sequential write).
Is there something in Passmark 7 advanced disk testing that is not valid for a SAS RAID5 array on HP RAID controllers?
When doing Advanced disk benchmarking, we consistently get sequential write numbers of 2-3MB on a 6-disk array when configured with RAID 5.
800MB test file
16K block size
Async with depth 20
This 3MB number does not reflect the actual disk performance when ballparked with simple file copies. It appears to be a testing anomaly. The sequential read numbers appear to be fine (120-140MB).
When the same drives are configured for RAID10, we get what appear to be reliable numbers (130-150MB sequential read, 100-110MB sequential write).
Is there something in Passmark 7 advanced disk testing that is not valid for a SAS RAID5 array on HP RAID controllers?
Comment