September 30, 201213 yr I have a test server that has 3 data disks, 2x 1 TB and 1x 750 GB The parity and cache disk are a 3 TB RAID0 volume for parity and and a 500 GB RAID1 Volume for cache on a ARC 1200 controller. My observatation when doing a parity built/sync is speed starting at about 80 MB/s slow dropping to 50-55 MB/s at 33% (1 TB point) after that when the data disks are no longer accessed speed seems to be "capped" at about 55-60 MB/s. Is this normal behaviour? A parity check runs at 98 MB/s at start to about 50 MB/s at 33% (1 TB point) and then increases to close to 200 MB/s. Should the parity disk size be limited to about the largest data disk size?
September 30, 201213 yr If the parity disk is larger than the data disks, then obviously some time is being spent checking or rebuilding parity where there is no equivalent data on the other drives. So you can argue that is time which is wasted. On the other hand, because (from unRAID's point of view) you can expand the array at any time with a larger data disk, it has to have the entire parity disk in a known state just in case it is needed later. I see nothing wrong with the speeds that you have reported.
Archived
This topic is now archived and is closed to further replies.