HP MSA1000 RAID5- Organization

[enroute2]

Anyone have a clue on how the data is specifically laid out- algorithm? It seems parity is on each drive, but not sure if in groups or data-parity-data, etc.

[phishin_ca]

This will have the typical delayed parity you see with all smart array controllers. Try with a delay of 2 first.

[dmarques]

How many drives?
I agree with phishin_ca, most probably you'll have delayed parity like all HP servers.

[hddguy]

Try with a delay of 2 first.

Delay in HP always will be higher than 2...

Anyone have a clue on how the data is specifically laid out- algorithm? It seems parity is on each drive, but not sure if in groups or data-parity-data, etc.

Data and parity is written as in standard RAID5, except parity is written to the same disk for a set number of stripes - this is the delay.

[ici_lemmy]

For information, WinHex can handle delay for HP RAIDs...

[hddguy]

I usually use R-Studio and create a custom block layout to deal with these.

[dmarques]

Both fits for the same goal, euth a rewind differences between them.
Winhex is much better for analysis and r-studio is better for rebuilding.

[phishin_ca]

I am just partial to the WinHex way. With a larger array you have to do a lot more work in RStudio. Once I have the volume snapshot in WinHex, I feel no reason to leave

[hddguy]

I am just partial to the WinHex way. With a larger array you have to do a lot more work in RStudio. Once I have the volume snapshot in WinHex, I feel no reason to leave

Personally I always use both. Winhex is great for analysis of disks, has a great option to synchronise multiple windows while R-Studio has a nice interface, many useful features and is easily configured.

But then again, I do not think any engineer working on RAID arrays will limit themselves to 1 or even 2 software titles.

[dmarques]

Totally right, a good DR engineer will adapt itself to any app.