Parity Delay on Promise RAID

Jared

Administrator
Staff member
Has anyone ever come across a Promise RAID brand external RAID system that used parity delay?

I'm dealing with one now that seems to have some unexplainable anomalies. It's clearly using a single sector stripe size, I've confirmed that several ways including by using entropy analysis (and experience told me to expect this). However, I've built the RAID every possible way and none of them are correct. I've confirmed that the XOR is good from beginning to end across all three drives, so it's not a matter of an out of sync drive or anything like that either. An offset shouldn't make any difference with single-sector stripe size, so that's not really a concern.

So the only possibilities are that it's some sort of non-standard block rotation or a parity delay. Any ideas?
 

jmbach

New member
Jared":lby2pxq4 said:
Has anyone ever come across a Promise RAID brand external RAID system that used parity delay?

I'm dealing with one now that seems to have some unexplainable anomalies. It's clearly using a single sector stripe size, I've confirmed that several ways including by using entropy analysis (and experience told me to expect this). However, I've built the RAID every possible way and none of them are correct. I've confirmed that the XOR is good from beginning to end across all three drives, so it's not a matter of an out of sync drive or anything like that either. An offset shouldn't make any difference with single-sector stripe size, so that's not really a concern.

So the only possibilities are that it's some sort of non-standard block rotation or a parity delay. Any ideas?

What kind of anomalies are you experiencing with the RAID system. How is the RAID behaving.

My experience is mostly with internal RAID controllers on servers.
 

Jared

Administrator
Staff member
jmbach":15tjbxqt said:
[post]12972[/post] How is the RAID behaving.

The enclosure just died. So now I'm trying to rebuild it in recovery software. All drives are fine and 100% cloned with no bad sectors. No matter how I've built it using the known stripe size files aren't properly destriping. From analyzing the look of jpg files that I recover using the various builds it's clear that some things are right, but there's something off with the parity or block rotation in every possible build. I just can't figure out what it is.

Mind you, I've done a ton of RAID cases including some very large complex ones. A 3 drive RAID normally only takes me 20-30 min to figure out. Something is definitely amiss here.
 

Jared

Administrator
Staff member
OK, I finally solved the case. It wasn't a parity delay at all. I was correct in assuming a 1 sector stripe size. Seems there were some bogus MFT entries near the beginning of the drive that were to old deleted files that don't match the current RAID striping. Thus the anomolies.

After allowing it to scan overnight at several of the settings that I knew had to be close one finally found a good folder structure and all files seem good.
 
Top