Adaptec 6805 Critical error

[thatoneuser]

I created a RAID 5 VD with 4 drives on an Adaptec 6805. One of the drives failed, so I replaced it and completed the rebuild. When checking the drives individually and the VD in maxView Storage Manager all show "Optimal," however there is a critical exclamation point on the VD. I do see there is an event saying one or more logical devices contain bad stripe, so I've tried to run verify and verify with fix several times but the exclamation point just won't go away.

Does anyone have any idea what might be going on?
Thanks in advance for your opinion!

 

[i386]

I've tried to run verify and verify with fix several times

That are probably the information posts in the log ("Task priority changed from low to high").
Did you you check in the "task" tab if the first run finished? (I think in my case it took 48+ hours on a 6x 6tb hdds raid 6)

 

[moblaw]

Bad news. You have to destroy that array and create a new array. I had this happen on a 50tb r6 array during a drive and shutdown meltdown.

So move your data elsewhere. Destroy that array, create new and move the data back. You will encounter corrupt data, which you will need to check and replace if found. But you won't nessarcerly find any.

(Bad data or a raid that had miss alligment happen to the parity data)

 

[thatoneuser]

That are probably the information posts in the log ("Task priority changed from low to high").
Did you you check in the "task" tab if the first run finished? (I think in my case it took 48+ hours on a 6x 6tb hdds raid 6)

I think it did. All I see in the task tab is a scheduled task to check all spare disks for error.

 

[thatoneuser]

Bad news. You have to destroy that array and create a new array. I had this happen on a 50tb r6 array during a drive and shutdown meltdown.

So move your data elsewhere. Destroy that array, create new and move the data back. You will encounter corrupt data, which you will need to check and replace if found. But you won't nessarcerly find any.

(Bad data or a raid that had miss alligment happen to the parity data)

That sucks... If I were to just ignore it and keep using as-is would there be any problem? I guess I'll use this opportunity to clean up and redo the setup.

 

[moblaw]

That sucks... If I were to just ignore it and keep using as-is would there be any problem? I guess I'll use this opportunity to clean up and redo the setup.

You can ignore it, BUT data written now, will be mixed with the "BAD" parity data, and the probability of the new data going bad, is very high. I would stop writing to that array, because of that. It shoud be a "read only" until you can remove your stuff - How many TB are we talking about?

 

[Tom5051]

a bad stripe occurs when the array has lost the ability to regenerate the stripe data from parity or mirror disks. If you keep using it you will start to find files become corrupt, e.g. movie files will be missing sections or not play at all. This happened to me last year on my 6805 and I have to say the warning provided from the GUI is not very obvious. Just changes NO to YES.
As said, the only way is to destroy the array and recreate it.

 

[thatoneuser]

You can ignore it, BUT data written now, will be mixed with the "BAD" parity data, and the probability of the new data going bad, is very high. I would stop writing to that array, because of that. It shoud be a "read only" until you can remove your stuff - How many TB are we talking about?

a bad stripe occurs when the array has lost the ability to regenerate the stripe data from parity or mirror disks. If you keep using it you will start to find files become corrupt, e.g. movie files will be missing sections or not play at all. This happened to me last year on my 6805 and I have to say the warning provided from the GUI is not very obvious. Just changes NO to YES.
As said, the only way is to destroy the array and recreate it.

Thank you for the info. I have another pool I can dump these 9TB onto and redo from scratch. Unfortunately both systems are HP MS G8 so I can only do the transfer @ 2Gb. That's gonna take a while, hopefully nothing goes wrong during the transfer.

 

[Tom5051]

Yer that's the best way to go. The disks in my array were all perfectly fine and passed a thorough scan.
Has been running fine since I rebuilt the array. I was expanding my array when the bad stripe occurred, did yours just happen on it's own or were you making changes to the array?

 

[thatoneuser]

Yer that's the best way to go. The disks in my array were all perfectly fine and passed a thorough scan.
Has been running fine since I rebuilt the array. I was expanding my array when the bad stripe occurred, did yours just happen on it's own or were you making changes to the array?

It just happened on its own, I didn't tinker with anything.

 

[moblaw]

It should never happen on its own.

Can you check crc error count & aborted commands number?