question about replacing bad disk on dell powervault 220s raid 5
need to replace a "bad" disk on a powervault 220s, currently setup in raid 5
the disk is blinking amber on the array, we have sql data being written to this array, and want to make sure there arent any errors when rebuilding, its an extremely important production system.
any ideas on how to move forward? no support from dell, this device is EOL as of last year.
not the first time im replacing a bad disk in an array, but i need to be 110% positive i dont run into any problems.
in the dell openmanage array manager, this is what shows:
ARRAY DISK 0:0 MAXTOR
ARRAY DISK 0:1 MAXTOR
ARRAY DISK 0:2 MAXTOR
ARRAY DISK 0:3 MAXTOR (BAD DISK)
ARRAY DISK 0:4 MAXTOR
ARRAY DISK 0:5 SEAGATE
ARRAY DISK 0:8 SEAGATE
the ARRAY DISK 0:0 is unallocated, and i can right click on it and select "ASSIGN GLOBAL HOT SPARE", would this be sufficient right now?
also, why are disks 6 and 7 missing?
one more thing, these are all 36gb 15k drives, i also have a 36gb 15k drive from dell but its a hitatchi, this is the one i was going to use to replace the bad drive, would this work? i was told by dell at first, that it wouldnt, then i was told that if they were the ones that shipped out the drives, it would work. i need to make sure
thanks!
the disk is blinking amber on the array, we have sql data being written to this array, and want to make sure there arent any errors when rebuilding, its an extremely important production system.
any ideas on how to move forward? no support from dell, this device is EOL as of last year.
not the first time im replacing a bad disk in an array, but i need to be 110% positive i dont run into any problems.
in the dell openmanage array manager, this is what shows:
ARRAY DISK 0:0 MAXTOR
ARRAY DISK 0:1 MAXTOR
ARRAY DISK 0:2 MAXTOR
ARRAY DISK 0:3 MAXTOR (BAD DISK)
ARRAY DISK 0:4 MAXTOR
ARRAY DISK 0:5 SEAGATE
ARRAY DISK 0:8 SEAGATE
the ARRAY DISK 0:0 is unallocated, and i can right click on it and select "ASSIGN GLOBAL HOT SPARE", would this be sufficient right now?
also, why are disks 6 and 7 missing?
one more thing, these are all 36gb 15k drives, i also have a 36gb 15k drive from dell but its a hitatchi, this is the one i was going to use to replace the bad drive, would this work? i was told by dell at first, that it wouldnt, then i was told that if they were the ones that shipped out the drives, it would work. i need to make sure
thanks!
ASKER
so if i make 0:0 the global hot spare, it would start to rebuild right away, right?
im sure it would slow performance, so i would want to do this after hours,
should i do it straight from the array manager? is there another way to do it by taking the unit offline?
im sure it would slow performance, so i would want to do this after hours,
should i do it straight from the array manager? is there another way to do it by taking the unit offline?
Yes, make the hotspare change and it begins rebuild straight away.
Start it now, if you lose another disk on a RAID 5 you're in trouble, it's going to take a while anyway so chances of you completing the task completely outside working hours is slim to none.
As far as the multiple brands.. often times various brands of HDDs can be used in one array without much issue noticed by the user because of the controller's and the drive's error correction abilities. For an array to be optimal and in best form it should not only be of the same brand but should be the same firmware version across all disks as well. Sure you can get away without doing this while using the array, recovery though will prove to be a different story. So lets say another drive should fail before you fix this array, with different brands and I assume various firmwares across same branded disks, when you send them off to a recovery company I think you will find you should have been slightly more diligent about these little details being correct.
As far as the multiple brands.. often times various brands of HDDs can be used in one array without much issue noticed by the user because of the controller's and the drive's error correction abilities. For an array to be optimal and in best form it should not only be of the same brand but should be the same firmware version across all disks as well. Sure you can get away without doing this while using the array, recovery though will prove to be a different story. So lets say another drive should fail before you fix this array, with different brands and I assume various firmwares across same branded disks, when you send them off to a recovery company I think you will find you should have been slightly more diligent about these little details being correct.
ASKER
I think we are going to wait til after hours to start this, i was told by our development team that we need to get a full backup (86gb) of what is on the drive before we start the rebuild process with the hot spare. probably do it over the weekend, start friday night, or possibly start sunday morning after a full backup. what do you guys think?
I think:
A - This is a RAID 5 which means that with a disk failed you no longer have any protection. If another drive fails while you wait you lose everything and since you have a mixed soup of disks in the array recovery will be difficult if possible at all.
B - 85GB isn't much and it isn't a bad idea at all to grab a quick backup prior to rebuild.
C - This is a RAID 5 and with a disk failed you no longer have any protection.
Today is Wednesday - you want to wait till Friday to rebuild this array? Really?
A - This is a RAID 5 which means that with a disk failed you no longer have any protection. If another drive fails while you wait you lose everything and since you have a mixed soup of disks in the array recovery will be difficult if possible at all.
B - 85GB isn't much and it isn't a bad idea at all to grab a quick backup prior to rebuild.
C - This is a RAID 5 and with a disk failed you no longer have any protection.
Today is Wednesday - you want to wait till Friday to rebuild this array? Really?
I understand your situation (about have to wait).
some system, is likely not allow to have unschedule downtime.
so, the best you could do for right now is,
change recovery model be full and do database full backup & very often transaction log backup
or
if your recovery model is anything but full, then do database backup as often as you could
however, as many comment mention above, rebuild should take a while, but with 36g 15k, I'm sure it'll take less than a day
good luck!
some system, is likely not allow to have unschedule downtime.
so, the best you could do for right now is,
change recovery model be full and do database full backup & very often transaction log backup
or
if your recovery model is anything but full, then do database backup as often as you could
however, as many comment mention above, rebuild should take a while, but with 36g 15k, I'm sure it'll take less than a day
good luck!
It's RAID 5, why would there be down time for this rebuild?
ASKER
hey guys
i set the spare disk in the array to GLOBAL HOT SWAP and it started to rebuild with the new drive
when it rebuild was done (only took about 1.5 hours) i checked the status, and the drive that was previously bad/failed, was now showing as "READY" and not part of the array. could this drive still be good? why was it showing as bad in the first place?
i set the spare disk in the array to GLOBAL HOT SWAP and it started to rebuild with the new drive
when it rebuild was done (only took about 1.5 hours) i checked the status, and the drive that was previously bad/failed, was now showing as "READY" and not part of the array. could this drive still be good? why was it showing as bad in the first place?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
thanks for your help dude
regarding different brand, I'm not so sure, but in my environment, I've few brand of HDD working together in same unit
Seagate, Maxtor, WD
So far, I never face any problem.