The drives are predictive failure and should be replaced. If you have that many predictive failure drives then you have either been ignoring the problem for a long time or there is an underlying issue. Drives go predictive failure when they cross a threshold for bad blocks. This is determined by the SMART on the drive. If you have a lot of drives going predictive failure then it could be a compatibility(non-certified drives), cabling, controller, or backplane issue. It can also be a virtual issue with the RAID array.
If 3 of the 4 drives in your RAID 5 are predictive failure then you have a lot of bad blocks. It is likely that the array is corrupted/punctured. You can read more about replacing drives in the manual.
Please if i need to replaced a hardisk , what the procedure i guess can replace it by open manage execute a hardisk to be offline then replace it , its the recommended way ?
what the procedure i guess can replace it by open manage execute a hardisk to be offline then replace it , its the recommended way ?
Yes, the best practice to replace a predictive failure drive is:
Get a controller/TTY log (if your controller supports this feature)
Get a backup
Run a consistency check (if your controller supports this feature)
Offline the pred fail drive (if your controller supports this feature)
Insert the replacement drive
If a rebuild does not initiate automatically set the replacement as a hot spare
It can take 5 minutes or longer for the status of the drive to update in management applications like OpenManage Server Administrator, so make sure you give it enough time to update the status.
If you need to replace more drives then start over at step 3, be sure to wait until the rebuild of the previous drive completes before starting the process on another drive.
With the number of issues you have it is likely that one of the rebuilds will fail. You should plan accordingly with backups. When you start this process you may have an extended downtime if issues occur. If you would like to post the controller/TTY log I can review it to see if the array is punctured. If you provide a controller/TTY log then please give a URL or upload the text file, please do not copy and paste logs into a forum post.
I can't say for sure whether or not the array is punctured. There is nothing in the log to indicate a punctured array. The log only stores the last 10k lines, and the log only goes back to September 10th. PD 2 and 7 are the only two physical disks that have bad block information. The other drives were likely already predictive failure prior to 9/10.
What I see with the bad blocks on PD 2 and 7 is that most of the bad blocks are sequential. When you have blocks failing in sequence it typically means there was a head crash. It is very unlikely that multiple drives would suffer a head crash at the same time in a system. If multiple drives suffer a head crash at the same time it would likely be an environmental issue. I suspect that the system was moved in some way that caused the head to come into contact with the platters.
DELL-Daniel My
Moderator
Moderator
•
6.2K Posts
1
September 18th, 2017 10:00
Hello
The drives are predictive failure and should be replaced. If you have that many predictive failure drives then you have either been ignoring the problem for a long time or there is an underlying issue. Drives go predictive failure when they cross a threshold for bad blocks. This is determined by the SMART on the drive. If you have a lot of drives going predictive failure then it could be a compatibility(non-certified drives), cabling, controller, or backplane issue. It can also be a virtual issue with the RAID array.
If 3 of the 4 drives in your RAID 5 are predictive failure then you have a lot of bad blocks. It is likely that the array is corrupted/punctured. You can read more about replacing drives in the manual.
https://www.dell.com/perc
Thanks
faisal alshobaki
9 Posts
0
September 19th, 2017 06:00
thanks for reply ,
Please if i need to replaced a hardisk , what the procedure i guess can replace it by open manage execute a hardisk to be offline then replace it , its the recommended way ?
faisal alshobaki
9 Posts
0
September 19th, 2017 10:00
Thanks Daniel for reply its very helpful ,
i was uploaded TTY log, can you check it please .
1 Attachment
lsi_0919.log
DELL-Daniel My
Moderator
Moderator
•
6.2K Posts
1
September 19th, 2017 10:00
Yes, the best practice to replace a predictive failure drive is:
With the number of issues you have it is likely that one of the rebuilds will fail. You should plan accordingly with backups. When you start this process you may have an extended downtime if issues occur. If you would like to post the controller/TTY log I can review it to see if the array is punctured. If you provide a controller/TTY log then please give a URL or upload the text file, please do not copy and paste logs into a forum post.
Thanks
DELL-Daniel My
Moderator
Moderator
•
6.2K Posts
1
September 19th, 2017 11:00
I can't say for sure whether or not the array is punctured. There is nothing in the log to indicate a punctured array. The log only stores the last 10k lines, and the log only goes back to September 10th. PD 2 and 7 are the only two physical disks that have bad block information. The other drives were likely already predictive failure prior to 9/10.
What I see with the bad blocks on PD 2 and 7 is that most of the bad blocks are sequential. When you have blocks failing in sequence it typically means there was a head crash. It is very unlikely that multiple drives would suffer a head crash at the same time in a system. If multiple drives suffer a head crash at the same time it would likely be an environmental issue. I suspect that the system was moved in some way that caused the head to come into contact with the platters.
Thanks
faisal alshobaki
9 Posts
0
September 19th, 2017 11:00
ok daniel , so we need to replaced all HDD's that has a warnings ?
Please advise for the recommended path to solve these issues .
Thanks alot for your support .
DELL-Daniel My
Moderator
Moderator
•
6.2K Posts
0
September 19th, 2017 14:00
Yes, you should replace the predictive failure drives.
You can follow the steps I posted previously.
Thanks