Hi, I have updated the firmware of our MD3620i array to allow us to create a DDP. I have 48 x 600GB SAS 10K HDs in the DDP which has given me usable capacity of 19.253TB and preservation capacity of 1,668.000 GB (3 Physical Disks).
I then created 3 equally sized LUNs which uses up all of the disk pool space (about 6,572GB in size each). After I finished this the array started warning me that the DDP was full. Is this a problem?
I don’t want to create LUNs and leave free space that I can’t use, although if we have disk failure(s) will I then encounter issues? I would think the preservation capacity of 3 physical disks is what it would use to rebuild onto, so can I go into production with this configuration?
Let me clear this up for you. If you were using traditional disk groups then when you have a failed disk the hot spare will take over for the failed disk. Once the disk is replaced then the hotspare will go back to being a hotspare again.
When using a Dynamic Disk Pool it will use the reserved space that you provide for the DDP. If in the event that the space provided is full then it will take space from the disk pool itself to do the rebuild. Once the failed drive has been replaced then if it used any space in the disk pool it will return the space. If the space in the pool is not enough to allow the rebuild to complete the rebuild can run as long as it has all the matching segments need to keep the pool up in a degraded state.
Please let us know if you have any other questions.
No it is not an issue to use all the space. The error message can be disabled so that you don’t get that particular message. What you need to do is to go into the alerts & it should be under utilization warning threshold or capacity depletion.
Please let us know if you have any other questions.
So when you have a failed disk in you MD this is how it works.
Disk Group Rebuild - uses a hot spare or replacement physical disk
Disk Pool Rebuild
Uses reconstruction reserved amount. Enough member physical disks present, one for each piece of the same stripe. (example: 10-stripe disk VD, Raid 6, need at least 8 disks for the rebuild)
If reconstruction reserved amount is depleted, uses free capacity in the pool for the rebuild (if possible)
Please let us know if you have any other questions.
1) will it take space from the disk pool itself to do the rebuild? which is the "preservation capacity of 1,668.000 GB (3 Physical Disks)" ? Yes it use the preservation space first.
2) I could survive with a maximum of 3 failed drives at the same time? i know this is unlikely but just wondering. Yes unless they all failed at the same time then you might have an issue. But it is unlikely that you will have 3 drives fail all at the same time.
Thanks for your reply. Yes I've managed to disable the alert/warning so it's all green.
But just to double check, is the "preservation capacity of 1,668.000 GB (3 Physical Disks)" what it will use if 1 or 2 disks fail?
Is it max 2 disk failures at the same time or 3?
If, lets say, we have some HDs failures and it uses the preservation capacity disks to rebuild and we forget to replace the failed drives and then by chance have some more failures, won't it then try to rebuild and in doing so reduce the size of the LUNs/volume and corrupt some or all of the VM data? Or if not, what would happen?
Sorry not sure I understand, is the "preservation capacity of 1,668.000 GB (3 Physical Disks)" what it will use if 1 or 2 disks fail? Or does it actually then use the "reconstruction reserved amount" on the remaining disks?
Can you also answer this one please:
If, lets say, we have some HDs failures and it uses the preservation capacity disks to rebuild and we forget to replace the failed drives and then by chance have some more failures, won't it then try to rebuild and in doing so reduce the size of the LUNs/volume and corrupt some or all of the VM data? Or if not, what would happen?
DELL-Sam L
Community Manager
•
8058 Posts
•
34386 Points
1567
1
Posted September 11th, 2013 14:00
Hello Si,
Let me clear this up for you. If you were using traditional disk groups then when you have a failed disk the hot spare will take over for the failed disk. Once the disk is replaced then the hotspare will go back to being a hotspare again.
When using a Dynamic Disk Pool it will use the reserved space that you provide for the DDP. If in the event that the space provided is full then it will take space from the disk pool itself to do the rebuild. Once the failed drive has been replaced then if it used any space in the disk pool it will return the space. If the space in the pool is not enough to allow the rebuild to complete the rebuild can run as long as it has all the matching segments need to keep the pool up in a degraded state.
Please let us know if you have any other questions.