That configuration will create TWO RAID6 RAIDsets One for the SSDs and the other for the SAS drives. Plus a spare for the SAS. So that's 5 total drives for parity and spares.
The size won't change at 100%. Also some space is set aside from each drive. So the SSDs are about 1.8TB The Spinning drives will be around 7.7TB So that's correct.
The amount of reserve depends on how much write activity you have an how long you want to keep snapshots around. If you wanted to keep a years worth you would have to potentially allocate more than the size of the volume. Much less if you just want to have some for a few days or before major events. Resize, end of quater, end of year.
Re: Volume. I have found that multiple smaller volumes perform better than one massive volume. There are a number of reasons for this based on how SCSI works.
It's not the "size" that's important. It's that there is only one. Like only having a single checkout in a store. At times there will be a line at that register. If you had two or four registers working there wouldn't be a backup
If you haven't written much to that volume you can thn provision it. Create a smaller volume also thin provisioned and move the VMs to the smaller volume. Then delete the 6TB volume create a another smaller one and move VMs there. Depending on the capability of your hypervisor and how much space you have used so far.
dwilliam62
4 Operator
•
1.5K Posts
0
January 5th, 2021 12:00
That configuration will create TWO RAID6 RAIDsets One for the SSDs and the other for the SAS drives. Plus a spare for the SAS. So that's 5 total drives for parity and spares.
The size won't change at 100%. Also some space is set aside from each drive. So the SSDs are about 1.8TB The Spinning drives will be around 7.7TB So that's correct.
Regards,
Don
fred974
1 Rookie
•
40 Posts
0
January 5th, 2021 15:00
@dwilliam62 thank you again for your help.
I have now reach the volume creation part.
Do you have any recommendation for me? I am currently think to do something like this:
I have never used the snapshot reserve before so I have no idea on what value to give it.
The EqualLogic is used as a shared storage for 4 xcp-ng hypervisor and I have around 4TB of storage on all VM combine.
I do not use anything else, the array is purely for xcp-ng
dwilliam62
4 Operator
•
1.5K Posts
0
January 6th, 2021 11:00
Hello,
The amount of reserve depends on how much write activity you have an how long you want to keep snapshots around. If you wanted to keep a years worth you would have to potentially allocate more than the size of the volume. Much less if you just want to have some for a few days or before major events. Resize, end of quater, end of year.
Re: Volume. I have found that multiple smaller volumes perform better than one massive volume. There are a number of reasons for this based on how SCSI works.
Regards,
Don
fred974
1 Rookie
•
40 Posts
0
January 6th, 2021 15:00
Hi @dwilliam62 ,
I only just seen your reply. I have already create a large volume of 6TB. I tried to reduce it but it look like it is not possible.
Do you think 6TB is a 'large' volume?
dwilliam62
4 Operator
•
1.5K Posts
0
January 7th, 2021 07:00
Hello,
It's not the "size" that's important. It's that there is only one. Like only having a single checkout in a store. At times there will be a line at that register. If you had two or four registers working there wouldn't be a backup
If you haven't written much to that volume you can thn provision it. Create a smaller volume also thin provisioned and move the VMs to the smaller volume. Then delete the 6TB volume create a another smaller one and move VMs there. Depending on the capability of your hypervisor and how much space you have used so far.
Regards,
Don