RAID10 or RAID6.
RAID10 have the best performance, but it can tolerate only 1 disk fail, RAID6 have worst write performance but it can tolerate more hdd fail.
There is a calculator, you can check different solutions:
http://wintelguy.com/raidcalc.pl
http://wintelguy.com/raidperf.pl
But everything is depending on, you have more write than read, or reverse, you have BBU and cache or not, etc.
RAID10 have the best performance, but it can tolerate only 1 disk fail, RAID6 have worst write performance but it can tolerate more hdd fail.
There is a calculator, you can check different solutions:
http://wintelguy.com/raidcalc.pl
http://wintelguy.com/raidperf.pl
But everything is depending on, you have more write than read, or reverse, you have BBU and cache or not, etc.
Any Raid is a tradeoff performence versus redundancy versus raw capacity. Best performance and best redundancy is a mirror(raid1 with 8 disks) with all yours hdd but the usable space is only 1 tb.
Raid 10 is in the middle (decent performance, decent redundancy and decent capacity). For example in your case could be 4x mirror (2x 1 Tb hdd).
PS. Raid6/5 has a very bad IOPs.
Raid 10 is in the middle (decent performance, decent redundancy and decent capacity). For example in your case could be 4x mirror (2x 1 Tb hdd).
PS. Raid6/5 has a very bad IOPs.
Hi,
Thank you for your response. Sorry, but i didn't understand why the usable space is only 1 TB from all the 8 disks.
also for Raid 10, using 4x mirror the usable space is 4 TB, right?
can you explain more?
Thanks Again!
Last edited:
If you have 8 disks and uses raid 10 (1+0) you will have 4 disks per sub array netting you a 4TB data. (2+2) raid 1 and (2+2) raid striped so 4 TB usable.
If you chose RAID 1 on a 8 HDD hardware raid or a SAN : aggregate disk distribution would probably be (1+1) (1+1)(1+1)(1+1)
so 4TB also.
If you chose RAID 1 on a 8 HDD hardware raid or a SAN : aggregate disk distribution would probably be (1+1) (1+1)(1+1)(1+1)
so 4TB also.
It also depends if you have a big cache with battery of capacitor ... if you don't, get one. It will increase the write performance and can be used to get faster random IO.
i would go with ZFS Raid 10 and you can have 2 disk failure if there are on different mirror but you need to put ACHI Mode rather then RAID hardware
Sorry, you cannot. If the wrong disks fail, it'll fail too. Imagine you have a failed mirror, than nothing can be used. You're only "save" with RAID-6 (or z2) in a random-two-disk failure.
Thats true if the wrong disk fails the odds of that is rare but of course its possible but if your running MSSQL on a vm you might need to hit a raid 10 if not then raid 6 if your scared but end of the day if you got backups its all good
Depends on your performance/reliability requirements, your hardware, what types of content you want to store and what you want to use to create that raid.
For example when using ZFS it won't make much sense to partition it yourself, because you would work with datasets instead. Smilar thing with LVM/LVM-Thin on top of HW raid where it would be better to make use of LVs instead of partitions.
For example when using ZFS it won't make much sense to partition it yourself, because you would work with datasets instead. Smilar thing with LVM/LVM-Thin on top of HW raid where it would be better to make use of LVs instead of partitions.
Last edited:
I'd like to install pve on a ibm server with server raid controller. There are 2 raid structure on hardware raid, one is raid 1 with 2 disks, the other one is raid 5. Now which raid i should select while installing Proxmox VE? I'm close to pick the zfs ones. But which zfs?