Unsolved
9 Posts
0
1989
Scaleio Disk replaced by hot spare
Hi all,
I have a silly question, I am testing a scaleio deployed on esxi hosts, they are all dell servers.
As I only used some of the installed disks for the scaleio deployment, I just wonder what would happen if I set the rest of the unused disks as "hot spare" on Dell IDRAC platform.
Because when deploying scaleio, I need to select the devices to be used. In case of a disk failure, a hot spare would replace the failed disk automatically haredware-wise, but would the replaced disks be used in scaleio and initiate a rebuild automatically?
Thank you very much
SanjeevMalhotra
138 Posts
0
August 30th, 2018 06:00
Anything that slows the IO on any disk, shouldn't be used.
In normal scenario, if a disk is to be replaced (planned), the data from that disk is re balanced to all disks (one to many). If the disk suddenly fails (unplanned), data on the disk is rebuild on the other SDSs (one to many) so it is fast.
Hot spare may not be that fast. A small sized disk may also take several hours to move the data from the failed disk to the hotspare. ScaleIO will not like this latency. Hence we shouldnt be using it.
If i am wrong, experts may correct me.
pmng
9 Posts
0
August 31st, 2018 02:00
You are right, scaleio menu states that using hardware raid-ed disks is not recommended,
it may not a good idea to do that