Hi,
May i know if anyone have information on the following?
1) How many disks failure can VSAN accommodate?
With maximum number of failures tolerated FTT=3, am i correct to say if disks are the components to fail hypothetically, then maximum number of disk failures =3 at one time?
2) How dynamic can we scale the storage (with local disks) without operational impact/disruptions
Thanks and best regards,
CP
"Failures to tolerate" = how many failures can you tolerate (disk / host or network). So if FTT is set to 3 then you can tolerate 3 failures.
With other words - three hosts participating in vsan storage for that vm or that ftt=3-protected object can crash and your data will still be alife.
Regarding your second question: VSAN is not working like common RAID solutions, there is no parity distributed and written among certain disks or disk groups. Let´s try it this way: VSAN is "object oriented". If you add storage to a participating vsan-activated-cluster member the disk is there and it´s all good - no need to integrate it into some parity tree object or simliar. It is simply formatted with vmfs-l, 750MB-1GB subtract for metadata, and that´s it. No production impact/disruption.
Now: If you are not sure and you are kind of extreme cautious, you can always fully evacuate a host manually by placing it into maintenance mode and select full migration of all data.
Best regards,
Joerg