I'm trying to convince a colleague of mine that having a single voting file/disk on an external redundancy ASM disk group is a bad idea since a logical corruption would mean a painful outage and recovery operation. There is also the possibility that during certain cluster failures there would always be only a single available cluster node since there is only a single voting disk (so no majority voting to decide on the active cluster members). I would greatly appreciate some feedback on further arguments as well as the validity of mine own using the term 'logical' corruption and the voting to decide on the active cluster members.
I appreciate your time. Thanks in advance!
as one argument will be it's a waste of space, ask 5 for 1 GB lun, and create during the install a disk group called GRID, high redundancy, 5 members.
in case of corruption, having several copies, your cluster will start.
without that, you will have an extended outage that could have been avoided.
Oracle recommended for OCR files and VOTE file/disks ASM redundancy is HIGH with 5 way.
I think you must create a ASM disk group for example OCRVOTE, with HIGH redundancy.
And replace vote disk to new ASM diskgroup, and and change OCR files location to new ASM diskgroup.
Mahir M. Quluzade