always_exit_on_disk_error=enable always best practice

[Burt Rochon]

should set always_exit_on_disk_error=enable. We had a situation last week where there was an issue with the mount which caused a write error, so the active server immediately exited and failed over, but since the issue was with the mount, the second server also exited and the instance failed since the first server had already exited and there wasn't a server to fail over to. Is there anything we can do besids, or in addition to, setting always_exit_on_disk_error=disable

[Emmanuel Marchiset 2]

Hi, In a context where NFS is used to manage the shared file system supporting an EMS FT configuration I would recommend to keep always_exit_on_disk_error=enable and to put in place a mechanism to restart EMS instances automatically when they fail (can be done easily if EMS instances are managed has services in Linux or Windows).

In NFS context setting always_exit_on_disk_error=disable may lead to datastore corruption and loss of data, so at the end you have to do a trade off between availability and integrity.

I hope this helps

EM