Re: Help me tuning this wait event:log file sync

Yong Huang wrote:
>
> The original poster says he's using Veritas Quick I/O. We should
> assume data redundancy is already done at the lower level. I always
> propose 1 log member at the company I'm working at, because we either
> have Veritas or (on non-production servers) Solaris Disk Suite. Using
> 2 or more log members used to be prevalent.

It still is. Online redos are so important that most business critical systems I know of mirror at the Oracle level (multiple members per group) and at the disk level (multiple mirrors per member). They each protect against different failures - the former against more logical-type errors (e.g. corruption caused by the OS/disk-subsytem, etc.), the latter against hardware failure. Your disk mirroring mechanism (whether hardware- or software-based), like anything else in the computer world, is not immune to bugs.

If I had to pick one method though, it'd be mirroring at the Oracle level to members on non-mirrored, but physically separate disks. That way when someone goes crazy with the rm command, your database doesn't crash. Not that that's ever happened, of course. ;)

Regards,
Sean Received on Tue Jul 16 2002 - 18:46:06 CDT