Disaster recovery

Right now we have 2 machines of same configuration in a veritas cluster, one being the primary and other being hot standby for primary box. Besides this, we have raid 5 disk array hooked on to primary box and it gets export/imported to hot standby when primary box fails over (veritas clustering), redundant controller/power supply etc. etc. so we have a HA with a resilience of 5-10mins downtime if either box fails. (8.1.7.4 / solaris 8.0)

However, the above strategy still exposes failures from media corruption (via controller or due to disk array), problems with veritas cluster, catastrophe in data facility, etc. etc.

I know remote standby databases are the ones that comes to mind immediately (dataguard, when we get 9i), what approach one would take to implement this.

i'm kind of leaning towards having a remote standby with manual recovery mode and have archive logs ship to remote site via rsync/secure, as and when it generates (every half hour) and apply archive logs to standby once in 6hrs via scripts.

any comments are appreciated.

thanks Received on Wed Feb 05 2003 - 21:19:42 CST