Re: Question: ONE corrupted tablespace in standby db???

> Accidents happen - no one is perfect - not even the senior dba's :-) I
> believe 30% of all data loss or downtime is caused by human "errors" - the
> machines will take over eventually :-)

Yup, accidents happen - so the more one can do to mitigate them, the better.

I was involved in a job this past weekend which involved the reorganization of the disks of the SAN on which one of our production dbs sits.

I had it on all assurances from the SAN guys that the process of relocating volumes to new disk arrays was perfectly safe, and that he was perfectly happy to do it live, while the db was up and in operation.

Mmmmnah. My view is that these great bits of hardware, theoretically bullet-proof though they may be, are not a justification to be cavalier. So, I insisted that the work be done on a weekend, when the db could be taken down. Took a cold backup, made 2 copies of said backup, planned for a full dbv once the work was done, etc.

I asked the guy for a verification of the incremental tape backup before the disk reorg started - he said "Are you sure you need that?" Erf.

In my mind, a small bit of paranoia is a healthy thing, if it means you've covered your butt just that little bit more.

As it happened, the SAN reorg went fine. But I still believe in healthy paranoia when it comes to data security. Especially since it'll be me hung up by my b@!!$ if things go south. ;) Received on Mon Jun 19 2006 - 15:50:01 CDT