Hi All, We run a very antiquated Oracle 8i infrastructure that will get replaced by Postgres Real Soon Now(tm). Or so they said 5 years ago. Anyway, here's my problem:
We have two sites, Alpha, the production site, and Beta, a warm standby. We're doing transaction log shipping from the production instance at Alpha to the standby instance at Alpha, the standby instance at Beta, and off-site for backups. The database is around half a TB, so it's not something that we can sync across the WAN. The working disaster recovery plan involves activating the standby instance at Alpha (if possible) or Beta (if a site-wide event occurs) in the event that the production instance fails/dies/whatever. Once this happens, a new standby instance must be made, because the transaction logs reset, and because of that, our plan calls for immediately shutting down the database, copying the image to an external drive, and trucking it to the new standby site, wherever that is deemed to be. This way we can recover from a database loss at the secondary site. This is not convenient, but I'm willing to forgo convenience in a situation where our primary database has crashed and we're dealing with the fallout. As it happens, though, the powers that be are finally planning on a "hot" test of our secondary site, meaning Beta will be running production for a week. The inconvenient part of this is that when we initially switch to Beta, I'll have to make an image, throw it on USB disk, and run it to Alpha. When we switch back to Alpha after a week of testing, I'll have to do the same thing, but in reverse. This is more than cumbersome, it's damned irritating. Today, I came up with an idea that I think could eliminate the running back and forth. My plan for the switch is to ensure that the standby instances at Alpha and Beta are 100% current and have processed all transaction logs. I'm then going to shut down the Alpha standby, make an image of it for insurance, then I'm going to turn it back on, then activate both of them. At that point, I'll shut down the instance at Alpha, use the (now production) instance at Beta to create a standby control file, sync it to the Alpha instance, then I'll start up the Alpha instance in standby mode, which should be able to process the transaction logs from Beta. My question is, should my plan work? If two oracle instances are identical in standby mode, and are then activated, are they still identical? Will I be able to convert one of them to standby and reply the log files from the other? I appeal to those of you who know Oracle more than myself. Thanks! --Matt -- LITTLE GIRL: But which cookie will you eat FIRST? COOKIE MONSTER: Me think you have misconception of cookie-eating process. _______________________________________________ Tech mailing list Tech@lists.lopsa.org http://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
