I'm the poor schmuck that became VERY practiced at doing recoveries in this domain. We were initially told to move from three E5.5 servers to 1 clustered Dell SAN. My first comment was 'what if this unbreakable setup breaks in a way that the diagnostics on the SAN doesn't detect a problem... I'll end up with 8700 people without e-mail'... I was told "that can't happen"... well, it did... I did at least 6 recoveries in a 12 month span with 2 being all 8700 mailboxes and the others being 4300 each time. This FINALLY convinced the people with money that this wasn't a smart strategy and they finally allowed me to move back to the 'many small servers' approach that I prefer... yes, I take a hit on SIS (8700 mailboxes on currently 6 production servers... I'm backing up 200GB nightly.. full backups)... since our move back to off-the-shelf (but MS HCL hardware) servers, we've had no outages. Those in our org that are still using the Dell/Cluster approach continue to experience constant problems.
John is a little off with his description... this is our current approach and it seems to work very well in my test restores. (this was developed during my nightmare, er experience with the recoveries of last year). We keep an offline restore domain that contains our recovery servers... these same servers are the target for my backups of the production servers... this way in the case of a restore I am running the restore from the local drive on the recovery server as opposed to over the network or from tape (we backup to files on hard drive). We also dump VIP mailboxes to PST files nightly (perl script that uses exmerge... 93 mailboxes) Restore steps with our config... this is based on a couple MS whitepapers... I can dig them up if anyone is interested (don't remember the titles) 1. production db or sg goes down. 2. determine if the db/sg can be remounted without data loss... if not, continue. 3. copy production TLOGS to a safe location on the recovery server for that production server. 4. reset the dbs on the production server so that the people on these dbs now at least have mail service back (empty mailboxes) 5. Start restore of dbs on the recovery server. (making sure not to checkmark 'last backup' or 'mount db'... start exmerge of VIP mailboxes back into the reset mailboxes on the production server. 6. Once the restore completes, copy the production TLOGS into the templog dir and run eseutil /cc against the location that has the restore.env file (templog location).... this now gets your restored db back to the point in time just before the crash. 7. Mount the restored db on the recovery server to make sure all is well... if so, dismount and copy this restored db back to the production server as a different name (priv1.edb.rst for example) 8. Once all restored dbs are copied to the production server, dismount the reset dbs and rename them. (mark them for overwrite)... now rename the restored dbs back to their original names. 9. Mount the restored dbs on the production server.. your users now have their original mailboxes (rules, permissions and all) but are missing the mail that delivered in the time between (4) and (9). 10. Copy the reset dbs to the recovery server and mount them there. 11. Exmerge out the last 24 hours of the mailboxes on the reset dbs (that are now on the recovery server). 12. Exmerge the PSTs from (11) back into the production servers.... recovery complete. jeff e. -----Original Message----- From: Roger Seielstad [mailto:[EMAIL PROTECTED]] Sent: Wednesday, January 08, 2003 2:54 PM To: Exchange Discussions Subject: RE: Exchange 2000 Recovery I'm more interested in how often he has hardware failures. It sounds like a common event! ------------------------------------------------------ Roger D. Seielstad - MCSE Sr. Systems Administrator Inovis - Formerly Harbinger and Extricity Atlanta, GA > -----Original Message----- > From: Chris Scharff [mailto:[EMAIL PROTECTED]] > Sent: Wednesday, January 08, 2003 3:45 PM > To: Exchange Discussions > Subject: Re: Exchange 2000 Recovery > > > Doesn't play hell with your SIS? > > On 1/8/03 13:20, "John W. Luther" <[EMAIL PROTECTED]> wrote: > > > > Hey. > > We have multiple small exchange servers that do their backups > to recovery > servers that have several mirrored drives so no single > production server has > any of its backups on the same drive mirror. With our > database size limit > we have one recovery server for every three mail servers. In > addition we > have at least one "hot spare" mail server. > > When there is an outage we note which folks are affected and > then recreate > their (now empty) mailboxes on the recovery server to get > them back into > email. We then Exmerge the backed-up mail out of the backups > into the new > mailboxes. Some tlog juggling has to be done in order to > recover all mail, > but it is fairly strait forward. > > Each of our servers costs ~6K using "off the shelf" > components. We learned > the value of lots of small servers when our Dell PowerEdge > equipment crapped > out on us repeatedly early last year. > > You could probably do this with three servers, then. One > for production, > one for recovery/backups and one hot spare. Under your limit, though? > Well, I guess that would depend on your shopping ability and > the components > you choose. > > John > > John W. Luther > Systems Administrator > Computing and Information Services > University of Missouri - Rolla > > At 11:02 AM 1/8/2003 -0800, Newsgroups wrote: > >I am not aware of a budget but when I mentioned the solution from > >"Marathon Technologies" they almost fell off their chairs. > I think they > >want to spend somewhere from $3k to $7K (Not sure, as they > have not told > >me anything). I told them that for that price the best > thing they could > >do is have another server and do a daily restore of the > database on that > >box and if the main server dies put up the new one instead. > What do you > >think? Any other ideas? > > > >Thanks > > > >-----Original Message----- > >From: Chris Scharff [mailto:[EMAIL PROTECTED]] > >Posted At: Wednesday, January 08, 2003 10:48 AM > >Posted To: Exchange Newsgroups > >Conversation: Exchange 2000 Recovery > >Subject: Re: Exchange 2000 Recovery > > > >Seamless, transparent, automatic and cheap? Don't believe > such a high > >availability solution exists. Even overspeccing a single box > to ensure > >it's > >fully redundant gets rather expensive on a per user basis > for only 180 > >users. What are the actual requirements surrounding the solution and > >what > >budget has been proposed to implement it? > > > >On 1/8/03 12:27, "Newsgroups" <[EMAIL PROTECTED]> wrote: > > > > > > > >We are looking into different methods of recovery from > Exchange 2000. I > > > >know there are several ways of doing this. We want to be able to > >recover w/ out any user interaction (by that we mean it would be > >transparent to them and they don't want to be down for 4 to > 6 hours). > >We have about 180 users. I know we can cluster them but > they don't want > > > >to go that route because of the cost. Will software or hardware > >replication work and be transparent or are there any other > technologies > >that you may be aware of? > > > > > > > >_________________________________________________________________ > >List posting FAQ: http://www.swinc.com/resource/exch_faq.htm > >Archives: http://www.swynk.com/sitesearch/search.asp > >To unsubscribe: mailto:[EMAIL PROTECTED] > >Exchange List admin: [EMAIL PROTECTED] > > > > > >_________________________________________________________________ > >List posting FAQ: http://www.swinc.com/resource/exch_faq.htm > >Archives: http://www.swynk.com/sitesearch/search.asp > >To unsubscribe: mailto:[EMAIL PROTECTED] > >Exchange List admin: [EMAIL PROTECTED] > > > _________________________________________________________________ > List posting FAQ: http://www.swinc.com/resource/exch_faq.htm > Archives: http://www.swynk.com/sitesearch/search.asp > To unsubscribe: mailto:[EMAIL PROTECTED] > Exchange List admin: [EMAIL PROTECTED] > > > > > > > _________________________________________________________________ > List posting FAQ: http://www.swinc.com/resource/exch_faq.htm > Archives: http://www.swynk.com/sitesearch/search.asp > To unsubscribe: mailto:[EMAIL PROTECTED] > Exchange List admin: [EMAIL PROTECTED] > _________________________________________________________________ List posting FAQ: http://www.swinc.com/resource/exch_faq.htm Archives: http://www.swynk.com/sitesearch/search.asp To unsubscribe: mailto:[EMAIL PROTECTED] Exchange List admin: [EMAIL PROTECTED] _________________________________________________________________ List posting FAQ: http://www.swinc.com/resource/exch_faq.htm Archives: http://www.swynk.com/sitesearch/search.asp To unsubscribe: mailto:[EMAIL PROTECTED] Exchange List admin: [EMAIL PROTECTED]