joeware - never stop exploring... :)

I know some of you may read that and choke on your bagel or candy bar or coke or whatever else you were shoving down your throat as you relaxed and sat back to read what I have to say today…

But I mean it!

I had an unfortunate issue where one of my virtual host servers started having issues with one of the Western Digital drives in it. This particular drive was in a MOBO RAID (Promise RAID) "stripe" set.  This caused my server to act quite erratically and black screen (like a blue screen of death but worse because you have nothing to go on) and generally perform poorly. After a series of tests I worked out what happened. Unfortunately both my Exchange Server and my mail client machines were virtuals on that one host… So I lost Exchange and all of my PSTs in one fell swoop. I was devastated as you may imagine. Well you may not imagine it all… When I moved, my backup system "broke" and I never reset it all up. It was one of those, "I will get to it" items. Well busy schedule and Carbon Monoxide poisoning prevented me from worrying about it and voila, several months later, here I am with broken pieces and a need for a backup but that backup was almost 6 months old…

Long story slightly shortened, the Promise RAID "Stripe" set apparently wasn’t a stripe set because I noticed while trying to copy files off, some worked fine, some didn’t work as fine and actually hung the box. So I worked to get as much off as possible and soon realized that the symptoms added up to the "Stripe" set where the data is interspersed on both drives equally in stripes (hence the brilliant name) was actually implemented as a volume set… I.E. The info was written to one drive and then slowly crawled onto the second. Luckily, 3 DCs and my client’s Data logical virtual drives were all on the "good" disk and the Exchange server, one DC, and my client’s OS disk logical virtual drives were on the "bad" disk. Using the command line virtual disk mounting tool in Virtual Server I mounted my client data virtual disk and immediately copied the gigs and gigs of PSTs off and to about 10 different other machines. 😉

So only thing left was to rebuild a DC and the Exchange Server… Well I wasn’t relishing the Exchange Server rebuild as I had a lot of custom settings… Well my good friend and coworker and co-MVP buddy old pal Brian Desmond casually said in IM… why don’t you just use "setup /disasterrecovery" (dumbass)[1]. I was like… hmmm never heard of it, my Exchange servers don’t do that normally. Let me try it…

WOW. Very cool. I was very impressed. Worked like a dream. I was, at that moment, thinking, hey the Exchange team did something right here. I was quite happy as all special configs I had were all in place and bam things just worked. Very good.

Of course if I look on this in the slightly negative way I could say… why is it that one of the best implemented features I have found to date was the disaster recovery option??? Have to run that much around the world?  I will just assume they thought it was exceedingly important to get right and exceedingly easy to do because all of the data was in AD already. I am glad it ran that well. If it didn’t, it is quite likely I would be running a FreeBSD mail server right now because honestly, Exchange is running in my house for two reasons. First, it is for testing things to see how it impacts AD and of course for my ExchMbx utility and second, I need a SMTP/POP3 landing zone server and that is all it is.

BTW, some folks I consider to be  good friends now are over on the Exchange Dev team and I know we will see amazing things because of it. I look forward to when their influence is helping us all out.

    joe

[1] Brian didn’t say dumbass, I just inferred it from how it was said. 😉