May 30, 2013 at 10:09 am
I've recently started the adventures with SQL 2012. I work in a fairly large shop with over 100 servers, including several clusters, including one 3-node N+1 cluster with 2 instances, one with seven clustered instances on it, and another with a 6 TB "warehouse" (our developers call it that... it's not designed as one though). I'd heard one awesome thing today that I wasn't aware of yet, that tempdb can be housed on the local nodes of the cluster instead of on a shared drive. Awesome because it means I can stick a couple SSD's in the nodes and put tempdb there, and not have it beating up our SAN.
I was looking a little more into this when I spied:
http://msdn.microsoft.com/en-us/library/hh759341.aspx
Apparently... and I don't know if it applies to a cluster or not, don't really care yet... the system databases, including master, can now be housed on a UNC share. I... I'm not sure what to think of that. Awesome or catastrophically dangerous?
Once upon a time I had a really old NetApp, and everything had to go via shares. And with a handy little trace flag, I could put *user* databases on it. And I wanted to. But if someone had told me I could also put the system databases on it I would've asked them what they were smoking. Now that MS is actually supporting doing it though... I have to ask, is there any good reason I might want to? Or do I need to smoke something? 😉
Does anyone have a good reason they'd want to put system databases on a share?
Viewing 0 posts
You must be logged in to reply to this topic. Login to reply