3,870 reputation
817
bio website brentozar.com
location Chicago, IL
age 40
visits member for 5 years, 5 months
seen Sep 17 at 14:44

I make SQL Servers faster and more reliable. I'm a Microsoft Certified Master and Microsoft MVP.

Stuff I'm proud of:


Apr
30
awarded  Yearling
Mar
31
awarded  Nice Answer
Apr
30
awarded  Yearling
Feb
15
answered Dealing with Stretch (Geo) Cluster Node Failures
Apr
30
awarded  Yearling
Mar
16
comment How do I get SQL Server to release its memory?
@Massimo Can you point to some documentation on that?
Aug
13
comment Help me make sense:VLAN, Hyper-V, Intel VMLB
Make sure you install the Intel drivers. The default ones from Microsoft don't include VLAN tagging as a config option.
May
1
awarded  Yearling
Apr
6
answered Restoring SQL Server 2008 R2
Mar
11
awarded  Revival
Feb
23
comment Query Performance Degrades with High Number of Logical Reads
Yep, depending on how SQL Server and the T-SQL statement works, even dynamic SQL can get a cached plan.
Feb
23
answered Query Performance Degrades with High Number of Logical Reads
Feb
18
comment HA suggestions for geographically separated sites
And yeah, that mirroring connection string trick comes in handy, heh.
Feb
18
comment HA suggestions for geographically separated sites
OK, if it's more important to have availability than have it intact, I'd go the bidirectional replication route, but just know that there are a lot of ways it can break. Fortunately, it usually breaks "up" - both sides still accept reads & writes while it's broken, but the changes just aren't making it to the other node.
Feb
17
comment What is the standard method to keep track of the latest Cumulative updates for SQL Server?
Wow, that's nice. And to know when it changes, you can use WatchThatPage.com. It's got a crappy UI, but it works.
Feb
16
comment HA suggestions for geographically separated sites
Allan - why wouldn't replication combined with a geographic load balancer (like an F5 Big-IP) work? Clients connect to a DNS name that points to the Big-IP, and that device automatically routes them to whatever server is up. That's totally transparent to the clients.
Feb
16
comment HA suggestions for geographically separated sites
@Blootac - just to be clear, you're saying you can't tolerate 2ms of query slowdown, but you can tolerate 5 minutes of data loss? That seems a little odd, but in that case, log shipping is probably the easiest solution. Set up the client's connection strings to use the database mirroring failover partner option, and they'll automatically try to connect to the secondary server. You'll need to script something to do a RESTORE WITH RECOVERY on the secondary server if the primary has been down more than a couple of minutes.
Feb
15
answered HA suggestions for geographically separated sites
Feb
10
awarded  Citizen Patrol
Feb
9
answered Netapp snap mgr for sql vs. native backups