To pull off two really amazing conferences in a row is a wonderful indication of the attention to detail that Dell is taking with EqualLogic. Here is my first conference review
And now on to a review of the second conference. I’ll start with the bad. The weather. What is up with flying to San Diego and getting worse weather than in Seattle…
Thoughts\insights\advice I had\recieved during the sessions in no particular order.
Replication. Replication is a tricky subject. Some good advice I received: If possible use an applications native syncing\replication first before attempting an array replication. Some examples, use exchange 2010 DAG, and for SQL use Log Shipping. These applications will know exactly what data has changed in replicate it probably in the most efficient manner.
Riverbed seems to be (possibly) the replication box of choice (this is base entirely on feel, and no hard facts) they are probably also the most proud of their boxes….
Getting the network right is still probably the most complicated part of setting up your EqualLogic. Jumbo frames should be enabled after they are found to be needed (I seem to remember 4k Iops being thrown out there). Stacking is definitely recommended, I would also recommend testing out your switch failure as part of your install. New firmware for the powerconnect should improve the failover rate, as well as in the future being able to do a rolling update of the switches! (One of the main reasons people don’t like stacking I hear) also setting up two stacks and then LAGing them together is an interesting option.
But, pay close attention to your networking setup. It can make or break your EqualLogic install!
Another topic of interest was the three different ways you can attach to your EqualLogic storage (in reference to vmware) Basically, it really comes down to choosing your approach. For example, if you go the iScsi initiator from within your vm you get to take advantage of the really great EqualLogic snapshot\VSS tools for exchange, filesystems, and sql. In the case of a DR situation though, it makes things a little more complicated because you will need to make sure that the volumes that initiator is connecting too has been replicated.
A ‘normal’ vmfs volume has the advantage of being self contained (but is limited to the 2TB vmfs limit) I like this method because then I can use Veeam for my backup. Speaking of Veeam, I really hope that someday that there will some integration between Veeam and EqualLogic. This would involve scripting vsphere to take a snapshot, then the equalLogic to take a snapshot, then alter the ACL of that snapshot so the veeam box can access it, rescan, connect to the snapshot volume and then proceed with the backup! This would be a very nice method! Please, please, pretty please! Veeam and EqualLogic!
The VDI sessions are always intriguing! Virtual desktops are definitely in the future… someday when we can afford them!
SSD sessions are interesting, would love to dump our ERP database on it and take it for a spin. They also talked about putting tempdb on its on RAID 10 volume… hmmm I wonder how much that would help our environment.
All in all, a really great conference! Good speakers, great to have engineers there listening to our ideas (even better, they actually implement some of them)
Some of my recent ease of use complaints were neatly addressed during the NDA session. All I can say is, I’m glad we got on the EqualLogic train, cause it’s full steam ahead.
The dinner on the USS Midway was epic (what a great place for a bunch of geeks to totally geek out! everyone was walking around in reverence)
A tweetout from Michael Dell to all of the social media folks attending #eqlconf was entirely unexpected and greatly appreciated!