Comments on: Democratizing Storage http://blogs.digitar.com/jjww/2008/04/democratizing-storage/ thoughts & musings Sat, 24 Jul 2010 02:26:30 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Erek http://blogs.digitar.com/jjww/2008/04/democratizing-storage/comment-page-1/#comment-397 Erek Sat, 10 May 2008 19:08:10 +0000 #comment-397 Thanks so much for the excellent post on your switch to an open SAN. I'm monitoring the developments with Sun's storage, and hoping to use it for an upcoming project.<br /> <br /> The main sticking point for me is the failover between the primary and secondary storage arrays. You mentioned here that you do application-level failover.<br /> <br /> Can you elaborate a bit more on your failover scheme: How the two thumpers synchronize with each other, and how a failure is detected and a failover takes place? Thanks so much for the excellent post on your switch to an open SAN. I'm monitoring the developments with Sun's storage, and hoping to use it for an upcoming project.

The main sticking point for me is the failover between the primary and secondary storage arrays. You mentioned here that you do application-level failover.

Can you elaborate a bit more on your failover scheme: How the two thumpers synchronize with each other, and how a failure is detected and a failover takes place?

]]> By: Paul http://blogs.digitar.com/jjww/2008/04/democratizing-storage/comment-page-1/#comment-396 Paul Mon, 28 Apr 2008 21:16:12 +0000 #comment-396 ZFS rocks. We build Solaris's virtual zones each on it's own zfs. Then we send snapshots to a remote location, recieve on the other end to replicate the servers for backup. And since the "-F" option, we can do incremental snapshots and receive those on the remote server to save lots of bandwidth. And I wrote scripts that automate the whole process. And it's all done live. We have not had a single data corruption since switching to ZFS. This was not the case on ZFS and Metadb. ZFS is simply the best. ZFS rocks. We build Solaris's virtual zones each on it's own zfs. Then we send snapshots to a remote location, recieve on the other end to replicate the servers for backup. And since the "-F" option, we can do incremental snapshots and receive those on the remote server to save lots of bandwidth. And I wrote scripts that automate the whole process. And it's all done live. We have not had a single data corruption since switching to ZFS. This was not the case on ZFS and Metadb. ZFS is simply the best.

]]> By: Anantha http://blogs.digitar.com/jjww/2008/04/democratizing-storage/comment-page-1/#comment-395 Anantha Mon, 28 Apr 2008 14:41:30 +0000 #comment-395 Dude! I can vouch for the coolness of ZFS. I was an early adopter way back in September 2006. Look at my work described in Slide 17 of the "ZEBRA, ZAMBONI, ZEN & THE ART OF ZFS" paper presented at SAS Global Forum (thanks to Maureen for incorporating my work into her presentation). I also had a situation wherein we had silent in-flight corruption of data in the FC due to firmware issues. Good that I had ZFS to tell me I had bad data; Bad that I didn't mirror (or RAIDZ) the drive relying upon the SAN to provide me that cover.<br /> <br /> All I want now is lots of drives and a good connection, 'nuff of the SAN features. I'll use ZFS to deal with my needs. Dude! I can vouch for the coolness of ZFS. I was an early adopter way back in September 2006. Look at my work described in Slide 17 of the "ZEBRA, ZAMBONI, ZEN & THE ART OF ZFS" paper presented at SAS Global Forum (thanks to Maureen for incorporating my work into her presentation). I also had a situation wherein we had silent in-flight corruption of data in the FC due to firmware issues. Good that I had ZFS to tell me I had bad data; Bad that I didn't mirror (or RAIDZ) the drive relying upon the SAN to provide me that cover.

All I want now is lots of drives and a good connection, 'nuff of the SAN features. I'll use ZFS to deal with my needs.

]]>