We recently bought an 7410 Cluster and are very happy.

Blows away everything that NetApp and EMC had to offer by far! (price and features).

I am very happy about the movement of iSCSI target mode into kernel from user land!

Being able to share ZFS, NFS, CIFS, and iSCSI all from kernel offers a very strong benefit to using OpenSolaris:

- Having all of the major protocols integrated into the kernel with ZFS means that applications running on the server as well as different distributed OS's over varied protocols can all share the same data set without corrupting access lists and permissions.

- Placing ZFS on the same server that has kernel based NFS, CIFS, and iSCSI also means virtually unlimited file systems can be shared... virtually unlimited LUN sizes is a beautiful thing, especially when minimum file sizes are tiny in comparison to most other server competitors (i.e. 1K instead of 43K, 64K, 1M, or 2M, etc.)

- The doors are swung wide open for shared-nothing failover clusters on 2 (and now 3 servers) is made possible by integrating ZFS, 2-way mirroring, 3-way mirroring, and iSCSI COMSTAR

http://netmgt.blogspot.com/2009/08/multi-node-cluster-shared-nothing.html

- Moving applications into Zones also make failing over applications scriptable, inexpensive, and easy - avoiding costs of H-A kits for applications and hypervisors for virtualization.

Just to think that these capabilities are free under OpenSolaris and will be rolled into Solaris 11 sometime soon just makes me extremely pleased!

Somewhat related. Read a blog showing how seriously the Fishwork's team takes performance on this product line. Always nice when their improvements to Solaris get back ported to the enterprise distro for our servers.

http://blogs.sun.com/brendan/entry/7410_hardware_update_and_analyzing

Is this the last act of desperation before Larry gets full control and just brings in Pillar for some real storage hardware?