On Mon, 8 Jan 2001, Bennett Todd wrote:
But, presumably, you still care about amazing reliability, effortless manageability, snapshots, or some other unique attribute of the NetApp;
Precisely... we still need the reliability, the near-zero maintenance, the ability to centralize our backups, the capability for sharing out the data to both NFS and CIFS clients, the ability to arbitrarily create "filesystems" and instantly resize them up or down, etc. OTOH, we don't need 15000 ops/sec to write out log files. 72GB drives are great for this situation.
unless one or more of 'em are important, it seems to me like the way to go would be to get handful of cheap PC chassis', a short stack of cheap EIDE RAID controllers, and a large pile of the $250/each 80GB EIDE drives.
We have two applications that don't make economic sense for us to put on Netapps: our backup staging servers and our news servers. Both of those use 50GB and 72GB SCSI JBOD... probably about US$15K per terabyte, but still very fast and very reliable. I'd rather not have any IDE drives in my servers if I can help it (and Sun is foiling that plan with their IDE CD-ROM's!) ;-)