I was reviewing some end-systems in our lab environment the other day and came across something I found both amusing and impressive. Lab environments sometimes suffer from “The Carpenter’s House” syndrome, where the premium “white gloves” treatment is done when the carpenter is working with his clients; but at the expense of letting his own house get a bit run-down due to lack of resources and time.
Our lab environment has a data backup server that is using a direct-attached iSCSI SAN (A NexSAN SATABeast 2) as its backup target. Since it is direct attach, the management interface was only reachable via the backup server’s internet browser on an otherwise non-routable subnet. That system is not one that I log in to very often for reasons other than downloading updates and patches here and there, and configuring test backup jobs when I need to see if something will work for a customer. On a whim, I decided to do a health check on the SAN to make sure the system was not showing any kind of hardware errors. After all, I could not remember the last time I logged into the SAN.
What I saw was pretty impressive. The entire time these controllers have been online there has been a nonstop schedule of backup jobs running on this SAN for various systems and various applications. Fulls, incrementals, synthetic fulls, SQL transaction log backups, VMWare snapshots…you name it.
Now that’s uptime!
Wow! We’re approaching 3 years without a controller reboot with no signs of a problem. There were no hardware errors. The fans, drives, and controllers in this SAN have been rock solid. Do keep in mind that this level of hardware up-time is unusual in any kind of production environment. I highly recommend staying up to date on firmware & patches, which will require a rolling reboot of each controller (and subsequently will reset those up-time timers.) This particular SAN is a discontinued model so no new firmware has been released for it in years. It is also segregated from the rest of the network via being direct attach iSCSI to the backup server on non-routable subnets that the rest of the test lab network cannot access. This both minimizes the attack surface of the system and prevents cross-chatter from the data LAN bleeding over to the iSCSI network and causing performance problems.
If you are considering a storage solution for backup disk, file server storage, streaming video repository, or other capacity-oriented workload consider contacting Chi Corporation to hear about our NexSAN offerings. You too can experience this level of system stability in your SAN at an extremely affordable cost per Terabyte!
