Karl Cunningham wrote:
I did have one (only one) case where it cried wolf. In the days of 512M drives, on a month-old drive smart complained. We took the drive out of service but ran read-write tests on it for about two days. Smart stopped complaining after a few power cycles. After running more tests we put the drive back in service, and it worked flawlessly for several years until replaced to increase capacity.
wow, you must have had one of the late-model 512M drives. SMART was fairly new back in those days; it became a standard in 1995. Before that it was IntelliSafe (and only available on conner, seagate, and quantum disks). Maxtor now owns Conner and Quantum, for what it's worth.
It still only predicts about 30% of failures. I don't put much faith into it. Sure, any sort of proactive failure analysis is great, but i don't trust it.
That, and I've seen PFA go haywire on perfectly working disks, to the point where I've had to disable smartd for that disk in order to keep my logs from filling with garbage. I've even got some disks right now that either don't support SMART, or their support thereof is horribly broken. I have SMART off on this box for that very reason :(
anyhow, back to work. -kelsey -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
