I know I'm a bit late, but I figured I'd add a few comments.
Yes, once a drive starts acting weird migrating off it ASAP is the best course of action. Preferably in order of your most important data first. Smartd and related can help give you an early warning, to do a long test run the following: smartctl -d ata /dev/sda -t long Then after the time it mentions (often 1 to 2 hours) run: smartctl -d ata /dev/sda -l selftest Hopefully you will see something like: Num Test_Description Status Remaining LifeTime(hours) LBA_of_first_error # 1 Extended offline Completed without error 00% 11098 -i and -a can give interesting info, especially: SMART overall-health self-assessment test result: PASSED Seagate bought maxtor awhile back, I suspect the differences these days are mostly marketing and branding, not actual hardware differences. In general I suggest at least considering the enterprise level drives, better testing, better vibration resistance, they handle heat better, sometimes have better warranty, etc. When I was looking at 300GB drives the price difference for the maxtor maxline (enterprise version) vs the maxtor diamondmax (consumer drive) was $110 vs $100 or so. This is doubly so if your building RAIDs out of them, I've heard many stories about RAIDs having multiple failures, often tracked down to using consumer drivers or too small a power supply. Once you migrate off the disk then I'd use the manufacturers diagnostics to destructively test the entire drive, usually sector remapping only happens on writes, so if you hear strange noises while accessing particular drives you likely have sectors failing. Once you do that I'd run the diagnositics a few more times and if you continue to get errors pitch it (or RMA it), if not, I'd try it out for awhile without putting anything too important on it. I've had drives last for years after this with zero additional errors. I'd also watch the reallocated_sector_count, if it keeps increasing that is a very bad sign. Best of luck, buy good drives, keep them cool, and of course keep backups of anything important. Even the best preventative measures can't predict the occasional total failure. _______________________________________________ vox-tech mailing list [email protected] http://lists.lugod.org/mailman/listinfo/vox-tech
