Seagate ST2000DL003-9VT166

ndroberts1 wrote:

All. I ended up with an Ultra 6 and 8 of these drives with model # suffix 301 & FW CC32 During a support call for an attempted RMA seagate simply said that the drive is not supported in a raid config. (And its not their problem to fix) It's a desktop drive. Not a NAS drive. There are good articles on the net that covers TLER and RTS - in simplistic terms the drive times out and the raid config SW/HW marks it as dead. It times out as the drives internals try and recover bad blocks and this takes a bit of time. With 8 drives the rate of bad blocks recovery is surprising. I have had 3 double drive failure in 5 weeks. Swap them out - only to find others dead --- level of instability is scary to say the least with precious family home videos now lost as a result. I read a few threads on adjusting drive timeout values in Linux - noting that no one claims success & not having much else to do I applied the method and have flogged the unit for 2 weeks now with no dead drives. It seems pretty stable. I cant say if its just luck or if it actually works. Time will tell- if you would like to try it - here it is. After a reboot (can be automated via /etc/rc.local), I run the following commands:- echo 300 > /sys/block/sda/device/timeout echo 300 > /sys/block/sdb/device/timeout xxxx Default is 30 - I do this for every drive in the raid group. In the Ultra 6 it ends up being for sda, sdb, ..... sdf Give t a try..... If it continues to works ... it saves me going and getting WD red NAS drives.

Intesting, I like to hear how stable this is after a few months. Are there any caveats to changing the value above? Unfortunately, I already bought 2 WD 3TB RED drives and yanked the Seagates from slots 1 & 2. However, I still would be interested how this fix works long term since the Seagate drives may end up back in the NAS should a disk failure occur again.