[plug] Dying Samba Server

[Adrian Chadd]

I really would suggest using software raid. Combine software
raid with SMART monitoring to notice disk temperature, error rates,
errorlog, etc, and you're set.

Modern IDE disks seem quite hell-bent on telling you when they're
going to fail. Its nice to receive an email saying "disk logged
errors; failure in 5 days" which is another hint to replace the
disk (or tell software RAID to use the standby disk even before
it knows there's been a hard error) before it really -does- fail. :P

(out of interest: who here watches or graphs the temperature of their disks?)

[root at hosting-2 ~]# smartctl -A /dev/hda | grep Temp
194 Temperature_Celsius     0x0022   030   054   000    Old_age   Always       -       30
[root at hosting-2 ~]# smartctl -A /dev/hdc | grep Temp
194 Temperature_Celsius     0x0022   035   048   000    Old_age   Always       -       35
[root at hosting-2 ~]# 




Adrian