I have been setting up an old Prtoliant ML310 G1 with Debian Squeeze using a four disk RAID 1+0 array as storage
We had a few hiccups in the beginning, but seem to be nearing what resembles a stable system: every time we rebooted the system, the smart array 6402 controller would warn us that one of the drives (#4) required a rebuild, and requested that we allow automatic recovery process to kick in. Which we did, each and every time.
Since all hardware was second hand, we assumed that one or more components wasn't up to par. So we had them replaced : the hotplug backplane, the smart array controller, and of course the incriminated disk
There isn't much in terms of data on the drives since all we've done so far is install Debian, and a basic email system composed of Postfix, Dovecot, and Webmin. The server is still idle in our workshop
In the beginning I wasn't too observant of the logical drive's rebuild status before rebooting. Then I began to monitor it using HPACUCLI
The rebuiild process doesn't seem clear to me. Sometimes the progress seems linear and at other times it looks as if it keeps restarting from the beginning at different moments throughout the day.
I finally loaded the controller's Option Rom Confiuration Utility (F8) at boot and allowed the rebuild process to complete. Which it it did. The system rebooted normally, without any error or warning. Great !
Twelve hours later I decided to reboot the system (shutdown -r now) and lo and behold the smart array warning is back, same drive, same error.
So my question is: must I do anything special before rebooting the server in order to allow the logical drive to remain in proper shape during a reboot ?
I can obviously replace the 4th drive once again, who knows, maybe I still have a duff drive in the system.