whoop
March 3rd, 2010, 01:50 AM
I tested the reliability of a software raid 1 array "quite drastically" via formatting one of the partitions in an array (with a live cd) and then rebooting the machine. After starting the system with the degraded array and adding the formatted partition back to the array everything started working again.
So, I think I have two questions:
*I can understand that mdadm would figure out something was wrong with a disk if (smart for instance) was reporting it was failing, but when an array is out of sync (one file is slightly different for instance) how does mdadm know which partition is "wrong".
*If you have selected not to boot from a degraded array (during install of ubuntu server), how do you resolve the problem? I have only done this via booting the degraded array but I have read that this could lead to data loss in some occasions.
So, I think I have two questions:
*I can understand that mdadm would figure out something was wrong with a disk if (smart for instance) was reporting it was failing, but when an array is out of sync (one file is slightly different for instance) how does mdadm know which partition is "wrong".
*If you have selected not to boot from a degraded array (during install of ubuntu server), how do you resolve the problem? I have only done this via booting the degraded array but I have read that this could lead to data loss in some occasions.