I ran into a problem where a file system was corrupted beyond repair and wondered if anyone has seen anything like this before or has a reasonable explanation. Here is the scenario:
Linux was running on an Intel X64 system with two local drives mirrored in Linux containing the swap file system. The system was booting off a SAN drive where the rest of Linux was loaded. There were three other SAN LUN's being used that were: 1. 500 GB 2. 150 GB 3. 35 GB (Linux LUN.)
We swapped the system hardware to a different box and changed the HBA to be on the new box so the HBA Bios was still allowing the Boot from SAN. The new box had two local drives but they were mirrored at the hardware level with RAID 1 (so Linux would have only seen one drive drive.) The system was rebooted and crashed with numerous file system corruption errors. The 500 GB LUN on the SAN got severely corrupted on the reboot of Linux from the SAN with the new hardware to the point where it could not be repaired.
What I am wondering is, What caused the system to get corrupted?
Is it possible that the lack of a swap file system mounted would have caused this to happen? Or is it because the /dev devices were not the same as they were in the first configuration?
Any ideas?
Phillip Thayer