[CentOS] Error reading block x (Attempt to read block...)

Wed Mar 5 10:45:05 UTC 2008
Scott R. Ehrlich <scott at MIT.EDU>

A Centos 5 64-bit server reflected some SCSI kernel errors in the logs, so 
I opted to reboot, only to be prompted to perform a manual fsck, that an 
automatic fsck wouldn't be permitted.

As a result, the fsck is taking a long time, as I have to babysit the 
server for every bad block that is reported and needs a forced rewrite.

The exact error is "Error reading block x (Attempt to read block from 
filesystem resulted in short read) while doing inode scan.   Ignore 

Force rewrite(y)?"

I say yes to both.

Could I safely get away with booting from a CentOS 5 CD and performing a 
fsck with automatic system fixing to speed up the process, or do I need to 
be patient and just keep manually selecting Y for as long as it takes?