Re: Block allocation failed

[Graham Murray <graham@gmurray.org.uk>]

Theodore Tso <tytso@mit.edu> writes:

>> It was about 90s after the message showing the filesystem mounted
>
> Hmm, the most likely cause for that would be if the block group
> descriptors had an incorrect number of free blocks.  But you had just
> run e2fsck -f.

No, sorry I was obviously not clear as to the order of things. First the
system rebooted itself a few seconds after writing an oops (spinlock bad
magic) to syslog and subitting it to kerneloops. On the reboot, the
syslog showed a normal (ie no error and no recovery) mount followed
about 90s later by the error dump I posted. After this, I unmounted the
filesystem and ran 'e2fsck /dev/sdb3' which just replayed the journal
and said the fs was clean. I then ran 'e2fsck -f' which made several
prompts for problems.