Referencing: Odd INode problem on OSR505
This is an interesting post about someone who had a directory entry pointing at an unallocated inode. Usually those things get fixed by the system itself when it reboots, but somehow this one didn't. Or maybe it just wasn't rebooted.
The problem, as Bela Lubkin points out in his usual excellent manner, is that the original poster was chasing his own tail, just moving the problem from file to file as new files happened to use the inode number that originally had been left in the directory. The poster would remove the file, but that just decremented the inode count, leaving him with another directory entry pointing at nothing. Problem repeats, apparently randomly moving around the file system.
As Bela points out, clri is designed to deal with this kind of problem, but it can't quite.. and you need to immediately run fsck after using it. The original poster finds it hard to schedule downtime (those darn unreasonable users who expect to USE the machine!), but there is no way around it.
I'm amused by the fact that the users and apparently the brass have expectations that this box will stay running, period. Imagine trying to enforce that on a Windows server!
Got something to add? Send me email.
More Articles by Tony Lawrence © 2011-03-12 Tony Lawrence