Filesystem went crazy! 1

[tristan308]

A filesystem was growing rapidly. Added a gig of space. After a few days it ate this up. Issued the chfs and it came back with bad inode map and block map. Unmounted file system and remounted, which ran chfs, it fell to 5% utilization. WOW.

Question:

1) Why did it do this?
2) Tried to take this gig away from filesystem, but it said cannot reduce file system. Do I need to unmount?

 

[vickersdc]

You may get problems like this when an active file is removed from the filesystem.

In this case the blocks allocated to the file remain allocated and are not freed up. This can happen with a runaway process writing to a file and the file is then subsequently removed (in the thinking that it will free up space),

To reclaim the blocks allocated without having to stop the process thats writing to it you can redirect the output of another command to it....

date > myFile

...this will write the info to the file, but more importantly, reclaims the space back.

Now I don't know if this applies to your situation, but it's sure worth remembering for the future.

Cheers,
Dave V.

 

[tristan308]

Thanks. That helps a lot. I think that could be the problem when .log files are removed without cycleing the servers.

 

[tristan308]

Oops. Clarification. Ran the fsck command.

 

[bi]

Tristan308,

to answer your second question: you will have to back up the data in the large filesystem, blow the large filesystem away, recreate it in the smaller size you want, and then restore the data. Just make sure the new, smaller size is large enough to hold the data you backed up.

 

[tristan308]

Thanks for the help!