-   Linux - Server (
-   -   How to correct possibly wrong directory entry ? (

jbenezech 07-23-2012 09:31 PM

How to correct possibly wrong directory entry ?
The server I'm looking at had a directory with millions of file and almost 1GB space. This was due to a software bug and has been corrected since.
In order to delete the files, the host admin renamed the directory then ran a script to rm all files in it with sleep interval. It took about a month but the files have been deleted.
My problem is that the directory entry still shows the same 1GB disk space using ls -l.
I need to do a scan on the directory structure but commands like ls -R or grep -R get put in an uninterrupted disk sleep. I guess it comes from this huge directory (If you think the problem is different, please let me know).

What would be the best way to correct that directory entry or ideally delete it without crashing a quite busy production server to which I have no physical access ?
I've read somewhere that renaming it would work but I'd rather ask here first, make absolutely sure I don't crash it.

Linux version 2.6.18-194.11.4.el5PAE ( (gcc version 4.1.2 20080704 (Red Hat 4.1.2-48)) #1 SMP Tue Sep 21 05:48:23 EDT 2010

Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sda3 959228044 111934848 798567248 13% /

940552192 Jul 21 17:00 cache_old/

Thanks for your help

pan64 07-24-2012 03:53 AM

to check or fix filesystem problems you need to start with fsck. I'm not really sure, is this /cache_old dir now really empty?

chrism01 07-29-2012 08:31 PM

Try also

df -i which should confirm the inode usage/count.

All times are GMT -5. The time now is 11:58 PM.