[olug] My inodes ran out

Trent Melcher tmelcher at trilogytel.com
Mon Dec 16 16:31:14 UTC 2002


I have seen this too,  thats why I try to make it a point to reboot most
of my *nix server every couple months.  Cleans up lost inodes and
orphaned processes (Zombies) on most systems.

Trent


On Mon, 2002-12-16 at 09:34, Mike Peterson wrote:
> Watch your Redhat 7.3 systems closely.
>  
> On 2 out of 3 Redhat 7.3 systems I work with the following problem has
> occurred.
>  
> I could not send mail or modem connect to the system.
>  
> The logs for in and out mail and modem logs are in /var/log/.
>  
> The inbound mail is stored in /var/spool/mail/ also.
>  
> I create a separate /var partition so the root does not get filled up
> on production systems.
>  
> On the 2 out of 3 systems mentioned above I found that the
> /var/log/mailman error log extension names keep getting longer and
> even though logrotate is supposed to run weekly, the number of empty
> error log files exceeded 10,000 after less than 6 months of the system
> running.
>  
> On 1 of the 3 systems mentioned above I found that the /var/log mgetty
> error log extension names keep getting longer and even though
> logrotate is supposed to run weekly, the number of empty error log
> files exceeded 10,000 after less than 6 months of the system running.
>  
> Between the modem and mailman log files they filled the inode
> allotment for the /var filesystem.
>  
> I had to use
>  
> find /var/log -name "*mgetty*" -exec rm {} \;
> and
> find /var/log/mailman -name "*error*" -exec rm {} \;
>  
> to remove the files because it was to many files for ls and rm to
> remove any other way.
>  
>  




More information about the OLUG mailing list