Re: Very rapid inode leak in 2.2.14|13...

From: Keith Owens (kaos@ocs.com.au)
Date: Sun Jan 16 2000 - 04:42:01 EST


On Sat, 15 Jan 2000 18:01:43 -0800,
Chip Salzenberg <chip@valinux.com> wrote:
>I'm seeing the leak, and I've never used reiserfs. And IIRC, Joe is
>seeing his leak even with pristine kernel source.

Data point. 2.2.12 with minimal services, a few getty, minilog, kswapd
and friends. No squid, mail, httpd, named, X etc.

# cat /proc/sys/fs/inode-state
2430 28 0 0 0 0 0
# stat /dev/ttyS2
  File: "/dev/ttyS2"
  Size: 0 Filetype: Character Device
  Mode: (0600/crw-------) Uid: ( 0/ root) Gid: ( 5/ tty)
Device: 3,2 Inode: 2802 Links: 1 Device type: 4,66
Access: Wed May 6 06:32:27 1998(00620.13:04:26)
Modify: Wed May 6 06:32:27 1998(00620.13:04:26)
Change: Thu Jul 22 19:32:16 1999(00178.00:04:37)
# cat /proc/sys/fs/inode-state
2430 27 0 0 0 0 0
# find /etc > /dev/null
# cat /proc/sys/fs/inode-state
2610 7 0 0 0 0 0

AFAICT, doing stat on any inode will use up inodes. But only the first
time you stat a node, do it again and there is no effect. This occurs
on at least ext2 and vfat filesystems but the symptom does not occur
doing stat() on /proc.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Sun Jan 23 2000 - 21:00:13 EST