xfs
[Top] [All Lists]

Re: XFS dying when many processes copy many files/directories

To: Eric Sandeen <sandeen@xxxxxxx>
Subject: Re: XFS dying when many processes copy many files/directories
From: Adrian Head <ahead@xxxxxxxxxxxxxx>
Date: Fri, 21 Dec 2001 09:46:34 +1000
Cc: <linux-xfs@xxxxxxxxxxx>
In-reply-to: <Pine.LNX.4.33.0112161937450.12284-100000@xxxxxxxxxxxxxxxxxxxxxxxx>
References: <Pine.LNX.4.33.0112161937450.12284-100000@xxxxxxxxxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Although the CVS (18Dec01) did fix the problem for 80 background cp processes 
the system still hangs on 160 background cp processes.  I have tried this 
test with ext2 & reiserfs and neither of these filesystems have problems - 
they just chug away until finished.

The machine did not hang tight as I could still control the top session but I 
could not start a new console or ssh session.  I could still ping the box.

On Mon, 17 Dec 2001 11:39, Eric Sandeen wrote:
> From looking at these traces, it looks very much like the problem which
> was solved just recently... I'd bet $0.02 that the latest CVS code will
> solve your problems.  (Note that the cvs snapshot patch is not the latest
> CVS code.)

Below are the (shorthand) traces of kupdated and a cp process that 
deadlocked.  

PID6 kupdated =>
write_some_buffers
kupdate
kernel_thread

PID 3080 cp =>
schedule
_sv_wait
xlog_grant_log_space
xfs_log_reserve
cfs_trans_reserve
xfs_strategy
linvfs_pb_bmap
shrink_cache
try_to_free_paes
balance_classzone
__alloc_pages
linvfs_pb_bmap
linvfs_write
sys_write
system_call

This is just a screen dump of top

  7:28am  up 1 day, 23:37,  1 user,  load average: 164.17, 164.06, 164.01
186 processes: 183 sleeping, 3 running, 0 zombie, 0 stopped
CPU states:  0.0% user, 100.0% system,  0.0% nice,  0.0% idle
Mem:   383676K av,  381500K used,    2176K free,       0K shrd,     112K buff
Swap:  524624K av,   12144K used,  512480K free                  134136K 
cached

  PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM  CTIME COMMAND
    6 root      18   0     0    0     0 RW   99.6  0.0 505:40 kupdated
 3289 root      11   0   952  896   668 R     0.3  0.2   2:14 top
    1 root       9   0   208  152   136 S     0.0  0.0  36:14 init
    2 root       9   0     0    0     0 SW    0.0  0.0   1:14 keventd
    3 root      19  19     0    0     0 SWN   0.0  0.0   0:00 ksoftirqd_CPU0
    4 root       9   0     0    0     0 DW    0.0  0.0   0:54 kswapd
    5 root       9   0     0    0     0 SW    0.0  0.0  10:15 bdflush
    7 root       9   0     0    0     0 SW    0.0  0.0   0:04 pagebuf_daemon
    8 root      -1 -20     0    0     0 SW<   0.0  0.0   0:00 mdrecoveryd
  126 root      -1 -20     0    0     0 SW<   0.0  0.0  20:15 raid5d
  554 root       9   0   228  176   152 S     0.0  0.0   0:01 syslogd
  559 root       9   0   732   68    68 S     0.0  0.0   0:00 klogd
  573 rpc        9   0   352  264   264 S     0.0  0.0   0:00 portmap
  588 rpcuser    9   0   108    4     4 S     0.0  0.0   0:00 rpc.statd
  674 daemon     9   0   108   44    44 S     0.0  0.0   0:00 atd
  686 root       9   0   376  196   192 S     0.0  0.0   0:11 sshd
  754 root       9   0   176  120   104 D     0.0  0.0   0:46 crond
  782 root       9   0   136   76    76 S     0.0  0.0   0:00 mingetty
  783 root       9   0   136   76    76 S     0.0  0.0   0:00 mingetty
  784 root       9   0   136   76    76 S     0.0  0.0   0:00 mingetty
  785 root       9   0   136   76    76 S     0.0  0.0   0:00 mingetty
  786 root       9   0   136   76    76 S     0.0  0.0   0:00 mingetty
 1178 root       9   0   268  208   208 D     0.0  0.0   0:00 mingetty
 3080 root       9   0   664  612   440 D     0.0  0.1   0:08 cp
 3081 root       9   0   664  612   440 D     0.0  0.1   0:08 cp
 3082 root       9   0   664  604   440 D     0.0  0.1   0:08 cp
 3083 root       9   0   640  588   440 D     0.0  0.1   0:08 cp
 3084 root       9   0   664  604   440 D     0.0  0.1   0:08 cp
 3085 root       9   0   648  588   440 D     0.0  0.1   0:08 cp
 3086 root       9   0   664  604   440 D     0.0  0.1   0:08 cp
 3087 root       9   0   664  608   440 D     0.0  0.1   0:08 cp
 3088 root       9   0   664  604   440 D     0.0  0.1   0:08 cp
 3089 root       9   0   660  608   440 D     0.0  0.1   0:08 cp
 3090 root       9   0   664  608   440 D     0.0  0.1   0:08 cp
 3091 root       9   0   664  612   440 D     0.0  0.1   0:09 cp
 3092 root       9   0   640  592   440 D     0.0  0.1   0:08 cp
 3093 root       9   0   664  604   396 D     0.0  0.1   0:08 cp
 3094 root       9   0   660  600   440 D     0.0  0.1   0:08 cp
 3095 root       9   0   660  604   440 D     0.0  0.1   0:08 cp
 3096 root       9   0   664  608   440 D     0.0  0.1   0:08 cp
 3097 root       9   0   664  604   440 D     0.0  0.1   0:09 cp
 3098 root       9   0   660  608   440 D     0.0  0.1   0:08 cp
 3099 root       9   0   664  612   440 D     0.0  0.1   0:08 cp
 3100 root       9   0   640  592   440 D     0.0  0.1   0:08 cp
 3101 root       9   0   660  600   396 D     0.0  0.1   0:08 cp
 3102 root       9   0   640  588   440 D     0.0  0.1   0:08 cp
 3103 root       9   0   648  596   440 D     0.0  0.1   0:08 cp
 3104 root       9   0   660  604   440 D     0.0  0.1   0:08 cp
 3105 root       9   0   660  600   440 D     0.0  0.1   0:08 cp
 3106 root       9   0   664  608   404 D     0.0  0.1   0:08 cp
 3107 root       9   0   660  604   404 D     0.0  0.1   0:08 cp



- -- 
Adrian Head

(Public Key available on request.)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8Inhg8ZJI8OvSkAcRAq8LAJwMEIMNxYR4zeF1sV5DsXGnSWO9HACeNtgh
FOdmsz5eKIlJ4nv9jfWk19w=
=LcH8
-----END PGP SIGNATURE-----


<Prev in Thread] Current Thread [Next in Thread>