Re: BUG: jbd2 slowing down file copies even though no journalingfile system is used

[Ted Ts'o]

On Sun, May 13, 2012 at 01:21:00AM +0200, Björn Christoph wrote:
> 
> IOTOP shows this in some cases, mostly jbd2 is not there:
>   PRIO  USER     DISK READ  DISK WRITE  SWAPIN     IO>    COMMAND
>   416 be/3 root        0.00 B/s   97.65 M/s  0.00 % 13.74 % [jbd2/dm-1-8]

Well, the file system in that has the huge write bandwidth is
whichever device is associated with device mapper device dm-1 (i.e.,
with a device mapper minor number of 1).  What does "ls -l
/dev/mapper" show you?

Whatever file system is associated with it is clearly generating a lot
of journal activity.

Something that may help in determining what process is generating all
of this journal activity (which is likely the result of something
calling fsync a lot) is to try this:

echo 1 > /sys/kernel/debug/tracing/events/ext4/ext4_sync_file_enter/enable

... then wait for a minute or so, and then capture the output of:

cat /sys/kernel/debug/tracing/trace

See if that shows up anything useful.

Regards,

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