Re: syncer going nuts on 6.0

On Tue, Oct 25, 2005 at 08:29:59PM +0300, Gleb Kozyrev wrote:
> Kris Kennaway wrote to <current_at_FreeBSD.org> on Tue, 25 Oct 2005 11:59:33 -0400:
> 
> KK> I have 3 UP 6.0 machines that are currently stuck using 100-epsilon%
> KK> of CPU in the syncer:
> 
> KK> last pid: 17621;  load averages:  9.02,  7.21,  4.82                               
> KK> up 3+18:15:31 +08:49:50
> KK> 73 processes:  2 running, 47 sleeping, 24 waiting
> KK> CPU states:  0.0% user,  0.0% nice,  100% system,  0.0% interrupt,  0.0% idle
> KK> Mem: 27M Active, 149M Inact, 137M Wired, 4K Cache, 109M Buf, 684M Free
> KK> Swap: 2048M Total, 8K Used, 2048M Free
> KK>   PID USERNAME  THR PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
> KK>    38 root        1  -4    0     0K     8K getblk 180:21 431.30% syncer
> KK> 17474 root        1  -4    0  1928K  1428K getblk   1:13 28.71% rm
> KK>    11 root        1 171   52     0K     8K RUN     20.2H  8.94% idle
> 
> KK> rm -rf is just removing a single ufs directory tree, but it's taking
> KK> *much* longer than it should be (order of 5 minutes instead of 30
> KK> seconds it took for the other 25 machines), and the machine was
> KK> extremely unresponsive in the meantime (response times of 30 seconds
> KK> to run commands like top).  WITNESS is on, but it's also on on the
> KK> other 25 machines that aren't seeing a problem under what should be
> KK> analogous loads.
> 
> I've seen the same problem and it seems to be related with filesystem snapshots.

I'm not using them.

Kris