On Wed, Feb 9, 2011 at 12:02 AM, Richard Elling <richard.ell...@gmail.com> wrote: > The data below does not show heavy CPU usage. Do you have data that > does show heavy CPU usage? mpstat would be a good start.
Here is mpstat output during a network copy; I think one of the CPUs disappeared due to a L2 Cache error. movax@megatron:~# mpstat -p CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl set 1 333 0 6 4057 3830 19467 140 27 265 0 1561 1 48 0 51 0 > Some ZFS checksums are always SHA-256. By default, data checksums are > Fletcher4 on most modern ZFS implementations, unless dedup is enabled. I see, thanks for the info. >> Second, a copy from my desktop PC to my new zpool. (5900rpm drive over >> GigE to 2 6-drive RAID-Z2s). Load average are around ~3. > > Lockstat won't provide direct insight to the run queue (which is used to > calculate > load average). Perhaps you'd be better off starting with prstat. Ah, gotcha. I ran prstat, which is more of what I wanted: PID USERNAME SIZE RSS STATE PRI NICE TIME CPU PROCESS/NLWP 1434 root 0K 0K run 0 -20 0:01:54 23% zpool-tank/136 1515 root 9804K 3260K cpu1 59 0 0:00:00 0.1% prstat/1 1438 root 14M 9056K run 59 0 0:00:00 0.0% smbd/16 zpool thread near the top of usage, which is what I suppose you would expect. _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
