Sorry for the ludicriously late reply, but I figured out what the problem eventually was. While it was showing up in the ZFS code, the problem was actually fsflush. The machine has 64G of RAM and when memory started to get active, the scanning took a solid 7+ seconds to complete. Setting "autoup = 600" in /etc/system appeared to clear up the problem.
Thanks to everyone for the help! -- This message posted from opensolaris.org _______________________________________________ dtrace-discuss mailing list dtrace-discuss@opensolaris.org
