Thanks Colin, I believe that bountysource.com url is a scrape of https://github.com/openzfs/zfs/issues/6223
I hadn't seen much of this information before. Sadly my machine doesn't have enough memory to just keep turning up the knobs -- the working set far exceeds the memory of my computer. A recent archive-wide search I performed, I tried the following: echo 126491074560 >> /sys/module/zfs/parameters/zfs_arc_max # eight gigs less than RAM echo 117901139968 >> /sys/module/zfs/parameters/zfs_arc_dnode_limit # eight gigs less than ^^ I picked these numbers entirely at a guess. The zfs arc had previously been 0 -- for half of ram, 64 gigabytes. This workload seemed to run fine for longer than usual (arc_prune cpu numbers were low, good read iops reported in zpool status output -- not ideal indicator of actual progress, but a reasonable proxy) but eventually the arc_prune tasks started consuming most of the CPUs, and read iops fell to the dozens. I haven't tried a run with silly low numbers for the arc dnode limit, but I'm curious... Thanks ** Bug watch added: github.com/openzfs/zfs/issues #6223 https://github.com/openzfs/zfs/issues/6223 -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1814983 Title: zfs poor sustained read performance from ssd pool To manage notifications about this bug go to: https://bugs.launchpad.net/zfs/+bug/1814983/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs