Open GoogleCodeExporter opened 8 years ago
Here's the nabble link, if/when the macosforge list gets removed:
http://n3.nabble.com/ZFS-performance-tuning-td21704.html#a21704
Note that the bug listed in here was part of 124, but here's the change:
hg log -r 0e96dd3b905a
Diff (against the hg) is attached. We may not be able to roll forward to this
until later.
Original comment by alex.ble...@gmail.com
on 2 Nov 2009 at 11:37
Attachments:
Don't forget tuning is evil.
http://www.solarisinternals.com/wiki/index.php/ZFS_Evil_Tuning_Guide
Original comment by sammyomi...@gmail.com
on 3 Nov 2009 at 5:39
I don't think it's going to be too long before we roll forwards, which will
bring some of those enhancements. In
any case, a lot of performance tuning is going to be load/machine/disk
dependent.
What will be really cool is to experiment with a separate ZIL device, which we
don't have yet (but I think was
added fairly soon after the 72 split).
By all means, let's discuss possibilities here, but as for code optimisations,
it might be difficult to roll forward
those in the future if we then pick up a good split point for ZFS.
Original comment by alex.ble...@gmail.com
on 5 Nov 2009 at 9:49
I think the link now refers to
http://mac-os-x-zfs-discuss.19757.n3.nabble.com/ZFS-performance-tuning-td21704.h
tml
Original comment by alex.ble...@gmail.com
on 6 Jan 2011 at 1:52
Original issue reported on code.google.com by
jason.richard.mcneil
on 2 Nov 2009 at 7:24