https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=224496
Daniel Austin <freebsd-po...@dan.me.uk> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |freebsd-po...@dan.me.uk
--- Comment #48 from Daniel Austin <freebsd-po...@dan.me.uk> ---
I have this same issue too...
My setup is a LSI 9206-16e PCIe card (which is really just 2 x LSISAS2308 cards
with a PCIe switch) and BSD 12+13. I have this connecting to a QNAP TL-D1600S
16-bay SATA chassis with 12 x 8TB Toshiba SATA (non-SMR) disks.
I tried a workaround online (camcontrol tags daX -N 1) to disable NCQ per drive
and this is fine *once the server is booted*... however, if i ever rebooted, I
had a massive scroll of CAM errors as the kernel was trying to import a ZFS
pool before the camcontrol script had run... This even leads to ZFS reporting
too many errors and taking the pool offline (note: i'm not booting from this
pool)
I don't have the luxury of a firmware update as there hasn't been any updates
to the SAS2308 in a long long time! I am running 20.00.07.00.
I've also tried 3 different 9206-16e cards, so i'm happy it's not just a faulty
card. I also tried in 2 different servers with the same results so happy it's
probably not a hardware issue at all.
My final solution which is really a kludge but does fix the issue permanently
was to boot into a live ubuntu environment and use lsiutil 1.72 to disable NCQ
on the card itself. This is saved in the cards EEPROM.
Now when I boot bsd, I get zero CAM errors from any disks, zfs pool imports
straight away, and I can still max out the bandwidth to the drives - yay.
I also have an LSI 9207-8i card in a different machine running the same
firmware which has no issues at all even with NCQ enabled (but these are
directly connected to the card via SATA cables with no enclosure as such), so I
do think this is just some kind of incompatibility between
driver<-->card<-->enclosure of some kind.
It would be lovely if lsiutil could be ported to bsd... I did look at it
briefly but it's beyond my capabilities... the source code is online if anyone
wanted to try.
Hope that helps anyone else stumbling upon this PR and I appreciate it may only
fix some not all of the cases reported here so far... but some is better than
none :-)
--
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
freebsd-bugs@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-bugs
To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
