I can’t seem to see any reference to SATA, ATA or IDE - the only thing that exists is iscso_ibft and iscsi_boot_sysfs …
lsmod output in full - openSUSE Leap lsmod - Pastebin.com
Hi
I’m at a loss… think your going to have to wait on a ML reply…
I wouldn’t be surprised - this thread got 80+ views before I got the first suggestion.
From googling around before I first asked, it does seem like Marvell chips are infamous… but then the solution always seems to be to buy an expensive RAID card…
On the other hand, one would have thought that given that these chips are also built directly into mainboards that someone would have an idea…
A curious detail, the 9123 and 9128 chips seem to be treated as 1… at least they share an entry:
OK, found a nice German thread about tested hardware in the Dell T20 
- http://www.hardwareluxx.de/community/f101/dell-poweredge-t20-1031138.html
The CSL card with the Asmedia chip is listed - so I ordered that via Amazon.
(10.85 pounds)
However I would still love to see a resolution of the Marvell issue if possible - if only to know there is one less bug in Linux.
So, quick update: Got a new card (from CSL with an ASMedia 1062 chip according to openSUSE and the description said 1061… OK) and so far all is well, it is still rebuilding the array or rather re-adding the drive, 1TB out of 3TB are done so far. Speed varied, was as high as 170MB/s, now it is at 85MB/s… patience…
Anyway, I was considering keeping the Marvell card with an old 2.5" drive - maybe - for experimenting (it is removed for now) but just for the fun of it, I tried an old drive…
The old drive is a SATA2, 2.5" harddrive with an advertised 120GB, well, no change… - also searched around a bit and disabled the write cache… still failed…
Also collected dmesg while it was running prior to failure in case someone is interested…
As usual, files hosted by Pastebin.
Just running dd with defaults: http://pastebin.com/PkF5W7zr
dd while it was running: http://pastebin.com/eAgu7gXj
dd with noncq set manually via command line: http://pastebin.com/FKhFMMd8
dd with noncq set manuall and no write cache via hdparm: http://pastebin.com/icEgbKN4
On Wed 02 Dec 2015 04:06:02 PM CST, DetlevCM wrote:
So, quick update: Got a new card (from CSL with an ASMedia 1062 chip
according to openSUSE and the description said 1061… OK) and so far
all is well, it is still rebuilding the array or rather re-adding the
drive, 1TB out of 3TB are done so far. Speed varied, was as high as
170MB/s, now it is at 85MB/s… patience…
Anyway, I was considering keeping the Marvell card with an old 2.5"
drive - maybe - for experimenting (it is removed for now) but just for
the fun of it, I tried an old drive…
The old drive is a SATA2, 2.5" harddrive with an advertised 120GB, well,
no change… - also searched around a bit and disabled the write
cache… still failed…
Also collected dmesg while it was running prior to failure in case
someone is interested…
As usual, files hosted by Pastebin.
Just running dd with defaults: dd standard failure - SATA2 2.5" 120GB - Pastebin.com
dd while it was running: dd running hasn't failed - SATA2 2.5" 120GB - Pastebin.com
dd with noncq set manually via command line:
dd failed noncq - SATA2 2.5" 120GB - Pastebin.com
dd with noncq set manuall and no write cache via hdparm:
dd failed noncq no write cache - SATA2 2.5" 120GB - Pastebin.com
Hi
If your not going to use it, why not contact the Marvell driver
developer(s), email from modinfo and see if you can donate them the
card and maybe they can actually work out what the issue is…
–
Cheers Malcolm °¿° LFCS, SUSE Knowledge Partner (Linux Counter #276890)
SUSE Linux Enterprise Desktop 12 | GNOME 3.10.1 | 3.12.48-52.27-default
If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below… Thanks!
Well, ideally I would like to use it… (and the long term plan is to grow the server as needed)
I’ll see what I’ll do, for all I know StarTech might figure out the issue (they wanted to try too).