C2750D4I stability problems

[a13x3181]

Re: C2750D4I stability problems

OS : freebsd 10
RAM : kingston non ECC 2 * 8GB
zpool : 6 * 3TB WD Red

I did dd test for 10GB file, no issue
also, copied 450GB files from local disk ( OS disk ) to zpool, no issue.

[a13x3181]

Re: C2750D4I stability problems

guy. i did a test using iperf .

My NIC died at 14051 second and each second transfer 114-155MB
i'm successfully transfer the 14051 * 114MB = 1564GB data thru network.

I login to console ( no hung ) and restart the freebsd netif service, the network connectivity is back again.

[larch]

Re: C2750D4I stability problems

Just FYI in the end I RMA'd the board and the Rev 1.1 they sent back is totally stable even with speedstep etc switched on.

Much much better.

[ASRockRack TSD - Bill]

Re: C2750D4I stability problems

Good to know. Thanks for the update!

[Retteren]

Re: C2750D4I stability problems

Just FYI in the end I RMA'd the board and the Rev 1.1 they sent back is totally stable even with speedstep etc switched on.

Much much better.

larch, was that Revision 1.1 or 1.01? - I recently received my board, and the PCB version is 1.01, but with BIOS P1.50 and BMC FW 0.19.0.
Just curious

[Retteren]

Re: C2750D4I stability problems

OK, my turn to have stability problems with the C2750D4I's Marvell controllers :(

Setup:
Fractal Node 304 case. All three fans are connected to the motherboard.
Corsair RM450 PSU.
C2750D4I, Rev GA 1.01, BIOS P1.50, BMC FW 0.19.0.
Kingston 16GB ECC RAM (2 x 8GB KVR16LE11/8) in the blue slots.
bootdisk: Samsung 840 Evo 250GB on SATA_0 (Intel SATA3 controller)
disk array: 6 x WD60EFRX on the Marvell controllers/ports. (SATA3_M0-M1, SATAIII_M0-M3).
BIOS: C-state is disabled.

OS: Ubuntu 14.04 LTS (updated).
As a burn-in of the disks I am running "badblock" on the disks (badblock -b 4092 -c 16384 -ws /dev/<disk>).

If I run badblock on 2 disks at a time, I don't get any errors reported, and a full scan (1 W+R) takes just over 24hours (the default 4 runs [0x55, 0xaa, 0xff, 0x00] takes ~98-99h).
I have done this without issues on the first 4 disks, and the disks all report zero issues.

But if I run 4 badblock processes at a time, the kernel starts to spit out ata link errors and eventually it lost all contact with 2 out of the 4 tested disks. That promptly filled up my /var partition with messages in /var/log/kern.log and /var/log/syslog (30GB).

I have tried with SpeedStep both enabled and disabled, but it makes no difference. Errors start coming in less than 30 seconds of starting the test.

[ 1056.614092] ata12.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
[ 1056.614101] ata12.00: failed command: IDENTIFY DEVICE
[ 1056.614108] ata12.00: cmd ec/00:01:00:00:00/00:00:00:00:00/00 tag 22 pio 512 in
[ 1056.614108]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1056.614112] ata12.00: status: { DRDY }
[ 1056.614120] ata12: hard resetting link
[ 1056.934331] ata12: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 1056.935792] ata12.00: configured for UDMA/133
[ 1056.935905] ata12: EH complete
[ 1060.681285] ata9.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
[ 1060.681293] ata9.00: failed command: IDENTIFY DEVICE
[ 1060.681300] ata9.00: cmd ec/00:01:00:00:00/00:00:00:00:00/00 tag 4 pio 512 in
[ 1060.681300]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1060.681304] ata9.00: status: { DRDY }
[ 1060.681310] ata9: hard resetting link
[ 1061.001500] ata9: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 1061.002958] ata9.00: configured for UDMA/133
[ 1061.003065] ata9: EH complete
[ 1284.780087] ata11.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
[ 1284.780096] ata11.00: failed command: SMART
[ 1284.780103] ata11.00: cmd b0/d1:01:00:4f:c2/00:00:00:00:00/00 tag 8 pio 512 in
[ 1284.780103]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1284.780107] ata11.00: status: { DRDY }
[ 1284.780113] ata11: hard resetting link
[ 1285.112644] ata11: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 1290.116254] ata11.00: qc timeout (cmd 0xec)
[ 1290.124235] ata11.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[ 1290.124239] ata11.00: revalidation failed (errno=-5)
[ 1290.124247] ata11: hard resetting link
[ 1290.452510] ata11: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 1300.460315] ata11.00: qc timeout (cmd 0xec)
[ 1300.468302] ata11.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[ 1300.468307] ata11.00: revalidation failed (errno=-5)
[ 1300.468313] ata11: limiting SATA link speed to 3.0 Gbps
[ 1300.468316] ata11.00: limiting speed to UDMA/133:PIO3
[ 1300.468321] ata11: hard resetting link
[ 1300.796574] ata11: SATA link up 6.0 Gbps (SStatus 133 SControl 320)
[ 1300.798094] ata11.00: configured for UDMA/133
[ 1300.798218] ata11: EH complete
[ 1309.867675] ata12.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
[ 1309.867683] ata12.00: failed command: IDENTIFY DEVICE
[ 1309.867690] ata12.00: cmd ec/00:01:00:00:00/00:00:00:00:00/00 tag 10 pio 512 in
[ 1309.867690]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1309.867694] ata12.00: status: { DRDY }
[ 1309.867700] ata12: hard resetting link
[ 1310.195915] ata12: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 1310.197568] ata12.00: configured for UDMA/133
[ 1310.197705] ata12: EH complete
[ 1317.841921] ata12.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
[ 1317.841930] ata12.00: failed command: SMART
[ 1317.841937] ata12.00: cmd b0/d1:01:00:4f:c2/00:00:00:00:00/00 tag 17 pio 512 in
[ 1317.841937]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1317.841941] ata12.00: status: { DRDY }
[ 1317.841946] ata12: hard resetting link
[ 1318.170136] ata12: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 1323.174055] ata12.00: qc timeout (cmd 0xec)
[ 1323.182026] ata12.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[ 1323.182030] ata12.00: revalidation failed (errno=-5)
[ 1323.182037] ata12: hard resetting link
[ 1323.510382] ata12: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 1333.518130] ata12.00: qc timeout (cmd 0xec)
[ 1333.526096] ata12.00: failed to IDENTIFY (I/O error, err_mask=0x4)
[ 1333.526101] ata12.00: revalidation failed (errno=-5)
[ 1333.526106] ata12: limiting SATA link speed to 3.0 Gbps
[ 1333.526111] ata12: hard resetting link
[ 1333.854368] ata12: SATA link up 6.0 Gbps (SStatus 133 SControl 320)
[ 1333.855848] ata12.00: configured for UDMA/133
[ 1333.855955] ata12: EH complete
[ 1341.892701] ata12.00: NCQ disabled due to excessive errors
[ 1341.892709] ata12.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
[ 1341.892714] ata12.00: failed command: SMART
[ 1341.892721] ata12.00: cmd b0/d0:01:00:4f:c2/00:00:00:00:00/00 tag 3 pio 512 in
[ 1341.892721]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1341.892725] ata12.00: status: { DRDY }
[ 1341.892731] ata12: hard resetting link
[ 1342.220915] ata12: SATA link up 6.0 Gbps (SStatus 133 SControl 320)
[ 1342.222408] ata12.00: configured for UDMA/133
[ 1342.222521] ata12: EH complete
[ 1350.859692] ata10.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
[ 1350.859700] ata10.00: failed command: IDENTIFY DEVICE
[ 1350.859707] ata10.00: cmd ec/00:01:00:00:00/00:00:00:00:00/00 tag 0 pio 512 in
[ 1350.859707]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 1350.859711] ata10.00: status: { DRDY }
[ 1350.859717] ata10: hard resetting link

</disk>

[Retteren]

Re: C2750D4I stability problems

Update: I've been talking to Bill on the side, and so far I have been able to resolve the issues by 2 drives away from the SE9130 controller, to the Intel SATA2 controller. Ie: now I have 2 drives on the SE9178, 2 drives on the SE9230, 2 drives on the Intel SATA2, and the SSD boot drive on the Intel SATA3 controller. The move seriously messes up the drive detection order under Linux, but that is definitely Linux's issue (seems the Intel SATA2 ports are probed first, so now my boot drive is 'sdc'). I'll be picking up some Silverstone SST-CP11 cables, as some people from the unRAID forum (linked above) reported better success with these cables on the SE9230.