metalwolf

Hi, Just to fill in a few additional blanks. The cards running the latest firmware (I got it from 45drives guide on updating the card). There were other disks a few seconds earlier i clipped the last few as they were all the same. My test was to stop all my docker containers, spin down the array and start up a single container (plex). After a few seconds it started spinning up disks, so I'm assuming it was writing and they are the parity drives. I did follow a few other guides that had me increase the time the card should wait for the drive to be ready to 30 and 45 seconds for the two sections but this did nothing. Thinking back, it may have made things worse as the errors originally when i tried to fix it weren't showing up immediately on spin up. I might still have the errors somewhere.... I do have another identical card (running on an older firmware) that i had different issues with (1 disk reported thousands of errors - but i think the card wasn't to blame on this. I will give it a go with both cards and report back anything i find in case its helpful for someone else.

I've had it both ways sometimes it just shows up as an error and the drive works but sometimes it kicks the drives out. The drive itself works fine and the smart data of the drive shows no increment of the issue counters. In my recent testing I've stopped the array as soon as it showed 1 error to prevent it kicking a disk out. As mentioned going back to the 9211-8i and sas expander I get no issues at all. I have another pc I am going to set up with the raid controller and a couple of 16tb drives (spares) with an unraid demo license for testing. if theres anything you want me to try let me know, I need to format a couple of the disks with for the 520 vs 512 issue first. I'm reluctant to put the card back in my main array until I figure out if its solvable.

I am experiencing a similar issue to that the above poster. I have an unraid server I've had for a while that I've swapped the SAS card and drives at the same time. After doing some troubleshooting (replacing cables, repalcing the SAS card, changing spin up time outs on the card and upgrading the firmware) I am now getting the following error messages in the unraid log files: Jun 25 09:06:35 Tower kernel: sd 0:0:2:0: [sdb] tag#3806 UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte=DRIVER_OK cmd_age=0s Jun 25 09:06:35 Tower kernel: sd 0:0:2:0: [sdb] tag#3806 Sense Key : 0x2 [current] [descriptor] Jun 25 09:06:35 Tower kernel: sd 0:0:2:0: [sdb] tag#3806 ASC=0x4 ASCQ=0x1c Jun 25 09:06:35 Tower kernel: sd 0:0:2:0: [sdb] tag#3806 CDB: opcode=0x88 88 00 00 00 00 03 80 0f 17 00 00 00 00 08 00 00 Jun 25 09:06:35 Tower kernel: I/O error, dev sdb, sector 15033374464 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0 Jun 25 09:06:35 Tower kernel: md: disk1 read error, sector=15033374400 Jun 25 09:06:35 Tower kernel: sd 0:0:3:0: [sdd] tag#3808 UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte=DRIVER_OK cmd_age=0s Jun 25 09:06:35 Tower kernel: sd 0:0:3:0: [sdd] tag#3808 Sense Key : 0x2 [current] [descriptor] Jun 25 09:06:35 Tower kernel: sd 0:0:3:0: [sdd] tag#3808 ASC=0x4 ASCQ=0x1c Jun 25 09:06:35 Tower kernel: sd 0:0:3:0: [sdd] tag#3808 CDB: opcode=0x88 88 00 00 00 00 03 80 0f 17 00 00 00 00 08 00 00 Jun 25 09:06:35 Tower kernel: I/O error, dev sdd, sector 15033374464 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0 Jun 25 09:06:35 Tower kernel: md: disk4 read error, sector=15033374400 Jun 25 09:06:36 Tower emhttpd: read SMART /dev/sdd Jun 25 09:06:36 Tower emhttpd: read SMART /dev/sdb Jun 25 09:06:36 Tower emhttpd: read SMART /dev/sdf Jun 25 09:06:36 Tower emhttpd: read SMART /dev/sdc The array works perfectly if i don't spin the disks down and starts no issues at boot time. The issue happens whenever a disk spins up and it happens almost immediately. I have run disk verification on all the drives before putting them into the array. The card I'm using is an LSI 9305-16i using the firmware from 45 drives. The drives are all Dell Branded Seagate Exos X16 16TB drives with part number ST16000NM010G. Looking at the error (as I understand it) its asking the drive to spin up and then for data almost immediately resulting in a disk not ready reply that is causing the error. I'm not sure if this is correct? My old server used 12TB SAS drives from Dell and a 9211 card and I didn't have any issues like this. Any assistance is greatly appreciated. I am wondering if a fresh install of unraid would be the best thing or if I'm likely to have this issue again? Additional Note: I've kept the 16tb drives and gone back to the 9211-8i with a SAS expander and I have no such issues (from a quick test). So it appears to be something with 9305 cards and I don't understand what or how to further diagnose it. Any help on this would still be appreciated.

Thank you JorgeB, I will try that.

Hi, I have a system running unraid 7.2.4 with an LSI 9305-16i with a number of 16TB SAS drives attached. Since I set this system up and enabled the SAS Spindown plugin I see the error count on the disks increase when it is spinning up. The errors only occur at spin up and not at initial boot time. Looking in the logs I see errors like below. Jun 24 22:45:27 Tower kernel: sd 11:0:1:0: [sdc] tag#4212 UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte=DRIVER_OK cmd_age=0s Jun 24 22:45:27 Tower kernel: sd 11:0:1:0: [sdc] tag#4212 Sense Key : 0x2 [current] [descriptor] Jun 24 22:45:27 Tower kernel: sd 11:0:1:0: [sdc] tag#4212 ASC=0x4 ASCQ=0x1c Jun 24 22:45:27 Tower kernel: sd 11:0:1:0: [sdc] tag#4212 CDB: opcode=0x88 88 00 00 00 00 03 80 0f 16 e0 00 00 00 08 00 00 Jun 24 22:45:27 Tower kernel: I/O error, dev sdc, sector 15033374432 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0 I have carried out a number of trouble shooting steps to try resolve this: 1 - replaced the SAS cables 2 - replaced the power supply 3 - checked the drives are balanced on different power cables 4 - updated the SAS Card firmware 5 - Increased the timeout value on the card and by changing it in unraid using a script. None of these worked, but as I understand it the issue is happening because the drive is returning NOT_READY when its asked for a read while spinning up. has anyone had this issue and found a solution?