John_M Posted October 12, 2019 Share Posted October 12, 2019 There are longstanding problems with Marvell controllers and as newer Linux kernels are introduced the problems just get worse. Just Google "unraid marvell" and see for yourself. You can get second hand Dell H310s on Ebay. They are mostly pulled from servers and they are great value for money. That's what a lot of Unraid users are using now. The only nuisance is that they don't work "out of the box" as they generally come configured with RAID firmware. There are guides on this forum and elsewhere (they are popular with FreeNAS users and Serve The Home subscribers too) for crossflashing them to LSI IT firmware. The ASMedia ASM1062 is a simple two-port SATA card that works well (plug and play) with Unraid if you connect the disks to it directly, without a port multiplier. Quote Link to comment
JorgeB Posted October 12, 2019 Share Posted October 12, 2019 (edited) 11 hours ago, aurevo said: But how is it that only one hard disk was affected at a time? I can't say for sure the controller + pm is the problem, only that they are a common source of issues and the disk looks fine, but like I mentioned earlier that's a strange disk, white label? You can also swap that disk with one using the onboard SATA ports and see if the problem follows it. Edited October 12, 2019 by johnnie.black Quote Link to comment
aurevo Posted October 13, 2019 Author Share Posted October 13, 2019 11 hours ago, johnnie.black said: I can't say for sure the controller + pm is the problem, only that they are a common source of issues and the disk looks fine, but like I mentioned earlier that's a strange disk, white label? You can also swap that disk with one using the onboard SATA ports and see if the problem follows it. What do you mean with PM? Yes, it's a white label disc, you're right. That's exactly what I did yesterday, I connected the drive to an internal port and the other drive to the controller. I'll keep it under observation and report back otherwise. Thanks. Quote Link to comment
JorgeB Posted October 13, 2019 Share Posted October 13, 2019 Just now, aurevo said: What do you mean with PM? port multiplier Quote Link to comment
aurevo Posted October 16, 2019 Author Share Posted October 16, 2019 (edited) On 10/13/2019 at 12:36 PM, johnnie.black said: port multiplier Hello again, today the error occurred again. I just tried to shut down the server using the normal "Clean power down" function. After the timeout "force shutdown" occurred and after the system was rebooted and not shut down, the hard drive was gone again. The pattern is that the disk is no longer displayed after a shutdown. I didn't otherwise touch the case, touch any cables, or make any other changes to the hardware, I just wanted to turn it off to move it. After two reboots, the hard disk is visible again in the UnRAID, like the last times before. tower-diagnostics-20191016-1703.zip Edited October 16, 2019 by aurevo Logs attached Quote Link to comment
JorgeB Posted October 16, 2019 Share Posted October 16, 2019 Dis stopped responding and dropped offline, now on the onboard SATA ports: Oct 16 17:02:50 Tower kernel: ata1: link is slow to respond, please be patient (ready=0) Oct 16 17:02:55 Tower kernel: ata1: COMRESET failed (errno=-16) Oct 16 17:03:00 Tower kernel: ata1: link is slow to respond, please be patient (ready=0) Oct 16 17:03:01 Tower sshd[5655]: Accepted none for root from 192.168.178.44 port 63121 ssh2 Oct 16 17:03:05 Tower kernel: ata1: COMRESET failed (errno=-16) Oct 16 17:03:10 Tower kernel: ata1: link is slow to respond, please be patient (ready=0) Oct 16 17:03:40 Tower kernel: ata1: COMRESET failed (errno=-16) Oct 16 17:03:40 Tower kernel: ata1: limiting SATA link speed to 3.0 Gbps Oct 16 17:03:45 Tower kernel: ata1: COMRESET failed (errno=-16) Oct 16 17:03:45 Tower kernel: ata1: reset failed, giving up Oct 16 17:03:45 Tower kernel: ata1.00: disabled So likely a bad disk. Quote Link to comment
aurevo Posted October 16, 2019 Author Share Posted October 16, 2019 (edited) 25 minutes ago, johnnie.black said: Dis stopped responding and dropped offline, now on the onboard SATA ports: Oct 16 17:02:50 Tower kernel: ata1: link is slow to respond, please be patient (ready=0) Oct 16 17:02:55 Tower kernel: ata1: COMRESET failed (errno=-16) Oct 16 17:03:00 Tower kernel: ata1: link is slow to respond, please be patient (ready=0) Oct 16 17:03:01 Tower sshd[5655]: Accepted none for root from 192.168.178.44 port 63121 ssh2 Oct 16 17:03:05 Tower kernel: ata1: COMRESET failed (errno=-16) Oct 16 17:03:10 Tower kernel: ata1: link is slow to respond, please be patient (ready=0) Oct 16 17:03:40 Tower kernel: ata1: COMRESET failed (errno=-16) Oct 16 17:03:40 Tower kernel: ata1: limiting SATA link speed to 3.0 Gbps Oct 16 17:03:45 Tower kernel: ata1: COMRESET failed (errno=-16) Oct 16 17:03:45 Tower kernel: ata1: reset failed, giving up Oct 16 17:03:45 Tower kernel: ata1.00: disabled So likely a bad disk. That's what I'm thinking now. I ask for a return for the hard disk. If I can remove the disk from the array and distribute the data to the other hard disks, I would still have enough space on the disks. If so, how do I best proceed without risking data loss? I did not started the array until now, after restarting the machine. Edited October 16, 2019 by aurevo Quote Link to comment
JorgeB Posted October 16, 2019 Share Posted October 16, 2019 You can disable disk1 and copy everything from the emulated disk to the other disks, but it will require reeding all the other disks, one of them will read and write simultaneouly, so puts more stress than a regular rebuild, which IMHO is the best option, unless you don't plan on replacing that disk soon or ever. Quote Link to comment
aurevo Posted October 16, 2019 Author Share Posted October 16, 2019 6 minutes ago, johnnie.black said: You can disable disk1 and copy everything from the emulated disk to the other disks, but it will require reeding all the other disks, one of them will read and write simultaneouly, so puts more stress than a regular rebuild, which IMHO is the best option, unless you don't plan on replacing that disk soon or ever. I think it depends on whether I get the hard drive money back or a new one. So for now it would be best to start the normal rebuild and if the hard disk should be replaced again? Quote Link to comment
JorgeB Posted October 16, 2019 Share Posted October 16, 2019 I meant rebuilding to a new disk, you can use the array for a while with disk1 disable, but since you don't have dual parity if another disk fails in the meantime it will be unrecoverable. Quote Link to comment
aurevo Posted October 16, 2019 Author Share Posted October 16, 2019 13 minutes ago, johnnie.black said: I meant rebuilding to a new disk, you can use the array for a while with disk1 disable, but since you don't have dual parity if another disk fails in the meantime it will be unrecoverable. I think the most sensible way for me at the moment is to remove the disk and distribute the data to the other disks. Should I do this best with UnBALANCED? And if the data is copied, do I need a new configuration without the hard disk? Can I do that? So first copy, then new configuration without the hard disk. Quote Link to comment
JorgeB Posted October 16, 2019 Share Posted October 16, 2019 2 minutes ago, aurevo said: Should I do this best with UnBALANCED? That's fine. 2 minutes ago, aurevo said: So first copy, then new configuration without the hard disk. Yes, parity will need to be synced after new config. Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.