Currently a Parity-Check is running and its only hitting about 381 kbs. I cannot copy or move data on the system. I have disabled all my docker containers and it made no change. My CPU's are flip flopping between them and hitting 100%. From 2% to 100 in a second or two. In my trying to figure out logs, it looks like possibly a controller issue? This is all a bit over my head, any help would be greatly apprciated.

Thanks

tower-diagnostics-20261110-0443.zip

Edited 4 hours ago4 hr by ProblemWithField

Disk5 has a connection problem. Since boot, its link has failed and reset about every 30 seconds:

ata30: SError: { RecovComm PHYRdyChg CommWake 10B8B DevExch }

ata30.00: failed command: READ DMA EXT

ata30: hard resetting link

A parity check needs all the disks, so the whole array, including copies and moves, slows down to the speed of that link. The disk's SMART looks fine. These errors usually come from a bad SATA cable or a bad power connection.

What I recommend:

1. Cancel the parity check that is running now. At this speed it cannot finish.

2. Shut down the server. Replace disk5's SATA cable, and use a different power connector for it. If you use a power splitter or adapter, remove it.

3. While you are in the BIOS, set the onboard SATA controller to AHCI mode. It is set to IDE now. This is safe for the array, because Unraid identifies the disks by serial number.

4. Start the server, wait a few minutes, and check the syslog for more ata errors. If they continue, move disk5 to a different port to see if the problem follows the disk or the port.

5. When the log is clean, run a parity check again. It needs to be a correcting check, because there was an unclean shutdown.

Also, disk6 is connected through a port multiplier on the ASM1061 controller, and it is logging errors too. We do not recommend port multipliers. If you can, connect that disk directly to a port, or use an HBA.

After the changes, please post new diagnostics.

Thanks VERY much I will work on the changes you shared. Is there an HBA you recommend, I would like to get one ordered asap. I will work on this when I get back home tonight. Thank you so much.

You can see here: https://forums.unraid.net/topic/102010-recommended-controllers-for-unraid/

Actually I had time this morning. I see now I have read errors on disk 4, I haven't done anything yet, I wanted to see what your fedback was. I replaced the cable on disk 5. As for the controller that disk 6 was on, I realized I had a difffernt controller that had an open port. I am not sure if thats HBA controller or not. Let me know and I can upgrade if needed. I also made the change to the BIOS.

tower-diagnostics-20261110-0638.zip

Looks I am still geting 5% to 100% CPU hits as well. It seems to only be when I reinable the docker.

Did reseat the cable on drive 4 that had the error, and it seems to be corrected. I am not sure if I should spin up the dockers again. I also cannot copy any data to the NAS from elsewhere via drag and drop. It gets till it's just about finished and then erros out.

Edited 1 hour ago1 hr by ProblemWithField

Good progress. Disk5 is clean now, and disk6 is also clean on its new port.

The card you used is an ASM1064. It is not an HBA. The ASM1064 chip has only 4 SATA ports, but your card shows 6 to Unraid. On cards with more than 4 ports, the extra ports are usually behind a port multiplier, which we do not recommend. How many SATA ports does the card have? If it has 4, it is fine to use. If it has more than 4, it is not a good option. In that case, consider an HBA from the recommended controllers thread.

The new problem: disk4, disk1 and parity, all on the onboard SATA ports, now show the same kind of connection errors that disk5 had. Disk4 had the most. Its link reset many times and dropped to the lowest speed. Parity and disk1 had fewer resets. This started right after the work inside the case, so the most likely cause is that their cables or power connectors were moved or loosened. Another possibility is that these disks share a power cable or splitter that cannot supply them reliably.

About the disk4 read errors: they were caused by the connection drops, not by the disk. SMART for disk4 is clean. Unraid read that data from parity and wrote it back to the disk, so no data was lost and the disk stays enabled. The error count will clear on the next reboot.

What I recommend:

1. Shut down the server.

2. Reseat both ends of the SATA and power cables for disk4, disk1 and parity. Replace disk4's SATA cable if you have a spare.

3. If these disks share one power cable from the PSU, or use a splitter, spread them over different PSU cables if you can.

4. Start the server and leave the array idle for a few minutes. Then copy a large file to the array, or start a parity check briefly, so the disks get some load. Then download and post new diagnostics.

5. When the log is clean, run a full correcting parity check.

Sorry to bug you, I just want to make sure I do this right. This controller has the ASM1064 and only 4 ports,

https://www.amazon.com/10Gtek-Profile-Bracket-Controller-Expansion/dp/B09Y1NRHX3/ref=sr_1_1?crid=1PQ4EG4VDPNIL&dib=eyJ2IjoiMSJ9.2sWWGGvvvc-mX1oQm-rFM1FeN23Sl1zjP5ZTX0hjBVb45zwytoQpNny6SIckaA0vBLKMgiocspqXZSKcx0TsAVmgWfcdWR_1riYU2AFTRoVgj09pG562GgvvVIey0lI_Q_DzwfKtd2YEOo22PhmSm3F5NENxKidRAFZFk5e83zkS2-UNO8Yj-mGLb3GppXqqXXgtAULqkHWRv_pn9gbiALYhIVaf6yeHjI4s44NDz-g.jEOMtl1tZFwmWPrLS2bYfiPaT1JIILcqQ03LWcEUxjg&dib_tag=se&keywords=Asmedia%2BASM1064&qid=1794324336&sprefix=asmedia%2Basm1064%2B%2Caps%2C188&sr=8-1&th=1

The first diags were showing a port multiplier installed. Was there something between the disk and the controller:

Nov 9 19:19:21 Tower kernel: ata26.15: Port Multiplier 1.2, 0x197b:0x5755 r0, 5 ports, feat 0x5/0xf

1 minute ago, JorgeB said:

The first diags were showing a port multiplier installed. Was there something between the disk and the controller:

Nov 9 19:19:21 Tower kernel: ata26.15: Port Multiplier 1.2, 0x197b:0x5755 r0, 5 ports, feat 0x5/0xf

Interestingly the description in the provided link talks about port multiplier command compatibility. I wonder quite what that means?