Jump to content

TexasUnraid

Members
  • Content Count

    413
  • Joined

  • Last visited

Community Reputation

32 Good

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. It has a 320w gold power supply and rarely pulls over 150w from the wall according to the watt meter. The cable that this drive plugs into also powers 2 other array drives, I suppose it could just be this connector but kind of strange. I suppose I could swap the power slitter from those drives and see if things improve, it is the good kind without the molded ends. I am using an HBA and the breakout cable connects to 4 array drives. It has also never given me any issues in the past and the cable was new 6 months ago. The drive seems to of completely disappeared from unraid which is interesting since I can't even get a smart report (same as before), should come back when I reboot though. Thinking I will replace the power splitter and swap this drive with another one and if the issue happens again I will know if it is the drive or the connections.
  2. Ok, so rebuilt the drive and everything worked fine for a few days. Then I decided to do another file transfer like I was doing last time and the same disk dropped out again? In both cases I noticed that it seemed to happen during the file transfer from disk to disk when other stuff was also going on with the array. It is really strange since this drive has been working fine for the last 6 months in exactly the same setup, same cable, same card, same port. The error is the same thing as before but with different sectors: Oct 28 19:48:20 NAS kernel: sd 1:1:3:0: [sde] tag#938 UNKNOWN(0x2003) Result: hostbyte=0x01 driverbyte=0x00 cmd_age=9s Oct 28 19:48:20 NAS kernel: sd 1:1:3:0: [sde] tag#938 CDB: opcode=0x88 88 00 00 00 00 03 da 3d 15 a8 00 00 00 68 00 00 Oct 28 19:48:20 NAS kernel: blk_update_request: I/O error, dev sde, sector 16546338216 op 0x0:(READ) flags 0x0 phys_seg 13 prio class 0 Oct 28 19:48:20 NAS kernel: md: disk3 read error, sector=16546338152 Oct 28 19:48:20 NAS kernel: md: disk3 read error, sector=16546338160 I can RMA the drive (WD 12TB white label shucked drive) but that takes time and there is zero evidence that there is something wrong with the drive, so not even sure if an RMA claim will be approved?
  3. I finally had a little time to try out this script, had the same issues with it not being able to handle shares with spaces in the name. Oh well, I will just stick to the setup I am using now and see if someday a plugin is made. I do wish I could get the previous versions working in windows though.
  4. lol, a watched pot never boils, I get it. I think it was just leftover boot up tasks that were slowing it down, after 15 mins or so it settled down and has been going at the expected speeds since. The hardware is not fast and about 8 years old but fast enough, It is an old Del optiplex 7010 motherboard I hoodwinked into an ATX case. Sitting at around ~15% usage for the rebuild alone, ~20-25% with the dockers running.
  5. Thanks for the help everyone, I decided to just go ahead and rebuild the drive, in theory if there are any issues with the rebuild a BTRFS scrub should find them I realized. It is rebuilding now, I did notice that it does not seem to be maxing out the speed of my drives like a parity check, is that normal? Usually they start off around 200MB/s for a parity check but the rebuild is starting off around 150MB/s? CPU usage is not the issue at ~20% usage. Looking at the netdata stats, the IO is going up and down a lot, I assumed it would be like a parity check, pretty much stable speeds slowly dropping over time. Not a big deal but was just curious if that is normal or if there is an issue? Edit, speeds seem to be getting better as things progress, guessing it was just some kind of disk access left over from the startup.
  6. Ok, guess I will start the rebuild process. Although I have not run a parity check in ~2 months and have added 2 drives in that time, any reason to suspect parity would not be in sync? The process for a rebuild is still to: remove drive from array start array stop array add drive to array start array? Anything I should be aware of or do first? Still pretty new to unraid.
  7. That is strange since I just checked and debug logging is disabled in the settings? Yeah, I didn't expect the plugin to be able to do the first part. The second part of just straight up disabling all IO to the array (and maybe docker) on a drive failure, that is what I was thinking might be possible with a plugin?
  8. Yep, got an email about the issue and checked it right away. I didn't think the parity tuning plugin was an issue, just found it odd that message was spammed just before the issue as I had not seen that happen before. Given the choice, I would like to have the array setup so that if a disk was going to be disabled, all IO would be immediately halted and any pending operations written to some kind of log file. The idea being that in a situation like this where it was just a glitch in the system, things would not get out of sync and I could simply restart the server and if the disk checked out start up the array like normal. The log /temporary file could then be used to finish the pending operation when the drive was disabled. You can then run a parity check of course. Short of that, I still think I would prefer to have the array disabled completely should a disk be disabled with the option to re-enable it if I want. For example today this disk was disabled during a file move process, that process kept going after the drive was disabled and was trashing the drives as it tried to emulate the missing disk and write to another disk. It had another ~400GB to go in the move, that is a lot of strain to put on disks just before a rebuild operation had I not been here to stop it. Most important for me would be to stop any activity to the disks and possibly shut down docker until I can manually get on the server to sort things out. Last thing I would want to do is thrash drives just before counting on them to rebuild a drive.
  9. Thanks for the info. Yeah, I didn't run the extended test as it took it overnight to run last time I did it on a 12TB disk. These large disks are great on one hand but it takes forever to do any kind of big task. I did check all of the connections when I put it back together and even now didn't find any obvious cable issues with this drive (found a semi loose cable on another drive that threw a UDMA error the other day though). I just unplugged and replugged all the cables and hoping it is good to go. It has been running with this drive for the past 5 months or so 24/7 without an issue so pretty sure it is a cable issue as I have to re-wire most of the case. Yes, you are spot on, I was in the process of moving some data from disk 3 to disk 4 when the error happened. The data is just media files, I can easily replace them, I just have to know which ones need to be replaced. Thats why I was hoping I could compare the emulated disk to the physical disk but rebuilding might be best anyways since I would have to do a parity check either way. This particular disk is basically only backed up to the parity since it is just media that is replaceable (although not something I would want to do). I have all my important data on other drives and they are backed up to a few different places but can't afford the cost of backup drives for the media.
  10. I sent a PM with the post reboot diagnostics file. I have the server setup to save the diagnostics file on shut down but does it anonymize when it does this? Got to run an errand but pretty sure it was just a cable issue as I made quite a few changes inside the case a few days ago (added a 3.5" drive cage and it needed a bunch of changes in the case to make it fit).
  11. This is really inconvenient, is there a way to set the array to disable completely if a drive fails like this so that it would not get out of sync and thus could simply be re-added without a fuss? with 24 hour+ rebuild times rebuilding is not a simple thing.
  12. Just rebooted, Smart report looks good just like the other matching drives and smart test is good. So yeah, looks like it was just a temporary issue. I would really prefer not have to rebuild the whole drive, I really need to use the server.
  13. lol, it has been this way for the last week at least, possibly longer. I figured it was just a temporary issue at first as well.
  14. I tried a half dozen this morning but at best I was seeing ~10MB/s when I used to see 25-35MB/s
  15. Yep, switched to the new network and have tried going back to the old network as well but all servers I have tried seem to be much slower then before.