Any ideas? This is rather confusing to me, that it's offline in Unraid but passes the SMART extended self-test. Attached is the full report. Any help is appreciated!

 

Thank you.BadDrive.zip

This generally means a write to the disk failed and it needs to be rebuilt.

Please post the complete diagnostics, ideally after the disk was disabled and before rebooting: Tools - > Diagnostics

1 minute ago, francrouge said:

Here you go

 

thx

nas-diagnostics-20180121-1918.zip

That was to the OP, already responded to you on your thread.

Unraid v6.4. Attached are the diagnostics as requested, although I've already done a reboot since the error first popped up, I'm afraid. I shut down to check cables and connections. Thank you!

 

library-diagnostics-20180122-0755.zip

Edited January 22, 20188 yr by Flick

Will leave to Johnnie to look at the logs, but just wanted to reply on how a perfectly good disk can drop offline.

 

There are basically two common scenarios:

1 - The disk has a cabling issue, and under load the drive looses its connection to the controller. Once this happens the disk is inaccessible. Often a reboot will restore the connection. An extended test DOES NOT require contact with the controller (only to kick it off), so is not a good way to detect cabling issues. But if you start to see CRC errors on the SMART report, that is a very good indicator of cabling problems

 

2 - There are some controllers (Marvell) that will cause drives to drop offline, with no cabling problem. It is recommended to replace with LSI controllers (e.g., SAS9201-8i). Some users don't have issues while others do.

Without a pre-reboot syslog we can't see what happened, but the disk looks healthy so you'll need to rebuild, either using a new disk or the old disk, since you have dual parity it's not so risky to use the old one, just make sure that the contents of the emulated disk look correct, since whatever's there is what's going to be on the rebuilt disk, also I would recommend swapping/replacing cables/backplane slot just to rule that out in case the same disk fails again.

And I would just add another thing to consider for future support.

 

Don't hijack another user's support thread. Even if you have what you think is the same problem. Even if it really is the same problem. The details of your system will be different, and the details of the solution may turn out to be different. There is no good reason to create confusion by mixing up questions and answers for different users, especially when data is on the line.

 

/badcop

2 hours ago, trurl said:

And I would just add another thing to consider for future support.

 

Don't hijack another user's support thread. Even if you have what you think is the same problem. Even if it really is the same problem. The details of your system will be different, and the details of the solution may turn out to be different. There is no good reason to create confusion by mixing up questions and answers for different users, especially when data is on the line.

 

/badcop

 

Um, dude? This is *my* thread.

2 hours ago, johnnie.black said:

Without a pre-reboot syslog we can't see what happened, but the disk looks healthy so you'll need to rebuild, either using a new disk or the old disk, since you have dual parity it's not so risky to use the old one, just make sure that the contents of the emulated disk look correct, since whatever's there is what's going to be on the rebuilt disk, also I would recommend swapping/replacing cables/backplane slot just to rule that out in case the same disk fails again.

 

Yeah, I was afraid of that. Normally I'm pretty good at working through all the steps but I simply missed doing so this this time. I appreciate the feedback. I've never had to rebuild an existing disk; could you point me towards instructions to do so? I've only put in new disks in the past but, in this case, the drive has only been going for three months so I'd like to give it another shot.

 

What would be ideal, actually, would be if I could put it into a different slot. That way, should it fail again, it points to disk rather than cables/connections. Thoughts? I did go in and clean all cables, reseat connections, etc. This has been a rock solid system for several years now.

 

2 hours ago, SSD said:

Will leave to Johnnie to look at the logs, but just wanted to reply on how a perfectly good disk can drop offline.

 

There are basically two common scenarios:

1 - The disk has a cabling issue, and under load the drive looses its connection to the controller. Once this happens the disk is inaccessible. Often a reboot will restore the connection. An extended test DOES NOT require contact with the controller (only to kick it off), so is not a good way to detect cabling issues. But if you start to see CRC errors on the SMART report, that is a very good indicator of cabling problems

 

2 - There are some controllers (Marvell) that will cause drives to drop offline, with no cabling problem. It is recommended to replace with LSI controllers (e.g., SAS9201-8i). Some users don't have issues while others do.

 

1. Ah, I did not know that about the SMART kick off. TIL. Thanks!

 

2. Shouldn't be an issue, this has been a controller in play for many years with no issues.

 

Thank you for the assistance. Love this community!

11 minutes ago, Flick said:

Um, dude? This is *my* thread.

Of course it is. Did you not see that another user had  posted their diagnostics on your thread?

4 minutes ago, Flick said:

I've never had to rebuild an existing disk; could you point me towards instructions to do so?

https://lime-technology.com/wiki/Troubleshooting#Re-enable_the_drive

 

4 minutes ago, Flick said:

What would be ideal, actually, would be if I could put it into a different slot. That way, should it fail again, it points to disk rather than cables/connections. Thoughts? I did go in and clean all cables, reseat connections, etc. This has been a rock solid system for several years now.

Exactly, so if it fails again there could really be an issue with the disk, SMART is a good indication but an healthy SMART doesn't always equal an healthy disk.

1 hour ago, trurl said:

Of course it is. Did you not see that another user had  posted their diagnostics on your thread?

Yup, I did and was ignoring it. Since you didn't quote, I assumed you'd gotten us mixed up as I was the last "non-support person" to post in the thread. No worries, mate, I appreciate you helping to keep the forums clean!

 

1 hour ago, johnnie.black said:

https://lime-technology.com/wiki/Troubleshooting#Re-enable_the_drive

 

Exactly, so if it fails again there could really be an issue with the disk, SMART is a good indication but an healthy SMART doesn't always equal an healthy disk.

 

Perfect, I'll give this a go this evening. Thank you!

Done, looks like all went well. Fingers crossed!

 

Event: unRAID Parity sync / Data rebuild
Subject: Notice [LIBRARY] - Parity sync / Data rebuild finished (0 errors)
Description: Duration: 1 day, 5 hours, 52 seconds. Average speed: 57.5 MB/s
Importance: normal

Hi, I have the same issue - link above no longer works and I suck at searching - tried, no luck. Any updated link, or revised instructions for this scenario?

 

thanks

tiwing

Edited July 6, 20206 yr by tiwing

6 minutes ago, tiwing said:

Hi, I have the same issue - link above no longer works and I suck at searching

If you mean the link to re-enable a drive it's below, but make sure the drive is healthy and the emulated disk is mounting correctly and contents look correct before rebuilding on top.

https://wiki.unraid.net/Troubleshooting#Re-enable_the_drive