6.11.5 - No exportable shares???

[johnny2678]

I've been running Unraid without issue for 3 years.  Then I try to upgrade to 6.12.4 and have had nothing but problems since.  Got hit by the macvlan bug, so I went back to 6.11.5 where I thought it was safe.  But now all my shares disappear every few days.  When I go to the shares tab it says something like "No exportable shares".  Data is fine but most dockers/vms stop working and it takes a reboot to get the shares back.  My uptime sux now.  Can anyone ELI5 what's going on??  TIA

 

15620-beast-diagnostics-20231024-2042.zip 

Edited October 25, 2023 by johnny2678

[JorgeB]

Log is being spammed with NIC issues, difficult to look for anything else, I would suggest upgrading back to v6.12.4, see the release notes for how to fix the macvlan issue.

[johnny2678]

Thanks @JorgeB, the 6.12.4 macvlan fix didn't work for me. Despite following it to the letter, I would still get frequent lockups that would force a hard restart.  That's why I went back to 11.5.  I never really needed macvlan (was just the unraid standard when I started on 6.8), so when the system still wasn't stable after reverting to 11.5, I switched to ipvlan.  Now I don't have to hard restart, but I'm getting the "no exportable shares" error every so often. 

 

Frustrating for a system that used to stay up for months at a time with no issues.  I will try 6.12.4 the next time it goes down and report back.

[JorgeB]

3 minutes ago, johnny2678 said:

macvlan fix didn't work for me.

That is odd since AFIK it worked for everyone else, did you try 6.12 with ipvlan? If it still locks up macvlan is not the problem.

 

 

[johnny2678]

1 hour ago, JorgeB said:

it worked for everyone else

You found my special talent.  Somehow finding the loophole and breaking stuff that just works for everyone else 😂

 

I'm pulling the array down and going back to 6.12.4.  Will report back.

[johnny2678]

@JorgeB just rebooted to see if the NIC errors cleared and noticed this in the logs.  Does this say anything to you?

 

Oct 25 08:08:16 15620-BEAST kernel: ata1.00: exception Emask 0x10 SAct 0x4000 SErr 0x280100 action 0x6 frozen
Oct 25 08:08:16 15620-BEAST kernel: ata1.00: irq_stat 0x08000000, interface fatal error
Oct 25 08:08:16 15620-BEAST kernel: I/O error, dev sdb, sector 520 op 0x0:(READ) flags 0x80700 phys_seg 37 prio class 0
Oct 25 08:08:16 15620-BEAST kernel: ata1.00: exception Emask 0x10 SAct 0x8000 SErr 0x280100 action 0x6 frozen
Oct 25 08:08:16 15620-BEAST kernel: ata1.00: irq_stat 0x08000000, interface fatal error
Oct 25 08:08:16 15620-BEAST kernel: ata1.00: exception Emask 0x10 SAct 0x8000 SErr 0x280100 action 0x6 frozen
Oct 25 08:08:16 15620-BEAST kernel: ata1.00: irq_stat 0x08000000, interface fatal error
Oct 25 08:08:16 15620-BEAST kernel: ata1.00: exception Emask 0x0 SAct 0x10000 SErr 0x880000 action 0x6
Oct 25 08:08:16 15620-BEAST kernel: ata1.00: error: { ICRC ABRT }

 

[JorgeB]

21 minutes ago, johnny2678 said:

{ ICRC ABRT }

This is usually a bad SATA cable, replace it for that device.

[johnny2678]

2 minutes ago, JorgeB said:

This is usually a bad SATA cable, replace it for that device.

/dev/sdb?  That's an NVME drive

[JorgeB]

/dev/sdb cannot be an NVMe device, if you post the diags we can tell you which device ATA1 is.

[johnny2678]

18 minutes ago, JorgeB said:

/dev/sdb cannot be an NVMe device, if you post the diags we can tell you which device ATA1 is.

 

Sorry, you are correct.  It's a SSD.  I'll check the cable/replace next time I open the box.

 

I haven't moved to 6.12.4 yet but curious if you can tell if the logs are still being spammed with NIC issues?  I rebooted hoping they would clear so we could see if there were any other issues that might cause the "no exportable shares" issue.  Worried if I move to 6.12.4 I'll just take the existing problems with me.

 

 

15620-beast-diagnostics-20231025-0912.zip

[JorgeB]

1 hour ago, johnny2678 said:

if you can tell if the logs are still being spammed with NIC issues? 

Nothing so far.

[johnny2678]

Just going to throw this out there - I run the official plex docker.  For a few days now, I've watched as it constantly runs transcode operations on the same show (kids show/~50 short episodes).  Runs for hours, then stops, then runs again the next day on the same show - was guessing this was due to the plex maintenance window.  I have >20TB of media, but plex stays hyperfocused on this show for some reason.

 

At the same time, I've been having problems with my plex backup script.  It keeps getting hung up on the bundles under:

config/Library/Application\ Support/Plex\ Media\ Server/Media/localhost/

 

It never completes.

 

Today, I tried to delete the offending show - never really cared about it anyways.  Then I went to Plex Setup -> Troubleshooting -> Clean Bundles... Plex has been "Deleting 51 Media Bundles for over 2 hours now:

 

 

The underlying disk(s) is an SSD cache pool in Raid0 (sdb & sdc):

 

 

During this 2+ hour bundle delete operation, the read counter is wildly out of sync on the cache pool between these two devices:

 

Is my BTRFS messed up?  Could that be causing my issues?  Adding a current diagnostic if it helps.  thx again for looking @JorgeB.  Would be happy to try any ideas you have to unstick this.

 

15620-beast-diagnostics-20231026-1022.zip

[JorgeB]

Pool looks fine, it's using raid0 but I assume that was a choice.

[johnny2678]

6 minutes ago, JorgeB said:

Pool looks fine, it's using raid0 but I assume that was a choice.

Any ideas on why the write counts are so lopsided?

 

Any ideas on why tar-ing (or even rsync-ing) plex bundles on the cache pool never finish?

 

Again, thx for the attention. Not your job to fix my stuff.  Nice to have someone to bounce ideas off of.

[JorgeB]

21 minutes ago, johnny2678 said:

Any ideas on why the write counts are so lopsided?

Metadata is raid1, so reads can just use one of the devices.

 

22 minutes ago, johnny2678 said:

Any ideas on why tar-ing (or even rsync-ing) plex bundles on the cache pool never finish?

I don't use Plex, don't know what Plex bundles are, but if they are made of a lot of small files it can take a while.