Unraid OS version 6.12.8 available

[MPC561]

49 minutes ago, drummerboi said:

Did I do something wrong while the RTL 8125 (B) PCI Drivers package was installed and the setpci commands had no effect?

I should have written that.

It works only with the drivers which are delivered with the Unraid Release. The special RTL driver from the "APP" Package are different. They were made that they also running/working whith "critical" Device/Revisions of RTL Network chips, which normally not running with the normal Unraid RTL drivers. Due to this they not support ASPM really (although it is shown with the lspci instruction). Don´t ask me why this is but you will get this drivers not below C3.ZTh

 

If you don´t see instabilities of your Network connection you can use the "setpci" way to activate ASPM (as you did already).

 

My 2 systems are stable since 2 days with this "hack".

 

 

Long term I expect a clean solution from limetech. As limetch uses the official kernel drivers (at least this was told to me) they are also depending from the Kernel development. So they may react not so fast due to this dependency.

 

So me keeps my Boards. I have 2 small UnRaid server only SSD based with ATOM processors (7/24 running with some dockers) which you can nearly only buy with RTL network chipset. 

 

[Dave-Kay]

Coming from 6.12.7 RC2 i am never getting any further than this:
Does anyone know what to do?
 

[trurl]

33 minutes ago, Dave-Kay said:

Does anyone know what to do?

Without rebooting from that condition, attach diagnostics to your NEXT post in this thread.

[Dave-Kay]

16 minutes ago, trurl said:

Without rebooting from that condition, attach diagnostics to your NEXT post in this thread.

Thx for your help. Here's my diagnosticresult

unraid-diagnostics-20240224-1429.zip

[trurl]

13 minutes ago, Dave-Kay said:

Thx for your help

Nothing obvious.

 

Are you doing this from another computer or from a monitor and keyboard attached to your server?

 

If another computer, what browser? Adblockers or anything else that might interfere?

[Dave-Kay]

1 minute ago, trurl said:

Nothing obvious.

 

Are you doing this from another computer or from a monitor and keyboard attached to your server?

 

If another computer, what browser? Adblockers or anything else that might interfere?

OK, thank you, that was it. Segmented my network recently and did not switch of that for the new IP.

[SShadow]

On 2/17/2024 at 5:08 PM, mrhanderson said:

Anyone running a 13th gen Intel & W680 board combo that has been able to successfully install this update? The previous 6.12.6 update resulted in loss of iGPU access and reboot hangs. Hoping things are better with this update, but also hoping to hear from others before I try again myself.

 

Updated my two servers over the weekend from 6.12.4 to 6.12.8.  First server (Asus WS-C621E-SAGE, Xeon 4216 CPU, ECC RAM) everything worked fine afterwords including my powertop auto tune script and Asus BMC IPMI remote KVM.  Second server (Asus WS-W680-ACE SE, i7-12700K, ECC RAM, 3102 BIOS) not so much.   Like others have reported in the forums with the W680 board I lost my iGPU after updating to 6.12.8 (I had the same issue with 6.12.6 and reverted back to 6.12.4).  Also my powertop auto tune script would not complete and would just hang forever.  The server would not properly reboot or shutdown.  

After looking through many forum posts by people I tried  echo "blacklist ast" > /boot/config/modprobe.d/ast.conf .  This brought back the iGPU in GPU Statistics, the powertop auto tune script would complete normally, and the server would reboot/shutdown properly, but the Asus BMC IPMI remote KVM would stop working (this same behavior others reported in forum posts).

I run my servers headless making remote KVM important to me.  I am not currently using the iGPU on the W680 for anything just testing for possible future use.  I deleted the /boot/config/modprobe.d/ast.conf file.  Then changed my BIOS settings from:
 

Primary Display            [PCIE] 
iGPU Multi-Monitor        [Enabled]
DVMT Pre-Allocated      [64]
RC6(Render Standby)   [Enabled]

to

Primary Display            [Auto] 
iGPU Multi-Monitor        [Disabled]

This appears to have to disabled the iGPU, it no longer shows in GPU Statistics or in system devices but everything works again: powertop script, Asus BMC IPMI remote KVM, and reboot/shutdown.  For now I am going to leave it this way and remove GPU Statisics and Intel-GPU-TOP.  I am sure for most people this is not an option.  I am not sure what changed from 6.12.4 to 6.12.6/8 but something did that looks to affect only W680 boards?  Hopefully the issue gets resolved in a future release.  I hope this post helps.

Edited February 25 by SShadow

[tgiannak]

hello i had problems last days with the server hanging....#W#$ i saw some call traces on the log but i cannot make sense of this..if anyone can help..im posting syslog and diagnostics..also i have updated to 6.12.8 after that..so the diagnostics may reflect this...

 

 

diagnostics-20240226-2028.zip syslog.txt

Edited February 26 by tgiannak

[trurl]

https://docs.unraid.net/unraid-os/release-notes/6.12.8/#call-traces-and-crashes-related-to-macvlan

[tgiannak]

16 hours ago, trurl said:

https://docs.unraid.net/unraid-os/release-notes/6.12.8/#call-traces-and-crashes-related-to-macvlan

im using macvlan without bridging and host access to custom networks on...it isnt right?

 

also what the heck is this???

 

Edited February 27 by tgiannak

[JorgeB]

19 hours ago, tgiannak said:

i saw some call traces on the log but i cannot make sense of this

Call traces appear to be related to the network/NIC, did the server hang after 20:21?

 

Feb 26 20:21:10 Umbrella kernel: RTNL: assertion failed at net/core/dev.c (2918)

 

Also, do you have a different NIC you could test with?

[mrhanderson]

On 2/25/2024 at 3:16 PM, SShadow said:

 

Updated my two servers over the weekend from 6.12.4 to 6.12.8.  First server (Asus WS-C621E-SAGE, Xeon 4216 CPU, ECC RAM) everything worked fine afterwords including my powertop auto tune script and Asus BMC IPMI remote KVM.  Second server (Asus WS-W680-ACE SE, i7-12700K, ECC RAM, 3102 BIOS) not so much.   Like others have reported in the forums with the W680 board I lost my iGPU after updating to 6.12.8 (I had the same issue with 6.12.6 and reverted back to 6.12.4).  Also my powertop auto tune script would not complete and would just hang forever.  The server would not properly reboot or shutdown.  

After looking through many forum posts by people I tried  echo "blacklist ast" > /boot/config/modprobe.d/ast.conf .  This brought back the iGPU in GPU Statistics, the powertop auto tune script would complete normally, and the server would reboot/shutdown properly, but the Asus BMC IPMI remote KVM would stop working (this same behavior others reported in forum posts).

I run my servers headless making remote KVM important to me.  I am not currently using the iGPU on the W680 for anything just testing for possible future use.  I deleted the /boot/config/modprobe.d/ast.conf file.  Then changed my BIOS settings from:
 

Primary Display            [PCIE] 
iGPU Multi-Monitor        [Enabled]
DVMT Pre-Allocated      [64]
RC6(Render Standby)   [Enabled]

to

Primary Display            [Auto] 
iGPU Multi-Monitor        [Disabled]

This appears to have to disabled the iGPU, it no longer shows in GPU Statistics or in system devices but everything works again: powertop script, Asus BMC IPMI remote KVM, and reboot/shutdown.  For now I am going to leave it this way and remove GPU Statisics and Intel-GPU-TOP.  I am sure for most people this is not an option.  I am not sure what changed from 6.12.4 to 6.12.6/8 but something did that looks to affect only W680 boards?  Hopefully the issue gets resolved in a future release.  I hope this post helps.

 

I don't really want to lose KVM access either but I also definitely need iGPU, so I guess I'll continue to remain on 6.12.4 for the time being and hope that eventually the issue will be resolved.

[johnieutah]

Thinking of jumping from 6.10.something up to this. Or should I take an intermediate release first?

 

I‘ve seen macvlan and ipvlan mentioned a few times for things to watch out for but have no idea what this setting is. I haven’t changed anything from the docker defaults I don’t believe…

 

Thanks for any help.

[BRiT]

4 hours ago, johnieutah said:

Thinking of jumping from 6.10.something up to this. Or should I take an intermediate release first?

 

I‘ve seen macvlan and ipvlan mentioned a few times for things to watch out for but have no idea what this setting is. I haven’t changed anything from the docker defaults I don’t believe…

 

Thanks for any help.

 

Last weekend, I upgraded straight from 6.10.x to 6.12.8 without issues.

[johnieutah]

16 hours ago, BRiT said:

 

Last weekend, I upgraded straight from 6.10.x to 6.12.8 without issues.

Just tried updating and it's failing:

 

Mar 2 17:44:07 Toaster kernel: usb 1-1.2: device descriptor read/8, error -110

 

[itimpi]

11 minutes ago, johnieutah said:

Just tried updating and it's failing:

 

Mar 2 17:44:07 Toaster kernel: usb 1-1.2: device descriptor read/8, error -110

 

That looks like a failure reading the flash drive.    You could try this to see if it helps.

[johnieutah]

Tried again and it worked.

 

Do I still need to worry about switching from macvlan or is that issue resolved now?

I am running a fritzbox and use a unifi docker...

 

Thanks.

[itimpi]

1 hour ago, johnieutah said:

Tried again and it worked.

 

Do I still need to worry about switching from macvlan or is that issue resolved now?

I am running a fritzbox and use a unifi docker...

 

Thanks.

If you want to continue using macvlan then make sure bridging is disabled on eth0.

 

[walterg74]

Are there any notes anywhere regarding considerations for the upgrade depending on your current version?

[elcapitano]

Update from 6.12.6 without stopping the array and ran the Docker patch.

4-5 min total - no issues.

Thanks guys.

[johnieutah]

17 hours ago, itimpi said:

If you want to continue using macvlan then make sure bridging is disabled on eth0.

 

Are there any significant differences? Everything appears to be working on first glance. Only issue I came across was with the Unifi docker and the host name not being resolved. Using the ip address has fixed it for now.

[seestray]

Went from 6.11.5 --> 6.12.8 with the realtek RTL8125, macvlan and a unifi docker that's on 8.0.24.

 

An hour in and all appear to be working at first glance - we'll see over the next couple of weeks if it remains stable.

[DivideBy0]

Well, I regret to inform you that this upgrade broke my CPU iGPU  on my ASUS Pro WS W680M-ACE SE which is pretty disappointing as I can't transcode anymore with Plex.  I guess something in the new kernel?

 

UPDATE: And never mind the fix was right here:

https://forums.unraid.net/topic/153787-unraid-os-version-6128-available/?do=findComment&comment=1377747

 

echo "blacklist ast" > /boot/config/modprobe.d/ast.conf

 

But I guess my KVM remote is broken now? 

Edited March 6 by DivideBy0

[walterg74]

I am running an old Version: 6.9.1

 

Only extras are Duckdns, Plex, binhex-minidlna, and OpenVPN (yes, I realize this last one is now obsolete and I should look for an alternate option).

 

Any considerations for the upgrade, or can I/should I just jump to 6.12.8 ?

[DevXen]

i didn't see it in the release notes. but does this resolve the adaptec 7000 series card issues that .6 had. i had to roll back to .4 to fix it,. it was an issue with the then current version of the kernel i guess. but if it's not fixed then I guess i'm stuck on .4 still.