Fix Unraid 6.6.0 the AMD GPU reset Bug in Linux Vm´s.?  (or with Kernel 4.19)

After reboot any Linux VM or shutdown. Unraid Server must reboot complete every time (I´ve Try a RX480 and Rx580)

Windows 10  VM´s with GPU Pass works great.

11 minutes ago, mostlydave said:

but is this a bug?

No, it is by design and prevents users to install outdated versions.

For example if you have set a weekly check schedule, it can happen that between the check which reports version X and the actual plugin installation there are intermediate updates done and a newer version is available. You don't want version X to be installed but the latest version at the moment of installation.

 

When the plugins page is opened, a one time internet connectivity test is done, before the information about the latest versions is retrieved unless you have opted for a manual check.

 

Ps. If you have a speedy Internet connection, the impact is minimal.

Edited September 20, 20187 yr by bonienl

There's a process running over and over on my Unraid system to gather docker container's CPU and memory usage.  The process uses about 9% - 10% CPU every time it gathers the usage, probably since I'm running so many containers.  The load is more than average of all my docker containers combined. I assume it's to provide the new feature docker webui advanced view usage display. Is there a way to stop this process from running?  Could it be set to start only when I need to view the data it provides?  Is it running constantly to capture data and save it to be viewed for historical observation? 

I'm not sure if this is a docker default thing or its always been there.

Thank you for the help and great update.

The process is started when the docker service starts and stopped when the docker service stops. It runs in the background to keep track of container cpu load and memory load.

 

You can manually start and stop this process from terminal (CLI)

/usr/local/emhttp/plugins/dynamix.docker.manager/event/docker_started/docker_load_start
/usr/local/emhttp/plugins/dynamix.docker.manager/event/stopping_docker/docker_load_stop

 

4 minutes ago, bonienl said:

The process is started when the docker service starts and stopped when the docker service stops. It runs in the background to keep track of container cpu load and memory load.

 

You can manually start and stop this process from terminal (CLI)

/usr/local/emhttp/plugins/dynamix.docker.manager/event/docker_started/docker_load_start
/usr/local/emhttp/plugins/dynamix.docker.manager/event/stopping_docker/docker_load_stop

 

Thank you! Worked like a charm.

I have upgraded the 6.6 version but now i can"t access the server neither on PC nor webGUI. Then i downgraded to 6.5.3 but the problem remains the same. It was working extremely fine before. Now i regret the upgrading. I need help getting it running again. Anyone? 

4 minutes ago, Lalit said:

I have upgraded the 6.6 version but now i can"t access the server neither on PC nor webGUI. Then i downgraded to 6.5.3 but the problem remains the same. It was working extremely fine before. Now i regret the upgrading. I need help getting it running again. Anyone? 

Just to clarify, when it was working before was that on 6.5.3?

 

Do you have an attached keyboard and monitor?

7 minutes ago, Lalit said:

I have upgraded the 6.6 version but now i can"t access the server neither on PC nor webGUI. Then i downgraded to 6.5.3 but the problem remains the same. It was working extremely fine before. Now i regret the upgrading. I need help getting it running again. Anyone? 

How do you try to access the server? By name (e.g. \\tower) or by IP address (e.g. \\192.168.1.2).

Maybe your server obtained a different IP address after the upgrade, and the name still refers to the old IP address.

 

Upgrade from 6.5 to 6.6 in general goes without major incidents.

4 minutes ago, bonienl said:

How do you try to access the server? By name (e.g. \\tower) or by IP address (e.g. \\192.168.1.2).

Maybe your server obtained a different IP address after the upgrade, and the name still refers to the old IP address.

 

Upgrade from 6.5 to 6.6 in general goes without major incidents.

i generally use name to access server. After this problem i set my ip to static and then tried to access by ip address but no luck.

Edited September 20, 20187 yr by Lalit
addition to the reply

Now i'll try to upgrade again and see if it fixes the issue.

Can you "ping" to the IP address of your server, is there any response?

2 minutes ago, bonienl said:

Can you "ping" to the IP address of your server, is there any response?

I tried pinging to the IP address but no response.

2 minutes ago, Lalit said:

I tried pinging to the IP address but no response.

Now i upgraded the os to 6.6 again and suddenly it's working just fine. I don't know what the issue was but now it seems to work fine. I can access the server on PC and on webGUI too. Thanks for all the support.

1 minute ago, Lalit said:

Now i upgraded the os to 6.6 again and suddenly it's working just fine.

I have no clue what happened, but miracles still exist 😄

32 minutes ago, bonienl said:

I have no clue what happened, but miracles still exist 😄

Yeah, and i am fortunate that this miracle happened to me.

 

upgraded z420 and ml30 w/out any apparent issues. 

 

 

COUGHmultiplecachepoolspleaseCOUGH

Upgraded from 6.5.3 to 6.6.0 and worked flawlessly.

 

The only question I have is how does CPU pinning work?

Before upgrading from 6.5.3 to 6.6.0 Quick question...previously I set the following in my syslinux.cfg file manually:

 

append pcie_acs_override=downstream vfio-pci.ids=10de:1402,1b73:1100,10de:13c2 isolcpus=1,2,5,6 initrd=/bzroot

With the new CPU pinning do I remove the isolcpus setting?

Same question for pcie_acs_override?  

 

How about the vfio-pci.ids for my 2 video cards and USB3 card?   Don't recall hearing if this is now in the GUI???

 

Thanks,

Dan

4 hours ago, bonienl said:

The process is started when the docker service starts and stopped when the docker service stops. It runs in the background to keep track of container cpu load and memory load.

 

You can manually start and stop this process from terminal (CLI)

/usr/local/emhttp/plugins/dynamix.docker.manager/event/docker_started/docker_load_start
/usr/local/emhttp/plugins/dynamix.docker.manager/event/stopping_docker/docker_load_stop

 

Thanks.  I'm going to add the stop command to my startup script as I just ran it manually out of curiosity and it was using around 7-10% of my CPU, when most of my dockers were off - I imagine it'd be higher if more were running.  I'll fire it up when I need to investigate odd behaviour

well CPU and Memory load in dockers is a bit sexy!

 

With so many little GUI tweaks it's hard to keep up with it and there seems to be these lovely little surprises when i click another button

 

It wold be really nice to customise what is shown on dockers/VMs tabs. Basic view is too basic, and Advanced view is too crowded for every day use. Stuff that changes like loads is really useful but a long list of ports isn't so useful unless you're trying to troubleshoot.

 

Amazing work guys, i've even reverted back to the default GUI. Spacing def needs tidying up, but really a great job.

 

Many thanks for your continued development!

Edited September 20, 20187 yr by Ockingshay

49 minutes ago, goinsnoopin said:

Before upgrading from 6.5.3 to 6.6.0 Quick question...previously I set the following in my syslinux.cfg file manually:

 

append pcie_acs_override=downstream vfio-pci.ids=10de:1402,1b73:1100,10de:13c2 isolcpus=1,2,5,6 initrd=/bzroot

With the new CPU pinning do I remove the isolcpus setting?

Same question for pcie_acs_override?  

 

How about the vfio-pci.ids for my 2 video cards and USB3 card?   Don't recall hearing if this is now in the GUI???

 

Thanks,

Dan

 

No. You can leave those alone. The CPU pinning and Docker pinning in the GUI won't touch those settings. Emulator pinning in your VM XML files (if you have them in there) will be left intact too if you edit your VM with the GUI.

Edited September 20, 20187 yr by GHunter

Upgrade results in none functional web ui. With or without SSL, page never loads after login via direct ip.

No errors in nginx error.log.

Edited September 20, 20187 yr by aterfax

7 minutes ago, aterfax said:

Upgrade results in none functional web ui. With or without SSL, page never loads after login via direct ip.

No errors in nginx error.log.

Try loading without plugins

13 minutes ago, Jerky_san said:

Try loading without plugins

Safe mode? Or is there a specific way to do so?

Removed all plugins - now have access via chrome on my phone but not from my desktop in Edge, Chrome or Firefox.

Has something like fail2ban been added that has blacklisted my IP?

Edited September 20, 20187 yr by aterfax

6.5.3 to 6.6.0 went smoothly. The new theme will take a bit to get used to. I do agree with a comment earlier that it seems to take up quite a bit of vertical space.

 

Only one real issue I noticed:

 

That white on white is a bit tough to read. Maybe just go with the same dark grey as is behind the User Key entry box. I didn't check any of the other notification agent setups, but I'd imagine they're all using the same configuration info.

 

As always, I could wait patiently for the unRAID team to take some time to sort out Microsoft's Win10 update process for them, they've got it nailed for unRAID!