When i start a VM my dockers become unusable

[Mirano]

Let me explain this weird situation i just cannot resolve.

 

I have a PC with 16GB Ram, 1 NIC, i5-6500. I have like 10 dockers running on the first 2 core, with the last 2 cores excluded exclusivly for VM's. I start up my VM (doesn't matter if its ubuntu or windows) and when its on my dockers become weirdly unusable. For example:

 

My Pi-Hole just stops answering to DNS requests

My Unifi controller sets all the access points in the controller to "Adopting"

 

When i stop my vm and its shutdown my Pi-Hole starts to run again and all the acess points on the unifi controller says connecte again.

 

This i just one thing i cannot resolve myself. Host access to custom networks is off btw, every dockers has its own IP\

 

Diagnostics is included. Hopefully someone can help me out

 

Edited June 24, 2020 by Mirano

[PeteAsking]

Make sure your vm’s are not running on the same disk as your dockers and retest. 

[Mirano]

2 minutes ago, PeteAsking said:

Make sure your vm’s are not running on the same disk as your dockers and retest. 

My vms vdisks are on a unnassigned ssd and docker are on the array

 

You have any other ideas?

[PeteAsking]

Can you temporarily disable cpu isolation and retest. You can leave the vm’s pinned to the last 2 cores. 

[Mirano]

1 minute ago, PeteAsking said:

Can you temporarily disable cpu isolation and retest. You can leave the vm’s pinned to the last 2 cores. 

It’s rebooting now!

[Mirano]

21 minutes ago, PeteAsking said:

Can you temporarily disable cpu isolation and retest. You can leave the vm’s pinned to the last 2 cores. 

Starting Ubuntu resolves into acces points going to adopt stage and pi hole not responding, this did not work unfortunately!

[PeteAsking]

Hmm ok I will think some more then. 

[PeteAsking]

It looks from the logs that you have promiscuous mode on. Can you try disabling that since the issues are network related?

 

un 17 23:17:09 HPUnraid kernel: br0: port 2(vnet0) entered disabled state
Jun 17 23:17:09 HPUnraid kernel: device vnet0 entered promiscuous mode
Jun 17 23:17:09 HPUnraid kernel: br0: port 2(vnet0) entered blocking state

Edited June 18, 2020 by PeteAsking

[Mirano]

On 6/18/2020 at 3:21 AM, PeteAsking said:

It looks from the logs that you have promiscuous mode on. Can you try disabling that since the issues are network related?

 

un 17 23:17:09 HPUnraid kernel: br0: port 2(vnet0) entered disabled state
Jun 17 23:17:09 HPUnraid kernel: device vnet0 entered promiscuous mode
Jun 17 23:17:09 HPUnraid kernel: br0: port 2(vnet0) entered blocking state

I have just disabled Promiscouous mode:

 

 

But that didnt solve the issue unfortinately. i send new d

 

Edited June 24, 2020 by Mirano

[PeteAsking]

Hmm I actually cant see the issue. At this point I would see what happens if you only have the vm's running then start only the unifi and pihole dockers and see if there is still a problem. Maybe one of the dockers you have is causing an issue.

[Mirano]

16 hours ago, PeteAsking said:

Hmm I actually cant see the issue. At this point I would see what happens if you only have the vm's running then start only the unifi and pihole dockers and see if there is still a problem. Maybe one of the dockers you have is causing an issue.

Hey, Thanks for answering. I tried what you mentioned, i stopped al my dockers except for my Nginx and unifi controller. When i start my VM the same thing happends unfortinately

[Mirano]

Okay, i've got the issue tracked down:

 

When my VM is using network bridge Br0 my unifi access points go unresponsive. When i use the only other one avaible: Virbr0 my access points will stay responsive, yet i have no internet on my VM. My network settings:

 

Anyone know what setting is wrong?

[Mirano]

My question is, how can i get Br0 to work that i can get a ip range in the 2.0/24 instead of virbr0 which sets me in 122.0/24 range. 

[itimpi]

I don't know if it is relevant but why do you have bonding enabled if their is only one NIC defined?

[Mirano]

29 minutes ago, itimpi said:

I don't know if it is relevant but why do you have bonding enabled if their is only one NIC defined?

Should i set "Enable Bonding" to no?

[PeteAsking]

Are the wg0 interfaces wireguard interfaces? I think some other people on the forums have had issues with wireguard and bonded nics but obviously I dont know if that is your issue, just mentioning this in case.

[itimpi]

4 hours ago, Mirano said:

Should i set "Enable Bonding" to no?

That is what I have on my system and I am not seeing the sort of problems you are reporting

[Mirano]

23 minutes ago, PeteAsking said:

Are the wg0 interfaces wireguard interfaces? I think some other people on the forums have had issues with wireguard and bonded nics but obviously I dont know if that is your issue, just mentioning this in case.

Yes they are, I will disable bonding either this evening or tommorow and see if that fixes things

[Mirano]

On 6/19/2020 at 3:19 PM, Mirano said:

Yes they are, I will disable bonding either this evening or tommorow and see if that fixes things

 

On 6/19/2020 at 2:56 PM, itimpi said:

That is what I have on my system and I am not seeing the sort of problems you are reporting

So i've tested by disabling bonding:

 

But no success, devices still go on adopting on the wifi portal and my pi-hole stops answering do DNS requests. i've included the new diag

hpunraid-diagnostics-20200624-0916.zip

[PeteAsking]

I really hope someone else will look at the logs because I cant see the issue and cant thibk what else to do at this point.