Casper's questions

Running Windows Server 2012 R2 I have noticed an annoying feature on RDP connections:

Windows keeps killing the Audio Service audiodg.exe if it is idle for more than 5 minutes.

The problem with this is that any new audio output will now suffer from a startup delay of 5-10 seconds, having to wait for audiodg.exe to spool up again before audio output can start.

I have seen audio delay problems being discussed with RDP connections numerous times on all Windows Server versions, but I have not yet seen anyone mention that this could be the cause of all these problems.

The spoolup time of audodg.exe will delay all audio on the server. It doesn't matter where the audio comes from. Interactive apps with audio feedback will be out of sync, and Youtube videos on Chrome will freeze until audiodg is running again.

On my server audiodg consumes 100% CPU at startup. I do not know what it is doing, but it does something at 100% CPU for about 5-10 seconds before normal audio resumes.

Once it is up and running all audio is instant. No delays or lag. Everything works fine as long as it keeps running.

The only way I found to fix this annoying "feature", was to create a repeating task that plays a few seconds of sound (silence) every 4 minutes to prevent Windows from killing audiodg.

It seems like a silly solution.

A couple questions comes to mind (in order of importance, I think):

1. How could I prevent audiodg from getting killed at all in the first place without resorting hacky solutions? Is there a registry setting for this somewhere? I have set the service to "manual" but that does not make any difference. Automatic/Manual..same problem regardless.

2. Why is audiodg startup so slow? I think this may also be some sort of bug or unintended feature.

Update

Looks like I found the answer to question 2 here:
The process audiodg.exe scans catroot and hogs IO

I have a Hetzner W2012 R2 Cloud VM, and I updated the VirtIO vioscsi driver on it. Now it only boots into recovery mode and I cannot access the C-drive even from command prompt.

Is there any way to save this VM or is it a goner?
I have no (recent) snapshot of it :(

I can mount the old previous VirtIO ISO, but how can I reinstall the old driver when I have no access to the C-drive?

Update 1:
Am able to get access to the C-drive with drvload vioscsci.inf.
Still trying to figure out how to reinstall the old drivers.

Update 2
I used dism /image:e:\ /add-driver /driver:vioscsi.inf to add the driver but it still won't boot.

Update 3
Startup repair gives error 0x490.

No clue what to do next or how to even troubleshoot why it is not booting.

Update 4

It appears all my system drivers have become unsigned. Disabling driver signature enforcement allowed the system to finally boot.

Do not know why this happened. It appears I will have to do a full reinstall, but at least I got back access to the VM now.

I have a problem where I need to manually set up a route on Ubuntu 18.04 during the init (ram) process. I cannot use any of the new networking tools, only Busybox is available.

The routing table should looks like this:

Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         172.31.1.1      0.0.0.0         UG    100    0        0 ens3
172.31.1.1      0.0.0.0         255.255.255.255 UH    100    0        0 ens3

The host is a VPS with an IP of 95.xx.xx.xx, so the gateway IP is not on the same network. I can't figure out what command to build the above table with. I get various SIOCADDDRT errors with everything I try.

Here's what I've tried so far:

> route add default gw 172.31.1.1
SIOCADDDRT: Network uncreachable
> route add default gw 172.31.1.1 via 0.0.0.0 
..invalid format
> route add 172.31.1.1 gw 0.0.0.0
SIOCADDDRT: Invalid argument
> route add 0.0.0.0 gw 172.31.1.1
SIOCADDDRT: Network uncreachable
> route add default gw 0.0.0.0
SIOCADDDRT: Invalid argument

Is there any way to make this work with Busybox? Or any other way to make it happen inside initram?