Code Monkey home page Code Monkey logo

Comments (8)

nwoodmsft avatar nwoodmsft commented on August 20, 2024

Hi Ben, I think I saw this one time, but haven't been able to repro since....it seems to be related to clean-up/undo and the ovsdb conf file ending up with a corrupted copy. First to confirm, had you previously run SDNExpressUndo.ps1 against this system?

from sdn.

benpiper avatar benpiper commented on August 20, 2024

Yes, several times.

from sdn.

nwoodmsft avatar nwoodmsft commented on August 20, 2024

Thanks, it's probably the corruption issue which I mentioned or a case of the .conf files in your SDNExpress\AgentConf being mismatched with the OS. Either way, the 'undo' script shouldn't really be cleaning up by replacing those files (that will be resolved shortly).

Aside from starting with a fresh host, the only workaround I am aware of is to make sure NcHostAgent service really is stopped and then navigate to C:\ProgramData\Microsoft\Windows\NcHostAgent and replace the Firewall.conf and ServiceInsertion.conf files with fresh copies from GitHub (https://github.com/Microsoft/SDN/tree/master/SDNExpress/AgentConf). You may already have these files locally from when you downloaded SDNExpress (they are in the AgentConf) directory) but given that you are seeing corruption/mismatch I think it would be a good idea to re-download those files from the SDN Master branch (i'm assuming that you are working on Server 2016 RTM and not earlier preview builds).

from sdn.

benpiper avatar benpiper commented on August 20, 2024

Thanks, I'll try that. I've rebuilt the host since then and am actually having a different issue with the script not provisioning the infrastructure VMs. I get a "Windows could not complete the installation" error. It only happens when using the script. Just booting the sysprepped master VHDX works. Is there something special the VHDX needs for the script to provision the VMs?

from sdn.

nwoodmsft avatar nwoodmsft commented on August 20, 2024

Hi Ben, Nothing special, no. The script creates a unattend file and copies it into the root of the VHDX so that on first boot the VM will come up with the right domain membership and management IP plumbed etc. I've not seen any reports of problems with infra VMs booting, so I would suspect an issue with the sysprep'd VHDX. When you syspre'd it, did you select both /oobe and /generalize ?

from sdn.

benpiper avatar benpiper commented on August 20, 2024

I resolved the issue with the VHDX. But after rebuilding the host and copying the Firewall.conf and ServiceInsertion.conf files, I'm still getting the Failed to start service 'NC Host Agent (NCHostAgent)' error. The event log shows the same A device attached to the system is not functioning error even though it doesn't specify which device.

from sdn.

nwoodmsft avatar nwoodmsft commented on August 20, 2024

Hi Ben, Thanks for trying that. Given that this is consistent I'm wondering if the .conf files are somehow mismatched to your build. Can you confirm which version of Windows Server 2016 you are running on the hosts? The .conf files from the Master fork of this GitHub repro is intended to be aligned with Windows Server 2016 RTM or later. If your build is RTM and your using the latest .conf files from Master, then the issue should not reproduce (but if the build is TP5 or the .conf files were from an earlier version of SDNExpress then you would see this issue).

Also, replacing the .conf files is not a required step...it's just intended to be a last resort clean-up step that the SDNExpressUndo.ps1 script performs to ensure that no tenant policies (e.g. firewall rules) are left behind when you undo the SDN fabric. If this step is the only thing causing problems for your environment, then it would be safe to just comment out the RestoreOVSDBConfFiles script block within SDNExpressUndo.ps1 to unblock yourself and ensure that this conf files don't get copied in future.

To fix your current environment, you can either copy the .conf files from another host which is healthy (i.e. nchostagent is able to start on the host) or reimage the host so that it starts with a clean/fresh .conf

from sdn.

benpiper avatar benpiper commented on August 20, 2024

I started over fresh with a clean install of RTM and the problem with the NC host agent is resolved. It seems that running the SDNExpress script unsuccessfully breaks something. If the first run is successful, then the NC host agent starts.

from sdn.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.