RVR for share network #10491

Activity

1. tuanhoangth1603 commented on Mar 3, 2025

   

   tuanhoangth1603

   on Mar 3, 2025

   Author

   More actions

   Also, I recently upgraded from 4.19 to 4.20. All existing networks (including isolated networks) must have a redundant VR; otherwise, I cannot create any VMs for these networks, and the management log will show the following error:
   Unable to contact resource. com.cloud.exception.ResourceUnavailableException: Resource [DataCenter:1] is unreachable: Can't find all necessary running routers!
2. weizhouapache commented on Mar 3, 2025

   

   weizhouapache

   on Mar 3, 2025

   Member

   More actions

   @tuanhoangth1603
   the VR for shared network only provides the service for Dhcp/Dns/Userdata
   In my opinion, it is unnecessary to make it redundant.

   anyway, the error message should be more clear and correct.
3. tuanhoangth1603 commented on Mar 3, 2025

   

   tuanhoangth1603

   on Mar 3, 2025

   Author

   More actions

   Thanks @weizhouapache, for your reply.
   But there are some issues that I don't fully understand:

   1. Why do I need to convert the VR to an RVR for an isolated network after upgraded from 19 > 20, even though the VR is still in a running state? After doing so, I can successfully create a new VM.
   2. Similarly, for the VR of a shared network, I am unable to create a new VM, even though the network's VR is also running. So then I think it's need to make redundant
4. weizhouapache commented on Mar 3, 2025

   

   weizhouapache

   on Mar 3, 2025

   Member

   More actions

   Thanks @weizhouapache, for your reply. But there are some issues that I don't fully understand:

   1. Why do I need to convert the VR to an RVR for an isolated network after upgraded from 19 > 20, even though the VR is still in a running state? After doing so, I can successfully create a new VM.
   2. Similarly, for the VR of a shared network, I am unable to create a new VM, even though the network's VR is also running. So then I think it's need to make redundant

   The version of scripts in VR should be consistent with the ACS version of the host and management server.
   So you need to upgrade the VR, you can patch the VR, but I suggest to restart the network/vpc with cleanup, which will

   • create new VR
   • destory old VR

   for shared network, no downtime as the VR is not VM gateway.
   for isolated network without RVR, there might be few seconds downtime.

   No need to make the network/vpc redundant
5. tuanhoangth1603 commented on Mar 3, 2025

   

   tuanhoangth1603

   on Mar 3, 2025

   Author

   More actions

   I restarted the shared network with cleanup, but it showed the following error:
   Resource [DataCenter:1] is unreachable: Can't find all necessary running routers!
   This issue occurs in both cmk and the UI.
6. tuanhoangth1603 commented on Mar 3, 2025

   

   tuanhoangth1603

   on Mar 3, 2025

   Last edited by tuanhoangth1603

   Author

   More actions

   I have found the cause.

   When I make the VR to an RVR with cmk, even though it failed to become redundant, but it still changed the value of the redundant column in the networks table to 1. As a result, the network was unable to restart
   In my opinion, if a shared network does not require an RVR, we should provide a clearer notification when someone attempts to convert a VR to an RVR and ensure that the redundant column value remains unchanged to 1.
7. tuanhoangth1603 commented on Mar 3, 2025

   

   tuanhoangth1603

   on Mar 3, 2025

   Author

   More actions

   Thanks @weizhouapache, for your reply. But there are some issues that I don't fully understand:

   1. Why do I need to convert the VR to an RVR for an isolated network after upgraded from 19 > 20, even though the VR is still in a running state? After doing so, I can successfully create a new VM.
   2. Similarly, for the VR of a shared network, I am unable to create a new VM, even though the network's VR is also running. So then I think it's need to make redundant

   The version of scripts in VR should be consistent with the ACS version of the host and management server. So you need to upgrade the VR, you can patch the VR, but I suggest to restart the network/vpc with cleanup, which will

   • create new VR
   • destory old VR

   for shared network, no downtime as the VR is not VM gateway. for isolated network without RVR, there might be few seconds downtime.

   No need to make the network/vpc redundant

   Also, restarting the network helped fix the issue mentioned in this topic.
   Thanks @weizhouapache

   👍React with 👍1Reacted by Wei Zhou
8. tuanhoangth1603

   closed this as completedon Mar 3, 2025
9. weizhouapache

   added this to the 4.19.3 milestone on Mar 21, 2025