VMs Lose Network Connectivity After Upgrade (Ubuntu 22/24 LTS Nodes)

Symptoms 

Following a SolusVM 2 upgrade, VMs experience a network outage: network briefly appears functional immediately after the upgrade, then fails once a VM is started, restarted, or created.

Cause

The SolusVM 2 upgrade process restarts the OVS (Open vSwitch) service, which clears all existing flow rules from the host. This leaves the host briefly functional but unprotected (no ARP whitelist rules in place).

As soon as any VM is started, restarted, or created, that VM applies ARP rules operating in whitelist mode. Because the OVS restart removed all existing whitelist rules, traffic for VMs that have not yet had their rules reapplied is treated as non-whitelisted and is dropped - causing the network outage.

This is a known issue and has been logged with SolusVM R&D as a point of improvement, tracked under SVM2-6538. Monitor the release notes (https://docs.solusvm.com/en/release-notes/) of upcoming SolusVM 2 updates for reference to SVM2-6538, which will indicate that a permanent fix has been released.

Resolution

1. Log in to the SolusVM UI. 
2. Navigate to the Networking Rules page. 
3. Click Save (no changes to the rules are required — the Save action alone triggers reapplication of the correct settings). 
4. Networking on affected VPS instances resolves within 1–2 minutes.
   
   
    

Note: Restarting OVS is never an appropriate remediation step. Whenever the ARP filter is enabled, an OVS restart will reliably reproduce this issue, since new VMs apply ARP rules that drop packets for every other VM lacking the corresponding whitelist flows.

In summary, restarting OVS:

• Drops datapath flows
• Breaks ARP filtering logic
• Causes packet drops across VMs
• Requires re-learning/re-population of flows