Affected OS: Ubuntu 18.04
Client Version: 9.1r10
I may have discovered a bug which causes name resolution failure while connected Pulse Secure VPN.
The issue occurs when there is some brief interruption to network connectivity between the client and server. When this was reported by our user, I suspect they were having some intermittent ISP issue.
After the network interruption (and connectivity is restored), the /etc/resolv.conf file reverts back to default values and the client can no longer resolve hostnames for hosts in our internal network.
The tunnel is still open and connectivity is still working using IP address.
To replicate this:
- Establish VPN connection using Pulse Secure client on Linux
- Connect via SSH to a server over the VPN
- Disrupt network connectivity briefly (even a few seconds), then restore. In my testing I did this by running Ubuntu in a VM and disconnecting the host-machine from the internet briefly.
- Contents of /etc/resolv.conf should be reverted back to default values and can no longer resolve hostnames of internal servers over VPN.
Strangely, this issue seemingly only occurs if there is an active SSH connection (other TCP/IP connections may have the same effect but I haven't tested) to an internal server over VPN during the process.
If there is no SSH connection, the /etc/resolv.conf should be restored with the correct DNS information.
