I have two separate Linux VMs running Ubuntu 12.04+ deployed in West US. I used the official Ubuntu Gallery image for 12.04, not my own image. SSH was properly configured and successfully working fine (authentication: username / password) for several months through November 9, 2014.
Since November 10, 2014, I have no longer been able to SSH to both VMs.
I have tried all the troubleshooting steps listed in this thread: https://social.msdn.microsoft.com/forums/azure/en-US/538a8f18-7c1f-4d6e-b81c-70c00e25c93d/troubleshooting-endpoint-connectivity-rdpsshhttp-etc-failures?forum=WAVirtualMachinesforWindows.
That is, I have tried all of the following: restarting the VMs, deleting / re-creating the SSH endpoint, re-creating the SSH endpoint on the default port, resizing VMs to change VM host (verified Internal IP changes), deleting / re-creating the VMs.
I have tried connecting from multiple networks. I have tried from multiple computers. I have tried on multiple days.
I am using NMAP to test connectivity (port shows as 'filtered'). I am using Putty as the SSH client.
One major thing that I did to both VMs early in the day on Sunday, November 9th: I installed all available updates (almost all security updates) by manually executing the command 'sudo unattended-upgrades'. I checked all the relevant install logs - no errors were reported. I forced a reboot from within the VM. Following the reboot, I was still able to continue to use SSH successfully for at least 12+ hours -- this was for true for both VMs.
I did not try SSH again until very late in the evening on November 10th, 24+ hours later, and now the SSH port seems to be staying closed no matter what I do.
I do not see anything in any of the Operational Logs in Azure between November 9th and today (all the endpoint / re-sizing shows up).
It should also be pointed out that both VMs had web applications installed on them, and the HTTP / HTTPS ports continue to work just fine. Only the SSH port is not working on both.
It seems like the issue is persisting on disk given all the troubleshooting steps that have been tried in vain so far. The fact that this is happening on 2 VMs makes a random coincidence seem extremely unlikely. If this is somehow related to the applied updates, it seems very peculiar that the issue would have such a delayed onset -- I would have expected update issues to manifest immediately or at least within 12 hours of the reboot.
Additional troubleshooting suggestions welcome.