UPSTREAM_NOT_FOUND [519]

Hi,

There are two things that you can do here, please let me know if either of those solutions work.

1. Verify that the RDP port (TCP 3389) is opened for your instance's firewall settings: https://lightsail.aka.corp.amazon.com/ls/docs/en_us/articles/understanding-firewall-and-port-mappings-in-amazon-lightsail

2. If that does not allow you to access the instance, attempt to create a snapshot of your instance, and then using the snapshot, create a new instance from it. Attempt to connect to the new instance:

https://lightsail.aka.corp.amazon.com/ls/docs/en_us/articles/prepare-windows-based-instance-and-create-snapshot
https://lightsail.aka.corp.amazon.com/ls/docs/en_us/articles/lightsail-how-to-create-instance-from-snapshot

Leslie

I am suffering from the same problem.
I did your solutions, 1, 2 all, but not work.
See this https://forums.aws.amazon.com/thread.jspa?messageID=920065󠨁
djenie found the tip to avoid this problem.
Yes! I clicked Firewall options include Port 22.
I didn't changed SSH port 22 at all, i did just clicked and saved.

So, I think this is the obvious technical problem.
If someone clicked Network > Firewall > Some Clicked more but nothing changed > Problem occur !!

I cannot do anything for my task.
Your engineer MUST fix it ASAP.

Edited by: HTMALL on Mar 1, 2020 5:41 AM

Are you 100% sure the sshd server is running on your instance?

I just ran a test on a completely functional Amazon Linux instance...

1. Logged in to the instance using ssh from a desktop client
2. Stopped the sshd server using 'service sshd stop'
3. Tried to log in to the instance from the Lightsail control panel.
4. Got the UPSTREAM_NOT_FOUND [519] error.
5. Restarted the sshd server from the desktop client session
6. Successfully logged in to the instance from the lightsail control panel

I totally agree.

I too experience this problem. I use AWS Support and they fix in a few minutes.

My environment is CentOS 7. New instances created from a snapshot inaccessible by SSH with "upstream_not_found 519" error.

The problem was because cloud-init was not installed on the original server.

On original server SSH terminal:
$ cloud-init --version

If it responds "cloud-init: command not found" then:
$ sudo yum install -y cloud-init

check installed
$ cloud-init --version
#response
/bin/cloud-init 18.5

Enable the service
$ sudo systemctl enable cloud-init.service

Take a backup and remove /var/lib/cloud files
$ sudo cp -a /var/lib/cloud /var/lib/cloud-backup
$ sudo rm -rf /var/lib/cloud

We then took a snapshot, create an instance from it and were able to access the new instance successfully.

It is a little worrying that AWS users may be relying on snapshots that prove to be inaccessible.

Edited by: socialblend on Aug 5, 2020 11:38 PM

Edited by: socialblend on Aug 5, 2020 11:40 PM