Installation on an AWS EC2 server (T2.Micro) at AWS China hangs

[andreas]

For a separate project I wanted to set up an instance with Cloudron for the Swiss parent company for the organization in Shanghai China at AWS in China.

Currently, the execution of the script is stuck. Maybe someone can help here.

[nebulon]

It seems it is just trying to update/installed required Ubuntu packages from the Ubuntu mirror. Maybe that has some temporary issues in that region. It is safe to rerun the script again.

Also we do have a pre-built AWS marketplace image. So that can be used as well to quickly get a Cloudron up and running on EC2.

[andreas]

It seems to be true. Meanwhile, it continues in slow motion.
The pre-installed version cannot be used at AWS China because not all Amazon services are available. Or do I have wrong information here?

[nebulon]

Actually I have never tried using AWS China. So not even sure if the marketplace image is available there.

Generally the t2.micro might not be the best fit, since it is based on those cpu credits, which might get depleted quickly during installation and then the whole system usually crawls in EC2

[andreas]

Okay, good to know. The decision came from our colleagues on site in China. What would you recommend for an EC2 instance?

[girish]

Very much depends on what apps you want to run. You can use the newer t3 instances, something like t3 medium (or the equivalent t2).

[andreas]

The installation finished after a few hours after starting the script. It seems to be successfully but I can't reach the setup panel of Cloudron after the installation process. I tested the public domain name of the EC2 server and also the public IP but I can't reach it. Also, to ping the address fails in the command line.

Chrome shows the error message "ERR_CONNECTION_TIMED_OUT"

Maybe it's a wrong configuration in AWS or a problem to install it at AWS China. Do you have an idea what kind of problem it could be?

After I know the right way, I will switch to the recommended server size.

[girish]

@andreas Are port 80 and 443 open in the security group? Have you also allocated an elastic IP to the instance? After that you have to visit, https://<elastic ip>

[andreas]

It set up the following inbound rule:

and outbound was the standard setting:

Now I also assigned as mentioned the elastic IP but the problem still persists. Also, the ping command gives the same reply with the lost packages.

Have the two steps to be done before the installation with the provided Cloudron script?

[girish]

For ping, you have to enable ICMP traffic (it's not TCP). See https://docs.amazonaws.cn/en_us/AWSEC2/latest/UserGuide/TroubleshootingInstancesConnecting.html#troubleshoot-instance-ping

Now, since you are able SSH, clearly port 22 is working. It's just that for some reason https is not working. Maybe there is some problem which is EC2 specific. Are you able to run cloudron-support --enable-ssh on the server? After that, can you send me the public IP to support@cloudron.io and I can take a look as to why it's not working. (I will only have access to your server and not any of your other aws stuff, of course).

[andreas]

I terminated the instance to test the whole thing again from the beginning with the recommended settings. This means that the installation will probably take several hours.

Unfortunately the message has overlapped with the termination. I will give you feedback if everything fits or if I have sent the mail to your support address.

With the additional rule the pinging worked, thanks for the great tips and the incredible fast support.

[girish]

Do you know why the install takes several hours? Cloudron should only take 10 mins to install .

[andreas]

I think the updates of the packages are the problem that slows all down. I can send the logs after installation if it would help?

[andreas]

Email is sent. Same result as before.

[robi]

can you validate by doing an apt update && upgrade before the install?

[andreas]

@robi I will try that after I have no more options with the Cloudron support team.

[andreas]

I suspect the problem is this:

https://stackoverflow.com/questions/29422709/cant-connect-to-ec2-instance-over-http-from-outside-aws

[scooke]

@andreas True, it doesn't matter where the host company, or head office is located. Anything in China, for China, is going to be under lots of scrutiny.

[andreas]

@scooke said in Installation on an AWS EC2 server (T2.Micro) at AWS China hangs:

@andreas True, it doesn't matter where the host company, or head office is located. Anything in China, for China, is going to be under lots of scrutiny.

Unfortunately true, we must accept this learning curve.