Self-Hosted Retool Server Locks During Install on EC2 VM

Self Hosted Retool
,
1

I'm installing Self-Hosted Retool using the standard scripts on an AWS EC2 VM running in our VPC. This had been straightforward till the latest releases. The VM hangs while creating the backend network. The VM becomes unaccessible and needs to be rebuilt. I see there are some errors related to Python 3.6. I'm wondering if this is causing the problem. Any suggestions welcome.

./local-trial
CONTAINER ID NAME CPU % MEM USAGE / LIMIT MEM % NET I/O BLOCK I/O PIDS
/snap/docker/2343/lib/python3.6/site-packages/paramiko/transport.py:33: CryptographyDeprecationWarning: Python 3.6 is no longer supported by the Python core team. Therefore, support for it is deprecated in cryptography and will be removed in a future release.
from cryptography.hazmat.backends import default_backend

INFO: updating Retool!
/snap/docker/2343/lib/python3.6/site-packages/paramiko/transport.py:33: CryptographyDeprecationWarning: Python 3.6 is no longer supported by the Python core team. Therefore, support for it is deprecated in cryptography and will be removed in a future release.
from cryptography.hazmat.backends import default_backend
Pulling db-connector ... done
Pulling db-ssh-connector ... done
Pulling postgres ... done
Pulling jobs-runner ... done
Pulling api ... done

INFO: running Retool! This can take up to 5 minutes
/snap/docker/2343/lib/python3.6/site-packages/paramiko/transport.py:33: CryptographyDeprecationWarning: Python 3.6 is no longer supported by the Python core team. Therefore, support for it is deprecated in cryptography and will be removed in a future release.
from cryptography.hazmat.backends import default_backend
Creating network "retool-onpremise_db-connector-network" with the default driver
Creating network "retool-onpremise_db-ssh-connector-network" with the default driver
Creating network "retool-onpremise_backend-network" with the default driver

2

Anytime I attempt to stop the docker container and restart it I lose network access to the VM and need to rebuild it. It always happens immediately after the following messages:

Creating network "retool-onpremise_db-connector-network" with the default driver
Creating network "retool-onpremise_db-ssh-connector-network" with the default driver
Creating network "retool-onpremise_backend-network" with the default driver

None of our AWS infrastructure engineers can figure out what the problem is. It is definitely related to how ReTool is creating the backend network. Would really appreciate some help or we'll have to drop our POC of the product.

3

Hi @nsteblay

Thanks for reaching out! Was this instance working previously? Are you upgrading Retool versions? Which version # are you upgrading to/from?

I don't see examples of other customers running into the Python warnings, but maybe this will help :crossed_fingers: https://askubuntu.com/questions/1439553/how-to-silence-or-fix-snap-docker-compose-cryptographydeprecationwarning-python

Are you able to send us all of the container logs (feel free to send as a DM)

4

Hi @nsteblay I wanted to check in here to see how things are going with your instance? If you're still blocked, keep me posted with these details! :pray: Thanks

5

Still having issues. I can run it on my personal account (AWS, GCP) but it hangs on our corporate AWS environment. I have a meeting with Wes and Alex later today. It seems to happen when the Retool is creating the on premise back end network. It's a black box to me as to what Retool is doing. I'm sure our corporation is doing some restrictive things with the VPC to ensure security. Maybe there is some conflict with that and how Retool is setting up the network. Not sure.