We've got a few apps linked up a MongoDb-compatible aws docdb server (among other things), and this connection has been working fine for weeks. Nothing has changed with the db server configs, and other (non-retool) apps of ours are working fine communicating with mongo + existing creds.
This morning, all of our mongo queries in retool are now returning the following: { message: "pool destroyed" }. I've seen this error in parts of our (non-retool) codebase where we were trying to close the db connection instance before some asynchronous task finished, for instance, but obviously in retool those aspects of the db driver are abstracted away.
Of note, when in the resources section, clicking "test connection" returns a valid, success response.
I'm curious if this may be an issue with Retool's mongo adapter, or might it be something else more sneaky/internal to our own configs?
We are using the on prem version. I didn’t install the container myself & I’m not certain of the version. Is that exposed somewhere in the application settings (I can’t find it), or might I need to ssh into the box to find that out?
Sorry for the delay on this. We’ve had a few significant issues surface today and will be able to dig deeper into the cause here tomorrow. We will look into this as soon as we can and reach out if we have trouble reproducing the issue on that version
We ended up syncing on this issue separately, but for the community at large if this issue comes up again: restarting the container running retool resolved this issue. We are continuing to investigate how to prevent and automatically resolve this problem.
This is of course just true while running retool on-prem, let us know if you see the same issue on cloud!
We are using Retool in the cloud, with a MongoDB database on a third party server. Version is 2.64.32. Restarted the server for the MongoDB database, and now get "pool destroyed" in the app, which was doing great before that. "Test Connection" on the resource works fine. All other tools can connect to MongoDB fine. I would restart the container, but it's your container.
I was able to work around this by creating a "new" resource with identical properties -- just a new name. Swapped that in for the "old" resource and the problem went away. Still, very odd, so I thought someone might find this information useful..
I would also try re-saving the original resource- I’ve seen a few cases where after a database or SSH tunnel goes down, the Retool resource needs to be re-saved to reconnect. Adding/removing a space from the name will enable to save button, but submit no actual changes
