Resource Manager - Temporary Failure in Name Resolution

Right so we thought we better give the Resource Manager another try.

Screenshot from 2024-06-03 22-56-40

Importing our Digital Ocean Database Cluster actually worked but importing the Droplets themselves the above is what we see. We know our domains all resolve. We use SSH every day for many other things outside of Wappler. We also connect to Digital Ocean with SSH in other ways without issue. We went as far as to entirely reset everything including turning off any firewalls, disabling Cloudflare and actually bypassing its DNS servers and returning all the way back to basics just to discount any of that being raised as an issue. We generated new SSH keys which work for everything else outside of Wappler just fine, and they are also deployed to Digital Ocean and all working fine. We can ping our hosts, they are included in our hosts file, we use Cloudflare for DNS with Google DNS as a fallback so all good there, we can lookup our host and tracert. The only thing that will not resolve is Wappler itself.

We have literally reverted to the absolute basics and still no success with Resource Manager and are now posting for advice?

This is a Wappler bug, it's trying to resolve "undefined"

1 Like

My thoughts too @Apple. We literally dragged the hosting environment back to the Stone Age to disregard our side as the issue. Wappler can retrieve the database cluster via SHH and the API token, and can connect via SSH in the Wappler terminal using SSH so not a problem there either (keys all work). Everything else aside from Wappler Resource Manager resolves just fine. Thanks for your response and insight.

Wappler probably weren’t able to fetch the droplet ip. Just click on it and choose refresh from the context menu.

Then check the properties afterwards that the ip is shown.

Hi @George sorry for the delayed reply been really busy, which is good and bad...

Unfortunately that did not work and resulted in the same error and also duplicates the dialogue.

Hi @Cheese

Well I checked again and could import all DO servers without any problem.

Did you select the server first on the dropdown?

I see there is small UI bug when you click on the server refresh list, so will fix it, but this shouldn't break the import.

So maybe try again without too much refreshing just pick on from the list choose the right key and import?

Hi George, yes I did select it from the dropdown.

I refreshed multiple times to highlight the repeating issue with the dialogue. Still if I refresh only once I still am unable to import my Droplet. Disabled everything that could complicate matters going even as far as to restore all DNS to defaults circumnavigating Cloudflare. Disabled firewalls (UFW and DO). Still with no success. However like I mention the Database Cluster imports just fine, and my SSH keys all work when connecting to DO through the Wappler SSH terminal so no problem there either, as well as the other keys we use for other purposes which all connect just fine also. At a total loss as what else to try..? This is not urgent at the moment but no doubt in the future when the Resource Manager is more integrated it will become a real issue.

Well not sure what kind of droplet you have created and why you can’t import it. Maybe indeed something with the network and ip definitions, hope is not behind firewall or so in private VPC network?

You can send me a debug log in a PM

Maybe try to import other droplet?

You can try to just create a new server with the resource manager se

It is a Droplet Wappler created prior to the Resource Manager.

No we stripped it back to the bare minimum. Wappler can connect to it fine, and Deploy to it, and SSH in to it in every other regard aside from the Resource Manager.

We have tried that with another Droplet created in Wappler. Same thing. Undefined error.

Will give that a try although not sure how it will resolve the issue we are experiencing?

Will do George. Thank you.

Ooh is it a docker machine droplet?

Yes, Docker Droplet created by Wappler using the old Targets method before Resource Manager was conceived. Have sent the log to you George. Contains the same undefined error and being undefined a bit difficult to debug, sent none the less though.

:slight_smile:

I created a new server/Droplet using the same SSH key that is used for the other Droplets and all connected fine. However I still can not connect to my other Droplets using the Resource Manager. I've also noticed that in the old Target method that when I click Manage and the Droplets are displayed they all contain 'error' for the 'State' whereby up until recent updates there was no error (nothing has changed aside from upgrading Wappler a few times).

I can stop and start the Droplets fine from the Manage dialogue.

As you can see, so the 'State' should reflect this... And not display error.

The only way we can figure out a work-around is to create new Droplets with the Resource Manager. Configure them to mirror our current Droplets that work fine, and delete them (the currently working Droplets). Forward our DNS to the new IPs for the Droplets, update Cloudflare to reflect these changes, pick the newly created Droplets for the Projects, and Deploy to them.

Surely this is not how we should be migrating to the Resource Manager? There is quite obviously something wrong connecting to older Wappler created Droplets (Docker). We have gone through our side with a fine tooth comb and have no issues whatsoever.

A breaking change has occurred somewhere along the way. There should be no need for us to create new Droplets just to use Resource Manager...

Hi @Cheese,

I have also experienced issues when servers or droplets (any remote service provider) created via the now deprecated Sockets method (Manage Docker Machines) couldn't be migrated to the Resources Manager.

Even though the servers that were being imported in the Resources Manager accessed the same SSH key that was used to create these servers in the Sockets Method, it didn't work as expected.

I had to create new servers / recreate / import servers specifically in the Resources Manager and delete the previously created servers in the Docker Machines.

Hi @guptast.
We have realised this is definitely broken within Wappler and have thought about doing the same as you have done but to be honest we should not have to do this. It truly is a last resort. And am sure we are not alone in this situation as you have also experienced the same issues and am sure others will also experience the same. I'd hope the Team can rollback whatever has caused this problem and allow us to adopt the Resource Manager without having to recreate our Droplets. It should work but does not and really needs to be resolved with a fix as soon as possible. Especially with plans to remodel the way Targets operate and impending expansion of Resource Manager as the only way to manage deployment.

This suggests that George may have an idea of what the issue is. We have faith he will provide a solution. Fingers crossed!

Thank you for posting your experience I do truly appreciate it and it matches ours exactly. Nice to not be alone on this.

Found the problem!

Was indeed a special case when importing the old docker machines.

It will be fixed in the next update and we also have improved adding a new ssh key during the import so it is auto authorized

Do note that once when importing and if you run a system check firewall options will be set to access the server only through ssh

2 Likes

Thank you very much George!

:smiling_face_with_three_hearts:

All seems to be working fine for us @George. We can now import our DO Droplets in to the Resource Manager. Only have a slight issue where one is not showing the Traefik Service although it is installed. The others show the Services just fine. All in all resolved for us! Many thanks for the fine work and another great update!

:slight_smile:

1 Like