Support Forum

We are testing some new changes to how we interface with Chocolatey in our Habitat plugin that if shows real results then it will be pushed back to this plugin. We have identified areas where Chocolatey public repo limits are being hit causing lockouts from Chocolatey.org So by adding in code to stagger and loop limit is being tested to see if we can improve the delivery.

This is not an issue if using Chocolatey licensed repos, they do not force limits, any client location and pull as much as they need.

I don't think it's limits in the repository, I can't see it even trying to download the framework on some clients, it's just doing nothing.

The reason I say this is that we been seeing this allot lately as Chocolatey has modified their limit engine.


From Chocolatey's site

Installations/downloads of Chocolatey itself (chocolatey.nupkg) are rate limited at about 5 per minute per IP address - temporary ban expires after 1 hour.

All other packages are rate limited at about 20 per minute per IP address - temporary ban expires after 1 hour.

So if you have 30 agents at a location (IP address) then they all get choco at the same time to install, the first 5 will download the other 25 will get an error of exceeding limits and will fail out install. This is then repeated over and over again every few hours until all agents have chocolatey.

This was the old way, We are now monitoring a new piece of development that if successful will be backwards ported to Chocolatey For Automate. We just released this update out to the masses today for Habitat. If in a week or so we get good reviews and no major issues show up it will be pushed out to C4A.


https://docs.chocolatey.org/en-us/commu ... essive-use


They say to avoid this to host your own repos.

https://docs.chocolatey.org/en-us/guide ... ment-guide

Looks like installing Nexus Repository as a cache might be the answer. I'll look into it. Thanks.

And the same error is back again The location is not inactive according to the locations screen

Can I get some data around this ? Is this repeating in the logs or did it just happen a couple of time or is affecting a particular location or agent?

After another round of uninstalling, rebooting the Automate server (Followed by a complete loss of connectivity to the Control server which wasted a morning) and another install of the plugin, this error seems to have gone again, for now.

It was one site which had all locations enabled. It happened when I tried to roll out 7zip (The only app I've been trying to get rolled out to this client since day one). It's trying again but complaining yet again that the choco.exe is missing.

I've just written my own script to install the framework and tried it on a couple of machines and it seems to work OK. At least I can fix that problem I think.