I am not sure where the problem is, but more often than not my attempts to create hosts on DO through the rancher interface fail. The latest error is:
runtime/cgo: pthread_create failed: Resource temporarily unavailable
The failures are intermittent. Eventually I get lucky and it goes through. The problem is there is no “retry” here, other than going back to the hosts page, deleting the failed host, and doing it again, which is, well, tedious.
I wouldn’t mind the intermittent failures if I just had to wait for the issues to work themselves out.
For the immediate issue, is this a digital ocean issue, or rancher?
I’m running Rancher 1.6.0
Still failing, but new error:
fork/exec /usr/bin/docker-machine: cannot allocate memory
Exciting!
You can clone a host to pre-fill in all the previous info.
Both of those are you’re out of memory. How big is the host running the server container and what else is running on it?
I’m trying to create a 4GB host. My rancher manager is a 1GB host.
The server container (“manger host”) is running out of memory, so you either need a bigger instance for it or to run less on it. At 1gb it basically needs to be dedicated to rancher/server as that is the minimum requirement, particularly with an internal database (the default).
It is a dedicated droplet for the rancher manager, but you are correct its not enough memory.
I increased the RAM to 2GB and everything is now working well.
It would be nice if the messages were more clear, or rancher would inspect its host to make sure it’s big enough rather than running into problems down the road like this.
Thanks for the help!