CPU pegged, repeatedly outputting "INFO:haproxy:Websocket close" about dockercloud-haproxy

[evanp]

As a side note, my Docker Hub account keeps getting locked for "Too many login attempts." And I think this might be related.

[fermayo]

@evanp Did the haproxy container show the 95% CPU load? What was the time that passed between INFO:haproxy:Websocket close?

[evanp]

Yes, the haproxy container. I used "docker stats" to get the figure.

It was scrolling as fast as it could in the console (through an SSH connection, so probably pretty fast).

HERE'S MY THEORY: I think that the container is using some kind of inherited, transient authentication credentials to log into the docker cloud API.

I think that these credentials are expiring at some point after redeployment. However, I think that the container is ignoring that the credentials have expired and trying to re-connect over and over and over and over, which is why the "Websocket close" message keeps looping.

I'm going to try deploying with explicit Docker Cloud credentials using the DOCKERCLOUD_USER and DOCKERCLOUD_APIKEY environment variables that the python-dockercloud module uses.

At the same time, I'll look into doing something better when credentials expire, and send a PR.

[evanp]

Looking at the code, it looks like the library will always use the DOCKERCLOUD_AUTH environment variable passed by the Docker Cloud infrastructure for auth, if it's provided, and ignore the DOCKERCLOUD_USER and DOCKERCLOUD_APIKEY.

Oh, and: I'd love to be testing this, but my account is still locked out of
I'm not sure I can fix this; it seems to be a problem with the lifetime of the DOCKERCLOUD_AUTH.

[fermayo]

The DOCKERCLOUD_AUTH token only expires when the container is destroyed in Docker Cloud, not before that. Do you have that container running for troubleshooting?

[svenmueller]

We are facing the same issue #33

[tifayuki]

This is related to API throttling, and we are keep optimizing it on the server side.
Not really an error of this haproxy image, and I am going to close this issue at the moment.
If you still have such a problem, please contact support@docker