Closed jaypipes closed 11 years ago
With host resolution failing (#19), does this node described have an updated resolv.conf
file, and can it properly resolve rubygems.org
? If so, is there actually a proxy between you and the public net?
Chris, that's exactly what it was. I removed the http_proxy recipe, upload_roles'd, ssh'd to the ops node, added the 10.0.3.1 nameserver to resolv.conf and that got everything working. Closing this one out as it is a duplicate of Issue #19.
Thanks! -jay
Hi again, Chris,
I'm working on adding Vagabond support for a stripped-down Chef repo that is used as an example Chef repo for installing OpenStack with Chef [1].
I'm running into an issue that is making me wonder whether Vagabond needs to insert/inject some HTTP proxy settings in the various nodes in order to have the nodes be able to reach external addresses.
When I run provision on the ops node, I'm getting this:
Note that the ops node only includes a single role:
And the os-ops-database role is really just installing mysql-server in the node's runlist, and nothing more. The Berksfile for the Chef repo looks like so:
When researching the error that appears in the chef-client output during provision, I found this:
http://tickets.opscode.com/browse/CHEF-3912
which seems to indicate that the Illformed requirement error is due to a bad HTTP proxy setting. The suggested fix is to put the following:
into a recipe. So, I did that. I put the above into a cookbook recipe called testing::default and uploaded the cookbook into my Berkshelf and re-uploaded the cookbook. In that testing cookbook's attributes/default.rb file, I set the default['http_proxy'] setting to '10.0.3.132' (the IP of the local chef server). I added the cookbook to my base role's runlist and uploaded the roles to the local Chef server, I then reprovisioned the ops node but a) I did not see any output that indicated the Chef::Log.info line above was run, and I continued to get the same chef-client error about Illformed requirements.
Any ideas on what might be happening? You think it's related to the http_proxy stuff, or do you think that isn't really the source of the issue?
All the best, -jay
[1] https://github.com/stackforge/openstack-chef-repo