I just wonder if conda packages like https://repo.continuum.io
Can be able to mirror from Artifactory?
I tried mirroring it but it was unable to get cache..
Thanks in advance
Artifactory don't support it natively but you can use this plugin:
Artifactory Conda User Plugin.
You can install it in your Artifactory server (is a groovy file).
Also there are a request to include this funtionality in the future:
RTFACT-7872 Add support for Conda
A year later and things have changed (as they often do). So adding a new answer.
Artifactory has added support for Conda packages as announced in this blog. Here are the docs. Appears this requires Artifactory 6.3+.
Related
Newton user guide for openstack newton mentions the command show for a stack. Example: openstack stack show <stack_name>.
However, latest docs stein don't mention anything about this command. How can I show stack info in latest openstack/openstackclient release (stein at the moment of writing this question)?
make sure you have not outdated version of python-openstackclient
just install it with pip in venv and then install python-heatclient from pip as well
That same functionality is still there. Check out https://docs.openstack.org/python-heatclient/stein/cli/stack.html
If the command is missing in your environment, then you probably just need to install python-heatclient. It is not available by default with the unified openstack client.
What is the right procedure to deploy a debian package built for different distros into the same Jfrog debian artifactory repo?
Just uploading to the same path, but with different deb.distribution properties does not work, they all get uploaded to the same place and clobber the previous upload.
Including the distribution name into the package name is ugly, but would of course work. Is there a better way?
You simply post the different debian to different locations within the Jfrog artifactory repository. The trick is that the repository layout has nothing to do with the aptitude API, which retrieves debians regardless of their location according to the requested metadata (deb.distribution, deb.version etc).
The latest version of Nexus Repository Manager OSS supports a new artefact types: Docker, npm, RubyGems. But it appears to drop support for artefact formats supported in the previous version: Maven, P2, OBR, Yum. Why is that?
Maven should be there, in fact we setup default repos for it when it starts. P2, OBR, Yum have not been rebuilt for Nexus Repository Manager 3 as of yet. Yum will likely be the next up, P2 and then maybe OBR (jury is still out on if we have enough people using it to include it in the core product).
If you'd like to kickstart any of these, we would love that. A community member recently created an APT format for himself:
https://github.com/mpoindexter/nexus-repository-apt
You might be able to follow his example and create one for OBR for example.
I'm currently using:
mod 'nginx', :git => 'https://github.com/puphpet/puppet-nginx.git'
with a customised puphpet file that loads nginx but the version we get installed does support
–with-http_geoip_module.
All of the posts on installing geoip functionality, suggest that the only way to get geoip support is to recompile nginx.
Search for compiling packages with puppet suggests it's not the best way forward. Is there anywhere I can get the pre-packaged nginx module that supports geoip?
I'm using ubuntu 12.04
You'll need to find a repo that has this module already added. If you find it, you can simply use that.
This one may be what you need: https://launchpad.net/~nicolas-zin/+archive/ubuntu/nginx
Puppet workflows that rely on compiling your own packages are actually commonplace. What's not a good idea is for Puppet to do the actual compiling.
Build and maintain your packages with whatever tool chain fits your needs best. Maintain a package repository, and use Puppet to configure your nodes to use your repo. Puppet can then install your custom packages just like any other.
I'm trying to use prudentia to install postgresql. On the latest version of the develop branch, there is a task for posgresql, but that installs the wrong version for me.
I have found here that I need a special apt repository to get the latest (9.3) version. But I need some help installing it from prudentia.
I did see some example in code, but I couldn't find anything about adding repositories in the docs. How should I solve this?
To add an apt repository there are two ways.
1) The simplest one is to use the Ansible apt_repository module that requires the python-apt package installed on the target machine. This module accepts both deb and ppa repositories style.
A Prudentia task for installing PostgreSQL has been provided using this method and can be found here.
2) The second is to provide an apt source file. This approach doesn't have any dependency. The Prudentia Chrome bundled task uses this approach.