[ckan-discuss] Installation issues

David Read david.read at okfn.org
Thu Jun 24 11:19:54 BST 2010


Yes there are a fair few technologies used in the CKAN install and
you've highlighted that the readme doesn't explain things well enough.
You're not the first either - apologies. I've given the REAME a good
overhaul so do refresh your browser:

John mentioned a current problem with the install. The 'python-openid'
module (which CKAN depends) has had a broken package for the past
week. Here's the work around:

When pip fails with "ReadError: not a gzip file", you can install
python-openid from the source (i.e. not the package):
pip -E pyenv install http://github.com/openid/python-openid/tarball/2.2.5
Now restart the main install and it should complete fine. e.g.:
pip -E pyenv install -r pip-requirements-metastable.txt

I don't recommend installing python packages for CKAN in the main
python install (e.g. /usr/local/lib/python2.6) because you can easily
get caught out when pip installs a different version of the same
package in your virtual environment. I would delete them from the main
python install and persevere with installing them in your virtual

Do let us know how you get on, and you can always try and catch me on
IRC to resolve any quick issues. See: http://www.okfn.org/contact

All the best,

On 24 June 2010 10:07, John Bywater
<john.bywater at appropriatesoftware.net> wrote:
> Hi Will,
> William Waites wrote:me
>> It's somewhat of a swiss-army knife program that takes
>> care of running various common types of commands. You
>> use it to run a development web server "paster serve config.ini"
>> or to create a config file. It is also possible for a package
>> to "install" sub-commands -- see "paster --help" for a list.
>> It is basically a command-line wrapper so if you are writing
>> a program that does command-line stuff you don't have to
>> reinvent the wheel, you just implement a class that does
>> whatever the program is supposed to do.
> Since I'm already on my second cup of tea..... :-)
> Of course, the truth is that it is only the most abstract *idea* of the
> wheel which doesn't need reinventing, whereas actual wheels are reinvented
> all the time. A train wheel is almost an entirely different objects from a
> bicycle wheel, or a wheel on a push chair, or a steering wheel.
> We don't suffer from push chair manufacturers assembling push chairs with
> train wheels, or car steering systems with push chair wheels, under the
> slogan, "there is no need to reinvent the wheel!" :-)
> Therefore, we can confidently conclude, firstly, that it is only idealists
> who never see any reason to reinvent the wheel, pragmatists know they need
> to do this each time there is a new context where a sort of round thing on a
> stick would be good.
> Secondly, that common CKAN system admin tasks suffer from just this "train
> wheel on a pram" anti-pattern, when they could easily be presented as:
> ckan-install
> ckan-makeconfig
> ckan-runserver
> and so on. If it's good for the git goose [1], surely it's good for the ckan
> gander. :-)
> But that's all I've got to say on the matter! Hope you have/had a safe
> journey today.
> J.
> [1]
> $ git-
> Display all 130 possibilities? (y or n)
> git-add                  git-diff-files           git-lost-found
>  git-parse-remote         git-shortlog
> git-add--interactive     git-diff-index           git-ls-files
>  git-patch-id             git-show
> git-am                   git-diff-tree            git-ls-remote
>  git-peek-remote          git-show-branch
> git-annotate             git-fast-export          git-ls-tree    git-prune
>              git-show-index
> git-apply                git-fast-import          git-mailinfo
>  git-prune-packed         git-show-ref
> git-archive              git-fetch                git-mailsplit    git-pull
>                 git-sh-setup
> git-bisect               git-fetch-pack           git-merge    git-push
>             git-stash
> git-blame                git-fetch--tool          git-merge-base
>  git-quiltimport          git-status
> git-branch               git-filter-branch        git-merge-file
>  git-read-tree            git-stripspace
> git-bundle               git-fmt-merge-msg        git-merge-index
>  git-rebase               git-submodule
> git-cat-file             git-for-each-ref         git-merge-octopus
>  git-rebase--interactive  git-symbolic-ref
> git-check-attr           git-format-patch         git-merge-one-file
>  git-receive-pack         git-tag
> git-checkout             git-fsck                 git-merge-ours
>  git-reflog               git-tar-tree
> git-checkout-index       git-fsck-objects         git-merge-recursive
>  git-relink               git-unpack-file
> git-check-ref-format     git-gc                   git-merge-resolve
>  git-remote               git-unpack-objects
> git-cherry               git-get-tar-commit-id    git-merge-stupid
>  git-repack               git-update-index
> git-cherry-pick          git-grep                 git-merge-subtree
>  git-repo-config          git-update-ref
> git-clean                git-hash-object          git-mergetool
>  git-request-pull         git-update-server-info
> git-clone                git-http-fetch           git-merge-tree
>  git-rerere               git-upload-archive
> git-commit               git-http-push            git-mktag    git-reset
>            git-upload-pack
> git-commit-tree          git-imap-send            git-mktree    git-revert
>             git-var
> git-config               git-index-pack           git-mv    git-rev-list
>         git-verify-pack
> git-count-objects        git-init                 git-name-rev
>  git-rev-parse            git-verify-tag
> git-daemon               git-init-db              git-pack-objects    git-rm
>                   git-web--browse
> git-describe             git-instaweb             git-pack-redundant
>  git-send-pack            git-whatchanged
> git-diff                 git-log                  git-pack-refs    git-shell
>                git-write-tree
>> HTH,
>> -w
> _______________________________________________
> ckan-discuss mailing list
> ckan-discuss at lists.okfn.org
> http://lists.okfn.org/mailman/listinfo/ckan-discuss

More information about the ckan-discuss mailing list