Yesterday I noticed we still have some tox jobs running on the permanent slave. An example is `tox-flake8`, we also have jobs suffixed with `-trusty` which was to use python 3.4.
This task is about:
[ ] identify all tox jobs still running on permanent slaves
[ ] update the JJB config to use Nodepool (node: ci-jessie-wikimedia and parameters: zeromq)
[ ] update jobs
[ ] verify it works on the various repo
[ ] celebrate
----
Status as of 8bf0ee342f1f21757d090313fd2f3db5a498a975 Feb 11 22:50 UTC
```
$ grep -l tox output/*|xargs grep -l assignedNode.*Ubuntu|cut -d/ -f2
mw-tools-releng-tox-doc-publish
mw-tools-scap-tox-doc-publish
pywikibot-core-tox-doc-publish
pywikibot-core-tox-nose
pywikibot-core-tox-nose34
pywikibot-tests-beta-cluster
pywikibot-wikibase-tox-flake8-py3
pywikibot-wikibase-tox-nose
pywikibot-wikibase-tox-nose34
tox-doc
tox-flake8
tox-flake8-bin
tox-py27
tox-py34
```
The publish jobs will be annoying they push to integration-publisher then trigger a job that grab the material from gallium. https://www.mediawiki.org/wiki/Continuous_integration/Documentation_generation Can be migrated later.