Hello all!
I had a few friends coming over this Sunday for a mini Hackathon in my kitchen. The goal was mainly to introduce them to what we do around Wikipedia, have fun, exchange ideas and maybe do actual work on a few tickets. Quick summary of what we did:
* Fix a few puppet unit tests [1][2] * Try to fix sending logs to logstash from elasticsearch [3][4]. We could not reproduce the issue with security manager with a local installation of elasticsearch and I did not prepare access to deployment-prep where we could have validated the issue. * Added a few tests to our elasticsearch admin tool (estool) [5] * Refactoring of Cassandra monitoring [6][7] (thanks to Luca for the fast code review!). This one prompted a great conversation. It was fairly unclear to us how to best integrate monitoring in a DRY way, while keeping the core of the Cassandra module free from external dependencies. We are all used to have a technical module (puppet-cassandra) that only manage core Cassandra and a higher level module (puppet-wmf_cassandra) that configure Cassandra in the context of WMF, with the all peripheral fonctions (monitoring, firewall, logging, ...). This higher level abstraction kind of exists in the roles, but multiple roles might use Cassandra, and might lead to some duplication. The change [6] is probably more the start of a conversation than something that can be merged as is. * We tried to have a go at adding unit tests to Kartotherian [8], but ran into an issue with mapnik and were not able to build the project. None of us had any experience with Node development or Kartotherian. We moved to other stuff.
The day was fun! Thanks to all who joined! Special +1 to Mara who put a lot of energy in learning Python for the first time (and succeeded!).
I'll try to followup on the changed opened during that day, make sure that those contributions, however small, are not forgotten.
We will be back!
MrG
[1] https://gerrit.wikimedia.org/r/#/c/295127/ [2] https://gerrit.wikimedia.org/r/#/c/295130/ [3] https://gerrit.wikimedia.org/r/#/c/295129/ [4] https://phabricator.wikimedia.org/T136696 [5] https://gerrit.wikimedia.org/r/#/c/290765/ [6] https://gerrit.wikimedia.org/r/#/c/295123/ [7] https://phabricator.wikimedia.org/T137422 [8] https://phabricator.wikimedia.org/T111950
That's certainly the most productive kitchen I've ever seen!
Thank you everyone!
Le 20/06/2016 12:55, Guillaume Lederrey a écrit :
Hello all!
I had a few friends coming over this Sunday for a mini Hackathon in my kitchen. The goal was mainly to introduce them to what we do around Wikipedia, have fun, exchange ideas and maybe do actual work on a few tickets. Quick summary of what we did:
- Fix a few puppet unit tests [1][2]
- Try to fix sending logs to logstash from elasticsearch [3][4]. We
could not reproduce the issue with security manager with a local installation of elasticsearch and I did not prepare access to deployment-prep where we could have validated the issue.
- Added a few tests to our elasticsearch admin tool (estool) [5]
- Refactoring of Cassandra monitoring [6][7] (thanks to Luca for the
fast code review!). This one prompted a great conversation. It was fairly unclear to us how to best integrate monitoring in a DRY way, while keeping the core of the Cassandra module free from external dependencies. We are all used to have a technical module (puppet-cassandra) that only manage core Cassandra and a higher level module (puppet-wmf_cassandra) that configure Cassandra in the context of WMF, with the all peripheral fonctions (monitoring, firewall, logging, ...). This higher level abstraction kind of exists in the roles, but multiple roles might use Cassandra, and might lead to some duplication. The change [6] is probably more the start of a conversation than something that can be merged as is.
- We tried to have a go at adding unit tests to Kartotherian [8], but
ran into an issue with mapnik and were not able to build the project. None of us had any experience with Node development or Kartotherian. We moved to other stuff.
The day was fun! Thanks to all who joined! Special +1 to Mara who put a lot of energy in learning Python for the first time (and succeeded!).
I'll try to followup on the changed opened during that day, make sure that those contributions, however small, are not forgotten.
We will be back!
MrG
[1] https://gerrit.wikimedia.org/r/#/c/295127/ [2] https://gerrit.wikimedia.org/r/#/c/295130/ [3] https://gerrit.wikimedia.org/r/#/c/295129/ [4] https://phabricator.wikimedia.org/T136696 [5] https://gerrit.wikimedia.org/r/#/c/290765/ [6] https://gerrit.wikimedia.org/r/#/c/295123/ [7] https://phabricator.wikimedia.org/T137422 [8] https://phabricator.wikimedia.org/T111950
That's awesome, but, was there food too? :)
Thanks for all your hard work, good job!
-- Deb Tankersley Product Manager, Discovery IRC: debt Wikimedia Foundation
On Mon, Jun 20, 2016 at 5:33 AM, David Causse dcausse@wikimedia.org wrote:
That's certainly the most productive kitchen I've ever seen!
Thank you everyone!
Le 20/06/2016 12:55, Guillaume Lederrey a écrit :
Hello all!
I had a few friends coming over this Sunday for a mini Hackathon in my kitchen. The goal was mainly to introduce them to what we do around Wikipedia, have fun, exchange ideas and maybe do actual work on a few tickets. Quick summary of what we did:
- Fix a few puppet unit tests [1][2]
- Try to fix sending logs to logstash from elasticsearch [3][4]. We
could not reproduce the issue with security manager with a local installation of elasticsearch and I did not prepare access to deployment-prep where we could have validated the issue.
- Added a few tests to our elasticsearch admin tool (estool) [5]
- Refactoring of Cassandra monitoring [6][7] (thanks to Luca for the
fast code review!). This one prompted a great conversation. It was fairly unclear to us how to best integrate monitoring in a DRY way, while keeping the core of the Cassandra module free from external dependencies. We are all used to have a technical module (puppet-cassandra) that only manage core Cassandra and a higher level module (puppet-wmf_cassandra) that configure Cassandra in the context of WMF, with the all peripheral fonctions (monitoring, firewall, logging, ...). This higher level abstraction kind of exists in the roles, but multiple roles might use Cassandra, and might lead to some duplication. The change [6] is probably more the start of a conversation than something that can be merged as is.
- We tried to have a go at adding unit tests to Kartotherian [8], but
ran into an issue with mapnik and were not able to build the project. None of us had any experience with Node development or Kartotherian. We moved to other stuff.
The day was fun! Thanks to all who joined! Special +1 to Mara who put a lot of energy in learning Python for the first time (and succeeded!).
I'll try to followup on the changed opened during that day, make sure that those contributions, however small, are not forgotten.
We will be back!
MrG
[1] https://gerrit.wikimedia.org/r/#/c/295127/ [2] https://gerrit.wikimedia.org/r/#/c/295130/ [3] https://gerrit.wikimedia.org/r/#/c/295129/ [4] https://phabricator.wikimedia.org/T136696 [5] https://gerrit.wikimedia.org/r/#/c/290765/ [6] https://gerrit.wikimedia.org/r/#/c/295123/ [7] https://phabricator.wikimedia.org/T137422 [8] https://phabricator.wikimedia.org/T111950
discovery mailing list discovery@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/discovery