search

stripe-archive / timberlake

Timberlake is a Job Tracker for Hadoop.
MIT License
177 stars 23 forks source link

Return back ability to read binary logs #82

Open maxtepkeev opened 6 years ago

maxtepkeev commented 6 years ago

Hi,

We're using Timberlake for several years already, we recently decided to upgrade to the latest version and found out that after cdacf5406e4419c0c472665a48b44b541e66cbfa was merged, Timberlake is not able to parse jhist in binary format anymore. We know that we can switch log format to json, but did you consider supporting both formats ? You have the code for both formats already there, what needs to be done is just to parse header and execute the right parser.

Thoughts ?

dug-stripe commented 6 years ago

Hi Max, thanks for flagging. This isn't on our current roadmap, but we're happy to review a PR to restore this functionality. Please include tests with any such PR – since we don't use this log format ourselves, the behavior is likely to regress again in absence of such tests.