Troubleshooting in Guzzle - ja-guzzle/guzzle_docs GitHub Wiki
- Guzzle repo is configured in $GUZZLE_HOME/confg/guzzle.yml and $GUZZLE_HOME/api/application.yml : ensure this are same and correctly specified. The first one is used by the jobs while second is used by UI to render the job run details, and support UI login and other details.
- Ensure Repo table structure ties back to the latest guzzle binaries and context columns and stages in guzzle.yml
- To ensure Spark env is working correctly by doing a simple file to file job - read a CSV and write as JSON in the same local dir
- Ensure connectivity is working from respective Spark env - for DB try using notebooks (after importing the lib)
- $GUZZLE_HOME/logs has the job and job group specific logs
- Spark Driver logs on Spark UI, Databricks UI will give additional details
- API console output which typically is in the nohup.out in the $GUZZLE_HOME/api has all the issues/error encountered for any failures in the job which was triggered from UI . If the job is run via Job group then the logs are not captured in API log but in the job group log file in $GUZZLE_HOME/logs
Lot of time the issues one finds during the job runs are due to missing
Some example of this errors are: The below error indicates that the GIT branch used for runing the job does not have the logical end point defined
And this one indicates that current Environment was not choosen when running the job
- API console output which typically is in the nohup.out in the $GUZZLE_HOME/api has all the issues/error encountered when performing any action on UI
- If there are errors in UI which are not getting trapped in API then its a UI level which typically will show up as error in java Script consol - Use Chrome dev too:
