FAQ
For tips and advices to deal with unexpected errors, please refer to Troubleshooting guide
Index
Where Can I Find More Assistance to Run source{d} or Notify You About Any Issue or Suggestion?
If you're dealing with an error or something that you think that can be caused by an unexpected error, please refer to our Troubleshooting guide. With the info that you can obtain following those steps, you could fix the problem or you will be able to explain it better in the following channels:
open an issue, if you want to
suggest a new feature, if you need assistance with a contribution, or if you
found any bug.
Visit the source{d} Forum where users and community
members discuss anything source{d} related. You will find there some common questions
from other source{d} users, or ask yours.
and talk with some of our engineers.
How Can I Update My Version Of source{d} CE?
When there is a new release of source{d} CE, it is noticed every time a sourced command is called. When it happens you can download the new version from src-d/sourced-ce/releases/latest, and proceed as it follows:
(You can also follow these steps if you want to update to any beta version, to downgrade, or to use your own built version of source{d} CE)
replace your current version of
sourcedfrom its current location with theone you're installing (see Quickstart. Install),
and confirm it was done by running
sourced version.run
sourced compose downloadto download the new configuration.run
sourced restartto apply the new configuration.
This process will reinstall source{d} CE with the new components, but it will keep your current data (repositories, metadata, charts, dashboards, etc) of your existent workdirs.
If you want to replace all your current customizations —including charts and dashboards—, with the ones from the release that you just installed, the official way to proceed is to prune the running workdirs, and init them again.
disclaimer: pruning a workdir will delete all its data: its saved queries and charts, and if you were using repositories and metadata downloaded from a GitHub organization, they will be deleted, and downloaded again.
sourced status workdirsto get the list of your current workdirsPrune the workdirs you need, or prune all of them at once running
sourced prune --allsourced init [local|orgs] ...for each workdir again, to initialize them withthe new configuration.
How to Update the Data from the Organizations Being Analyzed
There is no way to update imported data, and when a scraper is restarted, it procedes as it follows:
gitcollector
Organizations and repositories are downloaded independently, so if they fail, the process is not stopped until all the organizations and repositories have been iterated.
If gitcollector is restarted, it will download more repositories, but it won’t update any of the already existent ones. You can see the progress of the new process in the welcome dashboard; since already existent repositories won't be updated, those will appear as failed in progress status.
ghsync
The way how metadata is imported by ghsync is a bit different, and it is done sequentially per each organization, so if any step fails, the whole importation will fail.
Pull requests, issues, and users of the same organization, are imported in that order in separate transaction each one, and if one transaction fails, the process will be stopped so the next ones won't be processed.
Once the three different entities have been imported, the organization will be considered as "done", and restarting ghsync won't cause to update its data.
If ghsync is restarted, it will only import data from organizations that could not be finished considering the rules explained above. The process of ghsync will be updated in the welcome dashboard and if an organization was already imported, it will appear as "nothing imported" in the status chart.
Can I Query Gitbase or Babelfish with External Tools?
Yes, as explained in our docs about source{d} CE Architecture, these and other components are exposed to the host machine, to be used by third party tools like Jupyter Notebook, gitbase clients and Babelfish clients.
The connection values that you should use to connect to these components, are defined in the docker-compose.yml, and sumarized in the Architecture documentation
Where Can I Read More About the Web Interface?
The user interface is based in the open-sourced Apache Superset, so you can also refer to Superset tutorials for advanced usage of the web interface.
I Get IOError Permission denied
If you get this error message:
This may happen if you have installed Docker from a snap package. This installation mode is not supported, please install it following the official documentation (See #78).
Last updated