Housekeeping#

Archiving superseded or obsolete outputs#

To ensure storage is not wasted, owners of datasets produced by the Analysis Productions system and WG liaisons will occasionally need to clean up datasets that are no longer needed by analysts.

This operation can be performed on the Analysis Productions webpage.

Important

Archiving dataset(s) will preserve the production metadata, and will not immediately delete data from disk. However, the LHCb data management team may decide to delete archived datasets at any moment.

  1. Navigate to your production and click Select to begin selecting datasets.

  2. Click on the table row of each dataset you would like to archive.

  3. Click Finalise selection at the top of the table.

  4. Confirm your selection carefully! Then, select Archive.

  5. Confirm once more by ticking I understand, and clicking Archive again.

The web page should now confirm whether the operation succeeded or failed, and the archived datasets will disappear from the analysis table upon refreshing.

Reusing Samples#

Given the broad overlap in required samples for analyses it is possible that when collecting the data/MC for a new analysis many of the required tuples already exist in the Analysis Productions database. Therefore rather than retupling this data you can simply assign existing tuples to a new analysis. To do this:

  1. Open the Analysis Productions webpage

  2. Click Create new analysis.

  3. Select the working group and a name for your production.

  4. Use the filter tabs to filter the shown samples.

  5. Select any samples you wish to add to your new production.

  6. Click Add N samples.

If you simply want to assign a given production to many WGs you can do so by following the steps above for each new WG and selecting all samples in the original production.

You can also add samples from one existing production to another existing production by doing the following:

  1. Open the Analysis Productions webpage

  2. Navigate to the production you want to add samples to.

  3. Click Add samples

  4. Use the filter tabs to filter the shown samples.

  5. Select any samples you wish to add to your production.

  6. Click Add N samples.

Instructions for liaisons#

  1. Review the merge requests for your working group.

    • Size of the output file doesn’t seem excessive.

    • The analysis isn’t including too many variables in the ntuple.

    • Encourage people to share productions between analyses where practical.

    • There is no need to be too strict, it’s more important that productions are submitted promptly.

  2. Approve and merge! N.B. This requires membership in the lhcb-dpa-emtf-rta-liaisons egroup. You are responsible for subscribing to it yourself.

  3. When the CI pipeline finishes, ensure that a JIRA task was automatically created here.

    • The title should be “WG - ANALYSIS - AnalysisProductions VERSION”.

    • @lhcbdpa should post the link to the JIRA task on the Merge Request.

    • The link to the JIRA task is additionally shown in the logs of the CI pipeline.

Additionally, refer to the `DPA documentation for liaisons<https://lhcb-dpa.web.cern.ch/lhcb-dpa/liaisons.html>`__.