Database Meeting 2022-12-14

On-going planning for the upcoming DMLT in Chile (March 20223)

No major news from DMLT, except for this:

• (also from the data facilities meeting) an interesting discussion (mainly questions) on USDF support during the SLAC Winter shutdowns when LSST will get into operations

Fritz Mueller do we have a list of things to be wrapped up before the end of the year?

• the previous planning cycle ended on December 1st.
• hence, no specific action items are to be formally done before the end of the year

Colin Slater on the ongoing discussions on the data previews:

• an idea of updating catalogs as we're doing data reviews progressively. Periodically releasing the Object table as we add new data (regions of the Sky) for Q&A.
• the context: this is going to happen at the very start of the data-taking. Alerts will be released. People will need to see those. And a lack of static information will be a problem here.
• chances are that we will also need this for the Source table (to be updated nightly) as well

Action item for Igor Gaponenko to work on extending the Replication/Ingest system to support the incremental update of tables.

 Context:

• continue discussing the topic
• presently, a focus on improving worker-side Qserv monitoring:
  • DM-36965 - Getting issue details... STATUS
• next step - investigating and fixing the reason for the lockups:
  • DM-37334 - Getting issue details... STATUS

Action items:

• Igor Gaponenko will assist John Gates in deploying the test PRs in the slac6  

Fritz Mueller started playing with the API. Most of it works. However, status queries against the information schema hang. Something is wrong in czar 's the handling of the information schema queries.

Andy Salnikov there is unit testing for this. something might be broken.

Action item for Fritz Mueller - attach the corresponding JIRA ticket here for future reference.

Context:

Fabrice Jammes wanted to discuss 2 topics:

• improvements to the REST API (Igor Gaponenko)
• ran into troubles with the implementation aspects of some "entry points" (of the Qserv container). The binary of the replication service was not starting because some hidden error triggered in the entry point. Though, at the surface it all looked fine.