DUE TO SPAM, SIGN-UP IS DISABLED. Goto Selfserve wiki signup and request an account.
Date
Notification thread
TBD
Meeting Link
https://meet.google.com/mzq-iwcc-xvw
Attendees
- Jason Gerlowski
- Andrey Ukhanov
- Anshum Gupta
- Bryan Jacobowitz
- Carlos Ugarte
- Chan Chan
- David Smiley
- Dewane Woodley
- Hoss Hostetter
- Houston Putman
- Jan Høydahl
- Jude Muriithi
- Kevin Liang
- Luke Kot-Zaniewski
- Matthew Biscocho
- Nazerke S.
- Parveen Saini
- sanjay dutt
Discussion Topics
- Demo Solr port of OpenSearch Benchmark tool + discuss performance benchmarking follow up (Kevin / Jan)
- https://github.com/janhoy/solr-benchmark
- David pointed out naming conflict with solr/benchmark. Should it have "rally" in the name since its a fork of it?
- Solr Performance Benchmarking
- How are data sets defined? How are queries defined?
- Still requires a level of Solr integration, regardless of the tool
- Discussion - https://lists.apache.org/thread/s6rh06xlsy8mbno9qvy205s761mhm7th"Ranking Drift" Tooling Demo (Parveen Saini)
- Slides -
- Repo - https://github.com/parveensaini/solr-lucene-migration-correctness
- Has this issue been seen by others during major version upgrades?
- David mentioned Lucene changes from 7 to 8. Otherwise, things have generally been steady
- Hoss called out smaller changes that could amount to bigger diffs
- Luke mentioned changes to edismax parser in 9 (from 8?)
Benchmarking: New "SolrBackend" – and WIP for adapting solr/benchmark to use it, will share a PR soon. Similarly for SolrClientTestRule to use SolrBackend. (David)
- Ran out of time. Decided that we should continue the benchmarking discussion asynchronously (mailing list, confluence page)
Post-Meeting Summary
- ...