Skip to content

3. Troubleshooting

Jump to bottom
Silvan Heller edited this page Nov 20, 2018 · 2 revisions

My queries take too long for answering.

Try the following steps:

  1. Adjust the Spark parameters (in particular $ADAMPRO_MEMORY, $ADAMPRO_MASTER).
  2. Adjust the ADAMpro parameters; add the following parameters
  • defaultPartitions = 8
  • defaultPartitionsIndex = 8
  • filteringMethod = "semijoin" (for Parquet-based attributes, semijoin is the best option; otherwise choose isinfilter or bloomfilter)
  • VAGlobalRefinement = false (this option is only relevant for VA-File indexes; for large entities choose true)
  1. Vacuum your entities, this will also re-partition your entities according to the attributes set above.
  2. Ensure that your attributes are not distributed over multiple storage engines (see in the list of entities whether all attributes per entity are within one storage engine, otherwise costly joins have to be performed)

Solr - Could not load configuration from directory server/solr/configsets/basic_configs

This can be an issue for both the docker container and locally installed solr. Somehow, solr looks in the wrong config directory. The solution here is to simply copy server/solr/configsets/_defaults to server/configsets/basic_configs

Clone this wiki locally