Bug #4442
Spatial indexing should not run and block metacat
0%
Description
spatial indexing runs after a fresh install the first time someone accesses metacat servlet. It blocks any calls until it is done. This can take over thirty minutes in production knb.
It should run it the background, and possibly be kicked off by a scheduler.
Related issues
History
#1 Updated by Michael Daigle over 12 years ago
moving to 1.9.3 target
#2 Updated by Matt Jones over 10 years ago
Related to bug #3811.
#3 Updated by Redmine Admin about 9 years ago
Original Bugzilla ID was 4442
#4 Updated by ben leinfelder about 9 years ago
- Target version changed from Postpone to 2.1.0
- Assignee changed from Michael Daigle to ben leinfelder
Thoughts:
-if it runs in the background it is entirely possible to get incomplete results if you query as the index is being [re]built.
-we should consider how this search API can be replaced by the SOLR implementation. Showing the point/bounding boxes is a cool feature and one that we don't necessarily get with the coordinates being indexed in SOLR, but it seems like the spatial cache could easily be populated from the SOLR index and that would not block Metacat start-up because indexing will be running essentially stand-alone.
#5 Updated by ben leinfelder almost 9 years ago
- Target version changed from 2.1.0 to 2.x.y