Ticket #1616 (closed defect: fixed)
Catch exceptions when rebuilding the search index
| Reported by: | amercader | Owned by: | amercader |
|---|---|---|---|
| Priority: | major | Milestone: | ckan-sprint-2012-04-02 |
| Component: | ckan | Keywords: | search [a:1.5d] |
| Cc: | Repository: | ckan | |
| Theme: | none |
Description
Right now if an exception is found while reindexing, the whole process stops and the remaining datasets are left out of the index. The process should continue after logging the exception. If more than a certain number of exceptions occur in a row, the process should stop.
Change History
comment:2 Changed 2 years ago by amercader
- Milestone changed from ckan-sprint-2012-01-09 to ckan-v1.6
Sounds like a good option. I will move it to ckan-v1.6 anyway because it may need some work than the expected.
comment:3 Changed 2 years ago by amercader
- Milestone changed from ckan-v1.7 to current-ckan-sprint-2012-04-02
comment:4 Changed 2 years ago by amercader
- Keywords [a:1.5d] added
- Status changed from new to closed
- Resolution set to fixed
Fixed in 10cfd168413
Also new flags and fixes for other commands were added:
- Add -o option to only reindex datasets not already indexed
- Add -i option to ignore exceptions when rebuilding
- Add -r option to just refresh the index (not clearing it first)
- Fix show command to show the index stored for a dataset
- Add support for clearing the index of just one dataset
Note: See
TracTickets for help on using
tickets.
Maybe we just need a '-f' option to force it to continue if an exception occurs? It's a manual process and it's good to know loud and clear if there is a an exception setting up a new server say, but have the option to force it if the system is down temporarily is a critical service.