Automatically launching a re-index or brute force search from browser requests needs to be configurable
|Reported by:||Patrick Martin||Owned by:||anonymous|
In brief, being able to turnoff the auto-index behaviour for large and busy repositores would allow the post-commit script to handle the re-index, or a nightly task.
The update-index.py script is perfectly serviceable, and admins should have the option of only re-indexing using processes under their control, such as a cron job or Sheduled Task, to avoid a very large re-index from impacting apache's performance and stability.
A large re-index can blow the 2Gb address space, and thus never completes, which means once a repository has failed to index, it can never catch up. Being able to tune the re-index operations to ensure they succeed regularly would help mitigate this risk.
Tested against Trac and latest ReposSearchPlugin 0.10dev
Change History (8)
comment:5 Changed 7 years ago by
|Owner:||changed from Alec Thomas to Ryan J Ollos|
|Status:||assigned → new|
comment:6 Changed 7 years ago by
|Summary:||Automatically launching a re-index or brute force search from browser resquests needs to be configurable → Automatically launching a re-index or brute force search from browser requests needs to be configurable|
|Trac Release:||0.10 → 0.11|