Oracle Analytics and AI

User_RQLGR

Catalog Crawler consumes most of the CPUand slows down the BI Server response. I am trying to learn more about how the BI Search crawler works. I'm on OBI 12.2.1.4.230713. The sysadm documentation and KB articles I've found on the topic aren't very helpful.

1. Scheduling crawls: What table in the backend stores these crawl schedules?
2. When I disable crawls, if there are any scheduled they will run regardless. How do I prevent this other than clicking on the cancel button?
3. How do you control crawler configuration such as threads?
4. Has anyone seen this error from a webcat crawler: WebCatalogIndex close error: java.io.IOException:Input/output error? We are experiencing severe performance issues and core dumps and this error seems to precede the crash. Could this error mean the filesystem is full?

Steve F-2

Hi Bryan,

Welcome to Oracle Analytics community!

Some general comments.

1. Ensure you are patched, July 2023 is a bit old, but you may be on OBI 12.2.1.4 on FMW 12.2.1.3 which is out of error correction support.

2. Ensure your crawl frequency is not too frequent. Generally once a week.
Crawl only curated content, generally you do not want to crawl un-curated Users folders.

3. Schedule in off hours.

4. Update to Oracle Analytics Server as soon as you are able, as the indexing engine is different/improved.

You may need a service request to troubleshoot the issue in-depth, with stack trace analysis, thread dumps, etc.
and to answer the specific questions.

Other comments, welcomed.