Details
-
Question
-
Resolution: Cannot Reproduce
-
Major - P3
-
None
-
1.6.1
-
We are running MongoDB version 1.6.5 on CentOS release 5.5 (Final).
Description
On a mongo collection with 1.8 Million documents the following steps were taken:
> db.collection.dropIndexes()
> db.collection.ensureIndex() // ran this multiple times for the various indexes on it the collection
At this point the RPS of the application, and mongo's throughput were a quarter of what they were before the indexes were dropped and readded.
Then i ran a
> db.collection.reIndex()
Which brought the throughput and the RPS back up to the levels it was before the indexes were dropped.
What i wanted to know was is this expected behavior, and is it recommended to run a reIndex after an index is added to an existing collection.