Test inserting into a capped collection with multiple secondary indexes
Test
Batches of 1000 documents are inserted into a capped collection with multiple secondary indexes.
Results are reported as ops / sec.
Setup
- collection is capped at 1GB.
- collection count, not enabled by default, is capped at 1024*1024. The document size is 1KB, so this also results in a size of 1GB.
Notes
- For all the tests, a varying number of threads is used.
- Capped collection are not supported in a sharded configurations.
- Test time is 3 minutes.
- This is based on insert processes that MMS/Cloud Manager uses. Inserts like this caused a lot of headaches in the 3.0 timeframe.
- Typical results are unusual: Throughput over thread levels is flat for MMAPv1 (makes sense) and actually decreases with WT with more thread levels. This makes sense: the deletion is single threaded, so you cannot insert new stuff faster than that. (It's throttled on purpose.)
Owning-team
mongodb/storage-execution
- Source:
Members
(inner) thread_levels
The number of threads to run in parallel. The default is [1, 2].
The actual values in use are injected by run_workloads.py, which gets it from config file, see this hello world example.