Details
-
Bug
-
Resolution: Not a Bug
-
Major
-
Cheshire-Cat
-
7.0.0-4245
-
Triaged
-
1
-
Unknown
Description
We are seeing a drop in throughput for the test -
Functions executed/sec, 1 bucket x 100M x 1KB, 4KV + 1Eventing node, single function-BucketOp Source bkt mutation
The test is setup as follows :
- Create the source bucket (bucket1) and metadata bucket (eventing)
- Deploy eventing function which is listening on bucket1 for mutations and applying mutations on the same source bucket bucket1 after executing the function.
- Load 100M Docs on bucket1 and record throughput of the eventing function using the on_update_success stat of eventing.
- After loading 100M Docs, apply access phase mutations on the same bucket1
The regression is seen in 7.0.0-4245
Build | Throughput | Job |
---|---|---|
7.0.0-4241 | 102,596 | http://perf.jenkins.couchbase.com/job/themis/8626/ |
7.0.0-4245 | 74,711 | http://perf.jenkins.couchbase.com/job/themis/8627/ |
Stats Comparison : http://cbmonitor.sc.couchbase.com/reports/html/?snapshot=themis_700-4241_load_access_and_wait_1743&snapshot=themis_700-4245_load_access_and_wait_9f03
The CPU Utilization reduces during the access phase of the test where we apply mutations on the bucket1 as well as the ops rate
Attachments
Issue Links
- is caused by
-
MB-43559 Item paging stops at HWM if there are no replicas
- Closed