Details
-
Bug
-
Resolution: Fixed
-
Major
-
7.0.2
-
Untriaged
-
1
-
Unknown
Description
This test is a part of cost effective low-end clusters - product improvements.
http://showfast.sc.couchbase.com/#/timeline/Linux/cloud/lowend/idle
In this test, we do the following steps.
- launch aws ec2 instances
- set up cluster
- load data and create indexes
- sleep for 30 minutes
- measure CPU utilization of the bucket and individual component when the cluster is idle
We see higher CPU utilization reported by ns_server in 7.0.1/2. The issue is reproducible. I open this issue to investigate why CPU utilization is higher in 7.0.1/2 when the cluster is idle.
build: 7.0.0-5302
job: http://perf.jenkins.couchbase.com/job/Cloud-Tester/315/
logs:
build: 7.0.2-6511
job: http://perf.jenkins.couchbase.com/job/Cloud-Tester/314/
Attachments
Issue Links
- backports to
-
-
- Closed
-
| For Gerrit Dashboard: MB-48157 | ||||||
|---|---|---|---|---|---|---|
| # | Subject | Branch | Project | Status | CR | V |
| 160085,2 | MB-48017 Reduce idle time CPU utilisation of projector | cheshire-cat | indexing | Status: MERGED | +2 | +1 |
| 160118,4 | MB-48157 Reduce idle time CPU utilisation of projector | unstable | indexing | Status: MERGED | +2 | +1 |
| 160291,2 | MB-48157 Use lastFlushTime instead of lastActiveTime to decide on flush | unstable | indexing | Status: MERGED | +2 | +1 |
| 162642,9 | MB-48157 Do not invoke syncWorker when there are no mutations | unstable | indexing | Status: MERGED | +2 | +1 |