Time/Place
- Time: 11:00am Eastern Standard Time US (UTC-5)
- Dial-in Number: (712) 775-7035
- Participant Code: 479307#
- International numbers: Conference Call Information
- Web Access: https://www.freeconferencecallhd.com/wp-content/themes/responsive/flashphone/flash-phone.php
Attendees
- Nick Ruest
- Andrew Woods
- Aaron Elkiss (umich)
- David Wilcox
- Doron Shalvi
- Ed Fugikawa
- Esme Cowles
- Joshua Westgard
- Steve DiDomenico
- Martin Haye
- Scott Prater
Agenda
- Updates
- Tickets
- Results
- What lessons have we learned from tests already run?
- Nick Ruest, Martin Haye: Ingest slowdown after x-number or resources loaded
- Tickets
- Create a quick start guide for running the tests?
- https://gist.github.com/ruebot/f1979ba1e1ad1e519541c2aeb3b49f2c
- A vagrant box with Puppet that installs Fedora 4. Less complete than Nick's guide above for tests, but different environment (CentOS, OpenJDK). https://github.com/nulib/fcrepo4-vagrant
- YourKit profiling
- Modifying/adapting tests for Islandora and Hydra testing
- Hack-house event in the Fall?
- It would be great to have performance baselines before transitioning to Modeshape5
- ...
Minutes
Tickets
- Need to create tests 5 and 6
- Self-assign tickets for running tests
- Scott will do MySQL and PostgreSQL
- Martin to run Test - 4
- Not necessary to run on AWS
- Best to run on remote machine
- JMeter client also consumes resources, best to have on separate machine
- Test-1
- Aaron to update documentation for running this
- Remove Test-3
Results
Princeton
- leveldb - 737k resources before failure
- slightly faster than an untuned postgres
- postgres - never timed out, 3.7mil before network disconnect
- opportunity for tunning, was tested with stock install
- adding index on key would help
- opportunity for tunning, was tested with stock install
determine units of jmeter results
* Label - The label of the sample. If "Include group name in label?" is selected, then the name of the thread group is added as a prefix. This allows identical labels from different thread groups to be collated separately if required. * Num Samples - The number of samples with the same label * Average - The average elapsed time of a set of results * Min - The lowest elapsed time for the samples with the same label * Max - The longest elapsed time for the samples with the same label * Std. Dev. - the Standard Deviation of the sample elapsed time * Error % - Percent of requests with errors * Throughput - the Throughput is measured in requests per second/minute/hour. The time unit is chosen so that the displayed rate is at least 1.0. When the throughput is saved to a CSV file, it is expressed in requests/second, i.e. 30.0 requests/minute is saved as 0.5. * Kb/sec - The throughput measured in Kilobytes per second * Avg. Bytes - average size of the sample response in bytes. (in JMeter 2.2 it wrongly showed the value in kB)
Madison
- Focused on resource consumption
- Ran for ~48 hours
- CPU settled down after initial warm-up
- Will use jmeter graphing component next
- Will retrieve the number of resources created
York
- Timed out at 2.5mil resources
- Restarted tomcat and updated max file handles
- reached ~9.4mil resources
- Used leveldb
- It would be useful to run tests with alternate GC flags
Lessons learned
- -XX:+UseG1GC
- See http://www.oracle.com/technetwork/tutorials/tutorials-1876574.html
- Martin would like to run Test-4 in different memory configurations
Islandora/Hydra testing
- testing with interaction patterns based on these upper apps
hack house
- maybe in the fall
Actions
- Esmé Cowles to run other tests (1 and 2)
- Martin Haye to run test 4
- Nick Ruest to run other tests pending documentation from Aaron
- Aaron Elkiss to add docs next week
- Scott Prater, test 4, jmeter graphing
- Yourkit: Scott, UMd?, Doron, Nick
- Steve: vagrant to deploys F4
Next meeting
- May 16th next meeting
1 Comment
Andrew Woods
+1 on the quick start, Nick Ruest:
https://gist.github.com/ruebot/f1979ba1e1ad1e519541c2aeb3b49f2c