test(perf): Accumulated Block Test fixes + baselines #3940
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
roypat
added
the
Status: Awaiting review
pb8o
reviewed
Jul 6, 2023
kalyazin
reviewed
Jul 6, 2023
Make sure that fio workers inside the guest during block performance test are pinned to specific vcpus. Analyzing the raw data from block performance tests indicated that each fio worker has a unique performance characteristic, but that two workers sometimes "swap" their characteristics. I am hypothesizing that this is due to scheduling inside the guest. Signed-off-by: Patrick Roy <roypat@amazon.co.uk>
The variable determined how many fio workers should be started per vcpu. It was always set to 1, and its unclear why we would ever want it to be not 1, or what the impact of that would be. Signed-off-by: Patrick Roy <roypat@amazon.co.uk>
roypat
force-pushed
the
pin-fio-process-in-guest
branch
from
79545ee to
067b1ae
Compare
Makes its easier to see what tests are run/failing. Signed-off-by: Patrick Roy <roypat@amazon.co.uk>
The randrw fio mode tries to perfectly balance reads and writes. This can be seens from the perfect correlation between the bw_read and bw_write metrics (r=1) emitted by this test. This means we really only measure read OR write throughput (whicheverone is lower throttles the other). Therefore, to make sure we actually measure write throughput, use randread and randwrite as our two fio modes. Also removes the associated baselines. New baselines will be gathered in follow up. Signed-off-by: Patrick Roy <roypat@amazon.co.uk>
this caused pytest to crash during test collection if running on an unsupported kernel version. Fixes firecracker-microvm#3726 Signed-off-by: Patrick Roy <roypat@amazon.co.uk>
roypat
force-pushed
the
pin-fio-process-in-guest
branch
from
cf8f02b to
e374bcb
Compare
Collect new block baselines for all instances/kernel versions since we added a new test case, and the change of parameterization from baseline-level to pytest-level caused a change in performance behavior across the board (since previously, multiple tests would reuse the same microvm). Signed-off-by: Patrick Roy <roypat@amazon.co.uk>
roypat
changed the title
pb8o
approved these changes
Jul 7, 2023
kalyazin
approved these changes
Jul 7, 2023
3 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR accumulates some fixes to the block performance test that all each require new baselines. These include:
License Acceptance
By submitting this pull request, I confirm that my contribution is made under
the terms of the Apache 2.0 license. For more information on following
Developer Certificate of Origin and signing off your commits, please check
CONTRIBUTING.md.PR Checklist
CHANGELOG.md.TODOs link to an issue.rust-vmm.