fix: More migration tests and related fixes #2170
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
This expands the migration test quite a bit. It now tests alll the way until retirement of the original path, for both port-only (`rebinding_port`) and address-and-port (`rebinding_address_and_port`) changes. `rebinding_address_and_port` is succeeding, but `rebinding_port` is currently failing. That's because we treat it differently for some reason. If we replace `Paths::find_path_with_rebinding` with `Paths::find_path`, i.e., do proper path validation when only the port changes, the test succeeds. Leaving this out in case I'm missing something about the intent of the difference.
|
In a related QNS test, we have the problem that the client does not issue a new connection ID when the server retires the path, and then we time out because the client wont accept any more packets from the server. I was hoping to reproduce this with this test, but it's behaving differently. Sigh. |
Failed Interop TestsQUIC Interop Runner, client vs. server neqo-latest as client
neqo-latest as server
All resultsSucceeded Interop TestsQUIC Interop Runner, client vs. server neqo-latest as client
neqo-latest as server
Unsupported Interop TestsQUIC Interop Runner, client vs. server neqo-latest as client
neqo-latest as server
|
We need to do a path challenge even if only the remote port changes.
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #2170 +/- ##
==========================================
- Coverage 95.37% 95.36% -0.01%
==========================================
Files 112 112
Lines 36449 36427 -22
==========================================
- Hits 34762 34738 -24
- Misses 1687 1689 +2 ☔ View full report in Codecov by Sentry. |
larseggert
changed the title
larseggert
requested review from
KershawChang,
martinthomson and
mxinden
as code owners
mxinden
reviewed
Oct 19, 2024
Benchmark resultsPerformance differences relative to 978aa4e. coalesce_acked_from_zero 1+1 entries: 💔 Performance has regressed. time: [105.60 ns 105.99 ns 106.38 ns]
change: [+6.4418% +6.9710% +7.4688%] (p = 0.00 < 0.05)
coalesce_acked_from_zero 3+1 entries: 💔 Performance has regressed. time: [121.28 ns 121.54 ns 121.85 ns]
change: [+3.2560% +3.6974% +4.1287%] (p = 0.00 < 0.05)
coalesce_acked_from_zero 10+1 entries: 💔 Performance has regressed. time: [120.97 ns 121.34 ns 121.83 ns]
change: [+2.9226% +3.9609% +5.1498%] (p = 0.00 < 0.05)
coalesce_acked_from_zero 1000+1 entries: 💔 Performance has regressed. time: [100.34 ns 100.58 ns 100.86 ns]
change: [+1.7375% +2.5903% +3.4053%] (p = 0.00 < 0.05)
RxStreamOrderer::inbound_frame(): Change within noise threshold. time: [112.01 ms 112.06 ms 112.12 ms]
change: [-0.5045% -0.3071% -0.1760%] (p = 0.00 < 0.05)
transfer/pacing-false/varying-seeds: No change in performance detected. time: [26.330 ms 27.320 ms 28.314 ms]
change: [-6.4928% -1.2695% +4.2777%] (p = 0.65 > 0.05)
transfer/pacing-true/varying-seeds: No change in performance detected. time: [33.805 ms 35.468 ms 37.131 ms]
change: [-11.482% -5.5438% +0.0689%] (p = 0.08 > 0.05)
transfer/pacing-false/same-seed: No change in performance detected. time: [25.137 ms 25.895 ms 26.647 ms]
change: [-5.3764% -1.1632% +2.9220%] (p = 0.59 > 0.05)
transfer/pacing-true/same-seed: No change in performance detected. time: [40.075 ms 42.002 ms 43.979 ms]
change: [-9.1792% -2.7744% +4.0764%] (p = 0.44 > 0.05)
1-conn/1-100mb-resp/mtu-1500 (aka. Download)/client: No change in performance detected. time: [939.98 ms 948.91 ms 958.02 ms]
thrpt: [104.38 MiB/s 105.38 MiB/s 106.39 MiB/s]
change:
time: [-1.3558% -0.1261% +1.0940%] (p = 0.85 > 0.05)
thrpt: [-1.0821% +0.1263% +1.3745%]
1-conn/10_000-parallel-1b-resp/mtu-1500 (aka. RPS)/client: 💔 Performance has regressed. time: [326.60 ms 329.51 ms 332.51 ms]
thrpt: [30.074 Kelem/s 30.348 Kelem/s 30.619 Kelem/s]
change:
time: [+2.0918% +3.4934% +4.8172%] (p = 0.00 < 0.05)
thrpt: [-4.5958% -3.3755% -2.0489%]
1-conn/1-1b-resp/mtu-1500 (aka. HPS)/client: No change in performance detected. time: [33.511 ms 33.700 ms 33.905 ms]
thrpt: [29.494 elem/s 29.674 elem/s 29.841 elem/s]
change:
time: [-0.9949% -0.1428% +0.7046%] (p = 0.75 > 0.05)
thrpt: [-0.6997% +0.1430% +1.0049%]
1-conn/1-100mb-resp/mtu-1500 (aka. Upload)/client: 💔 Performance has regressed. time: [1.6682 s 1.6906 s 1.7141 s]
thrpt: [58.340 MiB/s 59.151 MiB/s 59.945 MiB/s]
change:
time: [+1.8420% +3.7645% +5.7478%] (p = 0.00 < 0.05)
thrpt: [-5.4354% -3.6279% -1.8087%]
1-conn/1-100mb-resp/mtu-65536 (aka. Download)/client: No change in performance detected. time: [112.25 ms 112.50 ms 112.74 ms]
thrpt: [887.01 MiB/s 888.91 MiB/s 890.87 MiB/s]
change:
time: [-0.4400% -0.1222% +0.2081%] (p = 0.47 > 0.05)
thrpt: [-0.2077% +0.1223% +0.4419%]
1-conn/10_000-parallel-1b-resp/mtu-65536 (aka. RPS)/client: No change in performance detected. time: [316.49 ms 319.67 ms 322.76 ms]
thrpt: [30.982 Kelem/s 31.282 Kelem/s 31.596 Kelem/s]
change:
time: [-2.0843% -0.6487% +0.7861%] (p = 0.38 > 0.05)
thrpt: [-0.7800% +0.6529% +2.1286%]
1-conn/1-1b-resp/mtu-65536 (aka. HPS)/client: 💔 Performance has regressed. time: [34.350 ms 34.533 ms 34.735 ms]
thrpt: [28.790 elem/s 28.958 elem/s 29.112 elem/s]
change:
time: [+1.2829% +2.1211% +3.0221%] (p = 0.00 < 0.05)
thrpt: [-2.9335% -2.0771% -1.2666%]
1-conn/1-100mb-resp/mtu-65536 (aka. Upload)/client: No change in performance detected. time: [279.66 ms 359.17 ms 453.86 ms]
thrpt: [220.33 MiB/s 278.42 MiB/s 357.58 MiB/s]
change:
time: [-13.622% +18.020% +62.614%] (p = 0.33 > 0.05)
thrpt: [-38.505% -15.269% +15.770%]
Client/server transfer resultsTransfer of 33554432 bytes over loopback.
|
Signed-off-by: Lars Eggert <[email protected]>
mxinden
reviewed
Oct 24, 2024
There was a problem hiding this comment.
A couple of clarifying questions. Sorry in case I am missing something obvious. None of them are blocking.
Other than the discussion in #2176 (review), this looks good to me.
mxinden
approved these changes
Oct 25, 2024
…st-more-migration
This reverts commit 97fafa6.
This reverts commit 302961f.
Signed-off-by: Lars Eggert <[email protected]>
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 expands the migration test quite a bit. It now tests all the way until retirement of the original path, for both port-only (
rebinding_port) and address-and-port (rebinding_address_and_port) changes, with and without zero-len client CIDs.This now includes #2182 and #2176.