-
Notifications
You must be signed in to change notification settings - Fork 3.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
release-24.3: roachtest: add failover
variants with leader leases
#133369
Merged
nvanbenschoten
merged 1 commit into
release-24.3
from
blathers/backport-release-24.3-131624
Oct 29, 2024
Merged
release-24.3: roachtest: add failover
variants with leader leases
#133369
nvanbenschoten
merged 1 commit into
release-24.3
from
blathers/backport-release-24.3-131624
Oct 29, 2024
+24
−4
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
Part of #132762. Leader leases have different availability properties than epoch leases under most failure modes. This patch adds failover test variants that use leader leases where possible. Initial test results: | test | lease=epoch (ms) | lease=expiration (ms) | lease=leader (ms) | parity with expiration | |:---------------------------------------------|-----------------:|----------------------:|------------------:|:----------------------:| | failover/chaos/read-only | 60,129 | 18,253 | 60,129 | ✔ | | failover/chaos/read-write | 60,129 | 20,401 | 60,129 | XXX | | failover/liveness/blackhole | 9,663 | 369 | 335 | ✔ | | failover/liveness/blackhole-recv | 11,274 | 402 | 369 | ✔ | | failover/liveness/blackhole-send | 9,663 | 385 | 469 | ✔ | | failover/liveness/crash | 8,053 | 352 | 318 | ✔ | | failover/liveness/deadlock | 24,696 | 385 | 369 | ✔ | | failover/liveness/disk-stall | 26,843 | 369 | 419 | ✔ | | failover/liveness/pause | 10,200 | 385 | 436 | ✔ | | failover/non-system/blackhole | 7,247 | 7,516 | 15,032 | XX | | failover/non-system/blackhole-recv | 12,348 | 10,737 | 18,253 | XX | | failover/non-system/blackhole-send | 6,979 | 6,979 | 8,053 | X | | failover/non-system/crash | 7,247 | 6,979 | 9,126 | X | | failover/non-system/deadlock | 60,129 | 60,129 | 60,129 | ✔ | | failover/non-system/disk-stall | 22,548 | 22,548 | 25,769 | X | | failover/non-system/pause | 7,247 | 7,247 | 9,126 | X | | failover/partial/lease-gateway | 8,589 | 19,327 [^1] | 60,129 | XXX | | failover/partial/lease-leader | 60,129 | 22,549 [^2] | 31,139 [^2] | XX | | failover/partial/lease-liveness | 8,589 | 301 | 318 | ✔ | | failover/system-non-liveness/blackhole | 369 | 402 | 352 | ✔ | | failover/system-non-liveness/blackhole-recv | 335 | 285 | 318 | ✔ | | failover/system-non-liveness/blackhole-send | 402 | 419 | 335 | ✔ | | failover/system-non-liveness/crash | 419 | 301 | 453 | ✔ | | failover/system-non-liveness/deadlock | 369 | 352 | 402 | ✔ | | failover/system-non-liveness/disk-stall | 402 | 318 | 453 | ✔ | | failover/system-non-liveness/pause | 369 | 385 | 335 | ✔ | _note: because of the way the test measures pMax, anything under 1,000ms is essentially "no impact"_ **Key _(comparing leader vs. expiration)_**: ✔ = parity X = minor regression XX = major regression XXX = unavailability [^1]: I don't understand why expiration-based lease perform worse than epoch-based leases on this test. [^2]: With #133214. Epic: none Release note: None
blathers-crl
bot
added
blathers-backport
This is a backport that Blathers created automatically.
O-robot
Originated from a bot.
labels
Oct 24, 2024
blathers-crl
bot
requested review from
nameisbhaskar and
vidit-bhat
and removed request for
a team
October 24, 2024 17:01
blathers-crl
bot
requested review from
arulajmani,
DarrylWong,
iskettaneh,
nvanbenschoten and
srosenberg
October 24, 2024 17:01
Thanks for opening a backport. Please check the backport criteria before merging:
If your backport adds new functionality, please ensure that the following additional criteria are satisfied:
Also, please add a brief release justification to the body of your PR to justify this |
blathers-crl
bot
added
the
backport
Label PR's that are backports to older release branches
label
Oct 24, 2024
arulajmani
approved these changes
Oct 25, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Labels
backport
Label PR's that are backports to older release branches
blathers-backport
This is a backport that Blathers created automatically.
O-robot
Originated from a bot.
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.
Backport 1/1 commits from #131624 on behalf of @nvanbenschoten.
/cc @cockroachdb/release
Part of #132762.
Leader leases have different availability properties than epoch leases under most failure modes. This patch adds failover test variants that use leader leases where possible.
Initial test results:
note: because of the way the test measures pMax, anything under 1,000ms is essentially "no impact"
Key (comparing leader vs. expiration):
✔ = parity
❌ = minor regression
❌❌ = major regression
❌❌❌ = unavailability
Epic: none
Release note: None
Release justification: testing only
Footnotes
I don't understand why expiration-based lease perform worse than epoch-based leases on this test. ↩
With roachtest: enable DistSender circuit breakers in failover/partial/lease-leader #133214. ↩ ↩2