Repository navigation
fix(self-managed): backport the NATS reconnect and Cassandra replica-down fixes [self-managed 1.1] - #3013
Merged
sbaum1994 merged 3 commits intoOct 29, 2026
Conversation
…ad timeout [self-managed 1.1] Cassandra 0.23.2 sets range_request_timeout to 5s (#2918), so a secondary index read that picked a crashed replica fails over in time for the driver to retry it on another coordinator. It also moves the migrations image to 0.18.1 (#2740, runtime dependency patch, no schema change). Relates to #2893 Signed-off-by: shobham <shobham@nvidia.com>
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. 🗂️ Base branches to auto review: 1
Please check the settings in the CodeRabbit UI or the
|
…[self-managed 1.1] Signed-off-by: shobham <shobham@nvidia.com>
This was referenced Oct 29, 2026
Closed
sbaum1994
approved these changes
Oct 29, 2026
sbaum1994
enabled auto-merge (squash)
October 29, 2026 19:37
sbaum1994
merged commit Oct 29, 2026
443651a
into
release-deploy/stacks/self-managed/v1.1
24 checks passed
Contributor
|
This PR is included in version 1.1.1. The release is available on GitHub release. |
This file contains hidden or 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
TL;DR
Backports two fixes to the self-managed 1.1 stack in one patch: the NATS reconnect fix for function workers (worker-utils 1.4.2) and the Cassandra replica-down fixes (#2918). Supersedes #3003.
Additional Details (optional for docs, build, test, refactor, ci, chore, style, and revert PRs)
helmfile.d/01-dependencies.yaml.gotmpl:cassandra0.23.0->0.23.2.range_request_timeout: 5000ms. A secondary-index read that picked a crashed replica now times out at 5 s instead of 10 s, leaving the driver time to retry it on another coordinator.0.18.0->0.18.1(fix(cassandra): patch migration image runtime dependencies #2740, runtime dependency patch, no schema change).helmfile.d/02-core.yaml.gotmpl:api1.33.0->1.33.2, the chart release cut from fix(cloud-functions): default nvcf-api to 1.28.2 for the Cassandra replica-down fixes #3012.RangeReadRetryPolicyin the ncp profile), and retriable Cassandra failures return 503 withRetry-Afterinstead of a 500 with CQL.For the Reviewer
utilsContainerImagekeep it.For QA (optional for docs, build, test, refactor, ci, chore, style, and revert PRs)
nats: no heartbeat receivedafterreconnected to nats.nodetool gettimeout rangereports 5000 ms on every node. On a 3-node HA install, run function lookup traffic andkill -9the Cassandra process on one node. Expect no stall longer than about 5 s, and no 500s from read timeouts.Issues
Relates to #2832 and #2893.
Checklist