[Backport release-1.5] fix(seaweedfs): adopt legacy workloads in place across the 4.31 rename - #3326
Closed
github-actions[bot] wants to merge 3 commits into
Closed
[Backport release-1.5] fix(seaweedfs): adopt legacy workloads in place across the 4.31 rename#3326github-actions[bot] wants to merge 3 commits into
github-actions[bot] wants to merge 3 commits into
Conversation
Author
|
Please cherry-pick the changes locally and resolve any conflicts. git fetch origin backport-3282-to-release-1.5
git worktree add --checkout .worktree/backport-3282-to-release-1.5 backport-3282-to-release-1.5
cd .worktree/backport-3282-to-release-1.5
git reset --hard HEAD^
git cherry-pick -x 72db724d5fae4b7be1235c328a6b660a10e611f2 5dc363d3980bdfb54eaff3e98e522c58c0cea5e0
git push --force-with-lease |
myasnikovdaniil
force-pushed
the
backport-3282-to-release-1.5
branch
from
July 16, 2026 17:21
f50af9d to
f2d21ed
Compare
myasnikovdaniil
marked this pull request as ready for review
July 27, 2026 15:24
Contributor
|
Caution The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased. |
Upstream chart 4.31 switched resource names from the chart name to the release name. The data-plane release is <name>-system, so every StatefulSet was renamed to <name>-system-*. StatefulSet names are immutable, so the upgrade could not rename in place: Helm stood up a second, empty set beside the running one while the data stayed on the original data1-seaweedfs-volume-* PVCs. Where the node count let both sets run, they shared one S3 Service and one filer metadata database, so the empty set answered reads and mixed volume IDs from an independent sequence into shared metadata. Pin fullnameOverride so the rendered names no longer follow the release name, matching the convention already used by other system packages (cozy-proxy, flux-operator, linstor-scheduler, victoria-metrics-operator). The s3-consumer suite pinned the name/fullname divergence this override removes by default, so it now re-creates that divergence explicitly and keeps guarding it regardless of the shipped default. Assisted-By: Claude <noreply@anthropic.com> Signed-off-by: Aleksei Sviridkin <f@lex.la> (cherry picked from commit 7bcd8e0) (cherry picked from commit 1da45ead8262334636919c097019afdf71da412e)
With the chart-based names pinned back, an upgrade adopts the running workloads and their volumes in place. Before 4.31 the chart ignored the release name, so this holds for an instance running under any name: its data sits on the chart-named data1-seaweedfs-volume-* PVCs either way. Two states cannot be adopted and stop the render instead, each pointing the operator at the matching recovery step: - Fresh on 1.5.x: only the renamed volumes exist, so adopting the chart-based name would rename the workloads away from that data and bring up an empty cluster. Helm cannot move data between PVCs, so the operator is sent to the PV re-bind procedure. - D-split: both sets exist AND the renamed volume servers are live, so they may hold objects written through the split endpoint. Adopting would strand them, so the operator is sent to reconcile the split first. A duplicate that never served (zero ready replicas) is safe to adopt and renders through. The guard mirrors the chart's fullname helper to recognise what 4.31 named this instance (the helper appends the chart name when the release name does not contain it, then truncates), and matches the renamed volumes by shape plus the chart labels on the StatefulSet rather than reconstructing the exact name. Assisted-By: Claude <noreply@anthropic.com> Signed-off-by: Aleksei Sviridkin <f@lex.la> (cherry picked from commit 052ce2e) (cherry picked from commit 5a724e0bc0d43c8afb621a6da4aab5b13a86e774)
Classify each tenant from its volume PVCs, re-bind the volumes of a tenant installed fresh on 1.5.x so the adopted names pick them up, stop a split cluster before upgrading it, and clear the leftovers the upgrade cannot remove itself. The recovery steps select the renamed set by a precise filter and derive secret names from the fullname helper, so they work for an instance running under a non-default name and for pooled or zoned volumes. Co-authored-by: Myasnikov Daniil <myasnikovdaniil2001@gmail.com> Assisted-By: Claude <noreply@anthropic.com> Signed-off-by: Aleksei Sviridkin <f@lex.la> (cherry picked from commit 088669fb5ae5106aa1d4d22e3226488f557e4b02)
myasnikovdaniil
force-pushed
the
backport-3282-to-release-1.5
branch
from
July 31, 2026 05:25
b0bb6a9 to
fefabbb
Compare
myasnikovdaniil
added a commit
that referenced
this pull request
Jul 31, 2026
Backport of #3339 onto release-1.5, on top of the #3282 adopt-in-place fix (#3326). Three differences from main, all forced by what release-1.5 does not carry: * The repair migration is renumbered 53 -> 45 and stamps 46, since release-1.5 tops out at migration 44 with targetVersion 45. platform values.yaml goes to targetVersion 46. * Migrations 43 and 45 keep release-1.5's inline labelled stamp instead of lib/cozystack-version.sh. That helper is the shared stamp refactor (#2980), whose backport (#3065) was closed, so it does not exist on this branch. Only lib/seaweedfs-db-adopt.sh -- the part that actually carries the fix -- is sourced. * The extra/seaweedfs post-delete cleanup hook and its three tests are dropped. The hook arrived with #3092, a breaking change that was never backported, so there is nothing here for #3339's changes to it to apply to. Comments in lib/seaweedfs-db-adopt.sh and the recovery runbook that promised the hook would reclaim an orphaned volume are corrected to say it does not on 1.5.x. The migration renumbering is carried through the shared helper, the bats suite (STAMP 54 -> 46) and the runbook prose. Verified locally: helm unittest green for both seaweedfs charts (51 + 16 assertions) including both blind-upgrade guard canaries, and 34 bats assertions across migration-seaweedfs-db-adopt, seaweedfs-naming-audit and seaweedfs-guard-parity. Assisted-By: Claude <noreply@anthropic.com> Signed-off-by: Myasnikov Daniil <myasnikovdaniil2001@gmail.com> (cherry picked from commit 88e9fdcbbbff8aeab317b08557ca6bd6c8b5706d)
myasnikovdaniil
added a commit
that referenced
this pull request
Jul 31, 2026
Backport of #3339 onto release-1.5, on top of the #3282 adopt-in-place fix (#3326). Three differences from main, all forced by what release-1.5 does not carry: * The repair migration is renumbered 53 -> 45 and stamps 46, since release-1.5 tops out at migration 44 with targetVersion 45. platform values.yaml goes to targetVersion 46. * Migrations 43 and 45 keep release-1.5's inline labelled stamp instead of lib/cozystack-version.sh. That helper is the shared stamp refactor (#2980), whose backport (#3065) was closed, so it does not exist on this branch. Only lib/seaweedfs-db-adopt.sh -- the part that actually carries the fix -- is sourced. * The extra/seaweedfs post-delete cleanup hook and its three tests are dropped. The hook arrived with #3092, a breaking change that was never backported, so there is nothing here for #3339's changes to it to apply to. Comments in lib/seaweedfs-db-adopt.sh and the recovery runbook that promised the hook would reclaim an orphaned volume are corrected to say it does not on 1.5.x. The migration renumbering is carried through the shared helper, the bats suite (STAMP 54 -> 46) and the runbook prose. Verified locally: helm unittest green for both seaweedfs charts (51 + 16 assertions) including both blind-upgrade guard canaries, and 34 bats assertions across migration-seaweedfs-db-adopt, seaweedfs-naming-audit and seaweedfs-guard-parity. Assisted-By: Claude <noreply@anthropic.com> Signed-off-by: Myasnikov Daniil <myasnikovdaniil2001@gmail.com> (cherry picked from commit 88e9fdcbbbff8aeab317b08557ca6bd6c8b5706d)
1 task
Contributor
|
Closing as superseded by #3370, which merged to This PR and #3370 were divergent reworks of the same seaweedfs base rather than a clean stack: #3370 carried its own copies of all three commits here plus the rename-fallout commit on top, and its versions were the newer ones — its Nothing is lost by closing rather than merging. The original PR this backported, #3282, is still correctly accounted for on the branch: |
1 task
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
2 participants
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.
Description
Backport of #3282 to
release-1.5.