[renovate]

This PR contains the following updates:

Package	Update	Change
opencloudeu/opencloud-rolling	major	7.5.0 → 8.1.0

---

Release Notes

opencloud-eu/opencloud (opencloudeu/opencloud-rolling)

v8.1.0

Compare Source

❤️ Thanks to all contributors! ❤️

@​NickWalters, @​aduffeck, @​butonic, @​dschmidt, @​fschade, @​maki5, @​micbar, @​pascalwengerter, @​pbleser-oc, @​rhafer, @​saw-jan, @​schweigisito, @​v-scharf, @​zerox80

🔒 Security

• feat(proxy): add optional OIDC access token audience validation [#​3466]

🐛 Bug Fixes

• Reindex disabled spaces once they get enabled again [#​3579]
• fix(search): tika key fixes [#​3651]
• fix(search): keep trashed and live folders at the same path apart [#​3602]
• Fix/uploads cli no async consumer [#​3603]
• Remove the timeout when reindexing spaces [#​3543]
• fix(shares): not auto accepting shares created by guest users [#​3533]
• fix(graph): fix PatchMe method to prevent password change [#​3526]

📈 Enhancement

• feat: add new editor roles [#​3637]
• feat(collaboration): let admins disable wopi extensions [#​3633]
• fix(collaboration): send LastModifiedTime and the EuroOffice file size [#​3636]
• feat(collaboration): mobile web view for EuroOffice [#​3635]
• fix(collaboration): harden wopi token handling [#​3630]
• add posixfs index command [#​3068]
• feat(proxy): add per-service metrics to the proxy service [#​3521]

✅ Tests

• fix(search): ellipsize parity matrix labels by runes, not bytes [#​3643]
• fix(collaboration): mint the mobile view test token with the token manager secret [#​3647]
• run cli tests with decomposed nightly [#​3580]
• [decomposed] test(api): remove passing notification tests from expected failure [#​3555]
• api-test: add CLI test for reindexing all spaces including disabled [#​3560]

📚 Documentation

• [SKIP CI] fix: add file_read documentation to audit log docu [#​3503]

📦️ Dependencies

• [full-ci] chore: bump web to v8.1.0 [#​3656]
• chore: bump reva to latest main [#​3642]
• chore: bump libre-graph-api-go to v1.0.8 [#​3645]
• build(deps): bump github.com/nats-io/nats.go from 1.53.1 to 1.54.0 [#​3584]
• build(deps): bump github.com/onsi/gomega from 1.42.1 to 1.43.1 [#​3582]
• build(deps): bump google.golang.org/grpc from 1.83.2 to 1.84.0 [#​3581]
• build(deps): bump github.com/beevik/etree from 1.7.1 to 1.8.0 [#​3583]
• build(deps): bump golang.org/x/image from 0.45.0 to 0.46.0 [#​3573]
• build(deps): bump github.com/nats-io/nats-server/v2 from 2.14.5 to 2.15.0 [#​3576]
• build(deps): bump golang.org/x/text from 0.41.0 to 0.42.0 [#​3552]
• build(deps): bump golang.org/x/sync from 0.22.0 to 0.23.0 [#​3551]
• build(deps): bump go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc from 0.70.0 to 0.71.0 [#​3427]
• build(deps): bump github.com/olekukonko/errors from 1.2.0 to 1.3.0 [#​3557]
• build(deps): bump github.com/prometheus/client_model from 0.6.2 to 0.6.3 [#​3550]
• build(deps): bump github.com/shamaton/msgpack/v2 from 2.4.1 to 2.4.2 [#​3459]
• build(deps): bump go.opentelemetry.io/otel/exporters/stdout/stdouttrace from 1.45.0 to 1.46.0 [#​3428]
• build(deps): bump github.com/sirupsen/logrus from 1.10.1 to 1.10.2 [#​3492]

v8.0.1

Compare Source

❤️ Thanks to all contributors! ❤️

@​aduffeck, @​maki5, @​pascalwengerter, @​pbleser-oc, @​rhafer, @​zerox80

🔒 Security

• feat(proxy): add optional OIDC access token audience validation [#​3466]

🐛 Bug Fixes

• Remove the timeout when reindexing spaces [#​3543]
• fix(shares): not auto accepting shares created by guest users [#​3533]
• fix(graph): fix PatchMe method to prevent password change [#​3526]

📦️ Dependencies

• build(deps): bump github.com/shamaton/msgpack/v2 from 2.4.1 to 2.4.2 [#​3459]
• build(deps): bump go.opentelemetry.io/otel/exporters/stdout/stdouttrace from 1.45.0 to 1.46.0 [#​3428]
• build(deps): bump github.com/sirupsen/logrus from 1.10.1 to 1.10.2 [#​3492]

v8.0.0

Compare Source

❤️ Thanks to all contributors! ❤️

@​butonic, @​dschmidt, @​fredrikblau, @​fschade, @​maki5, @​pbleser-oc, @​rhafer, @​saw-jan, @​sigurdvaa, @​v-scharf, @​zerox80

💥 Breaking changes

• fix(search): hierarchy tokenizer for bleve path fields [#​3510]
• feat(search): check the index schema on startup and refuse breaking changes [#​3197]
• refactor: reflection-based search mapping + location geopoint [#​3345]
• fix(search): make openSearch and bleve behave the same [#​3408]

✅ Tests

• api-test: retry token refresh on transient IDP failures [#​3507]
• fix(test): removeAccessToSpace test helper [#​3511]
• api-tests: add search retry to all search tests [#​3500]
• test(search): fail the parity suite when the committed matrix is stale [#​3423]
• ci: run search API and e2e suites against OpenSearch [#​3379]
• test(search): engine parity suite for bleve and opensearch [#​3418]

📈 Enhancement

• add insecure to search reindex [#​3505]
• feat(graph): expand thumbnails on driveItems [#​3471]
• graph: expose lockInfo on driveItems [#​3444]
• graph: expose @​libre.graph.shareTypes on driveItems [#​3438]
• feat(search): live photo facet [#​3202]
• feat: support $expand=children on the driveItem endpoint [#​3445]
• feat(search): motion photo facet [#​3200]
• feat(search): video facet [#​3201]
• graph: expose pendingOperations on driveItems [#​3437]
• feat(search): extract more data from tika 4 (if available) [#​3198]
• graph: expose following state, tags and allowed actions on driveItems [#​3113]
• feat(search): scope searches to a drive via the driveId field [#​3424]
• chore(policies): disable gRPC or event handlers by configuration + add metrics [#​3287]
• moved ShareCreated event consumer from frontend to shared service [#​3389]

🐛 Bug Fixes

• fix(search): extract facets from the main tika document only [#​3484]
• fix(thumbnails): bound declared image dimensions before decoding [#​3457]
• test(search): re-search until the expected files are in the result [#​3488]
• fix(config): correct pending version annotations [#​3487]
• Activitylog event handler split [#​3241]
• test(search): wait for expected properties and documents [#​3486]
• fix: log jwt expired on debug level instead of error [#​3463]
• fix(proxy): restrict JWT signed urls to the allowed HTTP methods [#​3481]
• fix: notification handling for share removal and space membership expiry [#​3257]
• fix: posix cli commands [#​3348]

📦️ Dependencies

• [full-ci] chore: bump web to v8.0.0. set introduction version v.8.0.0 [#​3528]
• build(deps): bump go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp from 0.70.0 to 0.71.0 [#​3441]
• build(deps): bump go.opentelemetry.io/contrib/zpages from 0.70.0 to 0.71.0 [#​3442]
• build(deps): bump go.opentelemetry.io/otel from 1.45.0 to 1.46.0 [#​3426]
• build(deps): bump google.golang.org/grpc from 1.83.1 to 1.83.2 [#​3429]
• build(deps): bump github.com/blevesearch/bleve/v2 from 2.6.0 to 2.6.1 [#​3417]
• build(deps): bump github.com/sirupsen/logrus from 1.10.0 to 1.10.1 [#​3409]
• build(deps): bump github.com/opensearch-project/opensearch-go/v4 from 4.6.0 to 4.7.3 [#​3216]
• build(deps): bump github.com/KimMachineGun/automemlimit from 0.7.5 to 1.0.0 [#​3404]
• build(deps): bump github.com/nats-io/nats.go from 1.52.0 to 1.53.1 [#​3403]
• build(deps): bump github.com/open-policy-agent/opa from 1.19.0 to 1.19.1 [#​3402]

---

Configuration

📅 Schedule: (UTC)

• Branch creation
  • At any time (no schedule defined)
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about these updates again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.

[claude]

Risk: MEDIUM — major semver bump (7.5.0 → 8.0.0), single-instance app but this HelmRelease has a documented history of major-version upgrade pain (see the terminationGracePeriodSeconds comment re: the v4→v5 NATS jetstream corruption incident).

Breaking changes: All 4 listed breaking changes are search-engine internals (bleve hierarchy tokenizer, reflection-based search mapping, bleve/opensearch parity, and a new startup schema check that refuses to start on incompatible search indices). This deployment sets no SEARCH_ENGINE/index config, so it uses the default bleve backend against the persisted storage-users/data PVCs.

Per upstream PR #3197: the new startup check auto-applies additive schema changes in place; for a true breaking change it would refuse to start with an explicit rebuild-procedure error (opencloud search index --all-spaces --force-rescan, or bypass via OC_EXCLUDE_RUN_SERVICES=search). Maintainers state indices are versioned (opencloud-resource-v<N>) specifically so normal released-version upgrades like this one shouldn't ever trigger the breaking-change path — it's mainly a dev-time safety net.

No HelmRelease values, envs, or CRDs need renaming — this PR is only the image tag bump.

Action needed: None required. Given this app's prior upgrade incident, recommend watching kubectl logs after rollout for search-index/schema errors on first boot before considering it done — otherwise safe to merge on green.

[claude]

• Risk: MEDIUM — major semver bump (7.5.0 → 8.0.1), but blast radius is a single app (opencloud HelmRelease, one container/PVC set, no CRDs touched).
• Breaking changes: This repo runs OpenCloud with no explicit search-engine config, so it uses the default embedded bleve index — directly hit by the two breaking search changes in 8.0.0:
  • fix(search): hierarchy tokenizer for bleve path fields (#3510)
  • feat(search): check the index schema on startup and refuse breaking changes (#3197) — the new binary can detect the old 7.x bleve schema and refuse to serve search until reindexed.
  • Storage/config PVCs and env vars (OC_*, IDM_*, NATS_*) in helm-release.yaml are unaffected — no values-schema changes apply to this deployment's config surface.
• Action needed: Per upstream's required 7.x→8.0.1 migration step, after this deploys, exec into the pod and run:
  opencloud search index --all-spaces --force-rescan --insecure
  This isn't automated anywhere in this repo (the container command is just opencloud init || true; opencloud server), so search results will be stale/broken (or the service may refuse to serve search) until this is run manually post-rollout. Otherwise safe to merge on green.

Release notes were reachable directly from the PR body (opencloud-eu/opencloud CHANGELOG + v8.0.0/v8.0.1 GitHub release notes), no WebFetch needed.

[claude]

• Risk: MEDIUM (major version 7.5.0 → 8.0.1; single-instance personal file-storage app, no HA/multi-tenant blast radius, but a real breaking change in search)
• Breaking changes: 8.0.0 changes the Bleve search index schema (hierarchy tokenizer fix + reflection-based mapping, #3510, #3345, #3197). This HelmRelease uses the default embedded Bleve search (no SEARCH_ENGINE/OpenSearch config), so it's affected. Per the upgrade guide: the service starts fine and a new versioned index (e.g. bleve-v5) is created automatically alongside the old one — no crash/refused startup. New activity is indexed immediately, but pre-existing, untouched files won't show up in search results until a manual reindex is run. No env var/values renames affect this deployment's OC_*/IDM_* config.
• Action needed: after this rolls out, kubectl exec into the opencloud pod and run opencloud search index --all-spaces --force-rescan --insecure (per 8.0.1's own release note tl;dr) to restore search for existing files, then optionally remove the stale old-version index per the upgrade guide's cleanup step. Not required for the app to function — safe to merge on green, but do the reindex afterward or search will silently miss old content.

[claude]

• Risk: MEDIUM — major bump (7.5.0 → 8.0.1). Single app (personal file storage), but the search subsystem now refuses to start on an incompatible index, so a botched upgrade can crashloop the pod rather than just degrade a feature.
• Breaking changes: 8.0.0 reworks the search index (bleve hierarchy tokenizer fix, reflection-based mapping, "refuse breaking changes" schema check on startup, bleve/OpenSearch parity fixes). This HelmRelease (cluster/apps/opencloud/opencloud/app/helm-release.yaml) doesn't set a SEARCH_ENGINE/OpenSearch config, so it's on the default bleve index built under 7.5.0 — exactly the index format that changed. 8.0.1 is just a bugfix on top (removes a timeout during reindexing), not a further breaking change.
• Action needed: Per upstream's "coming from 7.x" note, after this rolls out you must run, once, against the running pod:
  opencloud search index --all-spaces --force-rescan --insecure
  There's no existing job/hook in this repo doing that automatically, so it needs to be run manually (e.g. kubectl exec into the opencloud pod) right after the new pod comes up — don't assume green health checks alone mean search is usable. No values/CRD changes are needed otherwise.

[claude]

Risk: MEDIUM (major version 7.5.0 → 8.0.1; single-replica app, blast radius limited to this HelmRelease, but startup now validates the search index schema and can refuse breaking changes).

Breaking changes (per upstream release notes, v7.5.0...v8.0.1):

• Search index format changed (hierarchy tokenizer for bleve path fields, reflection-based mapping). This repo doesn't set OC_SEARCH_ENGINE/SEARCH_*, so it uses the default embedded bleve index stored on the NFS data/storage-users PVCs — it will need reindexing, not just a config change.
• OpenCloud now checks the index schema on startup and refuses to run with an incompatible/stale index — expect search to be degraded/unavailable until the reindex is run.
• No values/env renames apply here — IDM_*/OC_* vars and the decomposedfs/messagepack settings in helm-release.yaml are unaffected.

Action needed: After this deploys, exec into the opencloud pod and run the one-time reindex:

kubectl -n opencloud exec deploy/opencloud -- opencloud search index --all-spaces --force-rescan --insecure

(Upstream calls this "the whole migration path from 7.x to 8.0.1.") Until it completes, search results may be incomplete. No CRD or chart-version changes needed. Otherwise safe to merge on green.

[claude]

• Risk: MEDIUM — major bump (7.5.0 → 8.1.0, rolling channel) on a stateful file-storage app (NFS data, embedded NATS/search index).
• Breaking changes (from upstream CHANGELOG, v8.0.0–8.1.0; summarized by a fetch tool, so worth a skim of the raw changelog):
  • Search: the index schema is now checked on startup and breaking changes are refused. The existing bleve index may need opencloud search reindex. If the pod crash-loops with a schema error, that's the cause.
  • FRONTEND_ENABLE_FAVORITES is removed (favorites are always on). This repo doesn't set it.
  • search reindex needs --insecure for HTTPS without cert validation. Not relevant to the current config.
  • 8.0.1 and 8.1.0 add no further breaking changes. 8.1.0 adds optional OIDC access-token audience validation in the proxy, which is off by default.
  • None of the env vars set in helm-release.yaml (OC_*, IDM_*, PROXY_TLS, OC_DECOMPOSEDFS_METADATA_BACKEND, NATS_NATS_STORE_DIR) are listed as renamed or removed.
• Action needed: none in the manifests. Before merging, make sure the NFS data PVC is backed up or snapshotted, because a major upgrade can't be rolled back cleanly. After rollout, check the pod logs for search schema errors and run opencloud search reindex in the pod if needed. Otherwise safe to merge on green.

[claude]

• Risk: MEDIUM. Major bump, 7.5.0 → 8.1.0. OpenCloud is a stateful file-storage app, and the breaking changes are in search indexing.
• Breaking changes (from the 8.0.0 changelog in the PR body; I did not fetch upstream docs):
  • Four search changes are listed as breaking: the bleve hierarchy tokenizer for path fields, a reflection-based index mapping, an index-schema check on startup that refuses breaking changes, and bleve/OpenSearch parity.
  • This repo uses the default embedded bleve index under /var/lib/opencloud on NFS. It has no Tika or OpenSearch config, and the HelmRelease sets no SEARCH_* env vars.
  • The existing index will probably be rejected or stale after upgrade. The pod may crashloop on the schema check, or search may return incomplete results.
  • Nothing in the HelmRelease env needs renaming. The vars it sets (OC_*, IDM_*, PROXY_TLS, NATS_NATS_STORE_DIR, OC_DECOMPOSEDFS_METADATA_BACKEND) are not mentioned in the 8.0.0 or 8.1.0 notes.
  • 8.0.1 and 8.1.0 add an optional OIDC audience check (off by default), proxy fixes, and a Tika key fix. None of these affect this repo.
• Action needed:
  1. Back up or snapshot the NFS storage-users and data PVCs before merging.
  2. After merge, watch the pod logs for a search schema error. If there is one, stop the pod, remove the search index directory (likely /var/lib/opencloud/.../search, please check), and let it rebuild, or run opencloud search index --all-spaces in the pod.
  3. No values or CRD changes are needed. Otherwise safe to merge once green and a backup exists.

[claude]

• Risk: MEDIUM. Major bump (7.5.0 → 8.1.0) of opencloudeu/opencloud-rolling, and this is the file store with the user data. The change is a single image tag in cluster/apps/opencloud/opencloud/app/helm-release.yaml.
• Breaking changes (from upstream CHANGELOG.md, 8.0.0 "💥 Breaking changes"). All four are search/index changes:
  • #3510: the bleve path-field tokenizer changed.
  • #3197: the index schema is now checked at startup, and the service refuses breaking schema changes. Startup could fail or the search service could refuse the existing index.
  • #3345: search mapping was refactored, with a new location geopoint.
  • #3408: openSearch and bleve now behave the same.
  • The changelog gives no migration steps for any of these.
  • 8.0.1 and 8.1.0 list no breaking changes. They add optional OIDC audience validation on the proxy, which should be off by default, plus WOPI token hardening and bug fixes.
  • This repo sets none of the changed settings: no SEARCH_*, PROXY_OIDC_* or collaboration/WOPI env vars. It uses IDM_*, OC_*, PROXY_TLS, OC_DECOMPOSEDFS_METADATA_BACKEND and NATS_NATS_STORE_DIR, and none of these appear in the 8.x breaking list. The search index lives on the NFS volume at /var/lib/opencloud, so it will be hit by the schema check.
• Action needed:
  • Back up or snapshot the NFS PVCs nfs-opencloud-storage-users-pvc and nfs-opencloud-data-pvc before merging.
  • After merge, watch the pod logs. If the search service rejects the index schema, stop the pod and delete the search index directory under /var/lib/opencloud (the data/ subPath), then let it reindex. Verify the path in the logs first.
  • There are no values renames or CRD changes. Otherwise it is safe to merge on green.
• Caveat: only the first 100k of 115k characters of the changelog were read, which covers 8.x. Upstream's full release notes may have more detail.