Shipped · DevOps and infrastructure

What rook/rook shipped

The public repository of Rook · github.com/rook/rook

Written by FoxPlug from the repository's own releases; not affiliated with Rook. Paste your own repository in the box below and FoxPlug writes the same weekly update for your product. This page is an automatic summary of the releases, pull requests and commits in github.com/rook/rook; Rook did not write it and does not use or endorse FoxPlug. Every line links to the change it describes.

Get a weekly update like this for your product, free

Or Use it as a GitHub Action

See the update FoxPlug writes for rook, free

Week of November 2, 2026

What shipped

Why it matters

This week brings two patch releases with operator fixes plus tighter input validation on volume mounts and clearer duplicate OSD reporting. Several changes keep credentials out of logs and errors, and the default msgr2 behavior is enforced on more daemons.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

Rook v1.21.2 and v1.20.10 patch releases are out. Volume mount subPaths are now validated, duplicate OSD IDs are reported, RGW account quotas are supported, and msgr1 is disabled on OSD, MGR and MDS when msgr2 is required.

Post for LinkedIn

Rook v1.21.2 and v1.20.10 patch releases are now available. This week also added validation that keeps additional volume mounts inside their root directory, so a crafted subPath cannot mount a volume over arbitrary container paths. OSD prepare now reports an error when a duplicate OSD ID is found instead of skipping creation silently. RGW account quotas and resource limits can be configured through CephObjectStoreAccount, and the realm system user is created when a zone defines custom endpoints. When msgr2 is required, OSD, MGR and MDS daemons now disable msgr1. LUKS key rotation job failures now raise an alert. CodeQL scanning covers Go, Python and GitHub Actions code on PRs and master.

Week of October 19, 2026

What shipped

Why it matters

This week fixed several operator behaviors that caused repeated reconciles or broken reconnects, such as the mgr Prometheus interval comparison and the NVMe-oF Service ClusterIP. It also added configuration that users had to work around by hand, including scrape timeouts, secret templates for object store users, and a nofile limit annotation. Upgrade notes for v1.21 were added because the new StorageClass parameters make helm upgrade fail without first deleting them.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

This week: configurable Prometheus scrape timeout, secretTemplate for CephObjectStoreUser, stable NVMe-oF gateway ClusterIP, numeric interval comparison, deterministic RGW flags, and v1.21 upgrade guidance.

Post for LinkedIn

Changes merged this week in Rook: spec.monitoring.scrapeTimeoutSeconds for the mgr and ceph ServiceMonitors, secretTemplate labels and annotations on CephObjectStoreUser, and preservation of the NVMe-oF gateway Service so its ClusterIP stays stable across scale-down. The mgr Prometheus module no longer restarts every reconcile because the scrape interval is now compared numerically, and RGW command flags are emitted in a deterministic order. NVMe-oF gained an interrupt mode option to reduce idle CPU, and object stores gained an annotation to raise the radosgw nofile soft limit. Minimum failure domain detection now handles region correctly, and the v1.20 to v1.21 guide explains deleting StorageClasses before helm upgrade.

Week of October 12, 2026

What shipped

Why it matters

This week brings two patch releases, v1.20.9 and v1.19.12, plus fixes across object, OSD, and NFS reconciliation. Several changes target problems operators hit in practice: wrong dashboard credentials per realm, OSDs landing on unsuitable devices, stale noout flags from numeric topology names, and leftover config files in the operator pod.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

Rook v1.20.9 and v1.19.12 are out. Also this week: RGW service trafficDistribution, dashboard credential fix for object stores, device discovery skips nbd/drbd/zram, numeric topology validation, and NFS temp file cleanup.

Post for LinkedIn

Rook v1.20.9 and v1.19.12 are out this week. v1.20.9 enforces secure msgr2 mode for mon connections and rejects purely numeric topology values in validation. v1.19.12 allows missing lifecycle notification events and waits for the OSD upgrade before reconciling. Also merged: RGW services can set trafficDistribution through gateway.service.trafficDistribution; object stores ignore the dashboard's per-realm credential map when RGW keys are empty; OSD updates return early when all OSDs in a batch are skipped; device discovery now skips nbd, drbd, and zram devices; topology validation rejects numeric bucket names that Ceph parses as OSD IDs; the NFS operator removes its temporary ganesha config files; and external NVMe-oF client access documentation was added.

Week of October 5, 2026

What shipped

Why it matters

This week brings the v1.21.0 release and several changes that affect how clusters are built and secured. Secure msgr2 handling, per-device encryption and NFS pool creation now behave as expected, and CSI setup is left more to the csi-operator. Users upgrading should note the msgrv2 default and the Helm chart tag change.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

Rook v1.21.0 is out: msgrv2 required by default and Helm OCI chart tags drop the v prefix. Also this week: mon msgr2 modes enforced, per-device encryptedDevice respected, NFS no longer auto-creates the.nfs pool, and CSI defaults move to the csi-operator.

Post for LinkedIn

Rook v1.21.0 is released. Breaking changes to note: msgrv2 is required by default, and Helm OCI chart tags no longer include the v prefix, so update any scripts that reference the chart by tag. Other changes this week: mon-specific msgr2 mode options are now set to secure when connection encryption is enabled. Per-device encryptedDevice: "true" is respected when creating OSDs. The NFS controller no longer creates the.nfs pool automatically and instead asks you to create a CephBlockPool. Rook no longer manages CSI defaults and the imageset ConfigMap, leaving those to the csi-operator. A new ROOK_CREATE_CSI_OPERATOR_RESOURCES setting lets object-storage-only clusters skip CSI CRs. The csi-driver and ceph-csi-operator were updated, and the NVMe-oF gateway CPU alert now compares against the co

Week of September 28, 2026

What shipped

Why it matters

Rook v1.20.8 addresses critical OSD networking and monitoring issues that affect production clusters. The fixes for prepare pod host networking and rbd device filtering resolve deadlock scenarios, while the monitoring alert threshold correction prevents false alerts on growing pools.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

Rook v1.20.8 released with fixes for OSD prepare pod networking on host networks and rbd device filtering in raw-mode activation. Also included: pool growth alert threshold correction and toolbox pod configuration improvements.

Post for LinkedIn

Rook v1.20.8 is now available with important stability improvements for OSD operations. This patch release includes fixes for OSD prepare pods on host networks, rbd device handling in raw-mode activation, pool growth alert threshold corrections, and enhanced toolbox configuration options via Helm. These updates address production scenarios and improve monitoring accuracy.

Week of September 21, 2026

What shipped

Why it matters

This week includes important protocol changes for cluster communication with msgr2 becoming the default, security fixes preventing credential leakage in logs, and infrastructure improvements for better build reliability and CI behavior. These changes improve both security and operational stability for Rook deployments.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

Rook ships msgr2 as the default cluster protocol, security fixes for credential handling, improved CSI sidecars, and better build system support for sandboxed environments.

Post for LinkedIn

This week's Rook updates include significant security improvements that prevent S3 keys and Ceph secrets from leaking into logs and Kubernetes events, plus important protocol changes defaulting to msgr2 for cluster communication. We've also shipped CSI-Addons v0.15.1, fixed build system issues in sandboxed environments, and improved the retry logic to respect context cancellation.

Week of September 14, 2026

What shipped

Why it matters

V1.21.0-beta.0 arrives with improvements to object storage configuration immutability, better visibility into buckets via kubectl output, and TLS support for metrics collection. These changes address usability and security concerns for operators managing Rook clusters.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

Rook v1.21.0-beta.0 is out. New features include immutable RGW pool references, object bucket printer columns, and TLS support for Prometheus metrics.

Post for LinkedIn

Rook v1.21.0-beta.0 is now available. This release brings several improvements: CephObjectStoreUser gains placement and storage class configuration, RGW zone pools are now immutable to prevent accidental changes, Prometheus metrics support TLS, and object bucket CRDs display better information with kubectl. Additional lifecycle event support and security improvements included.

Weeks with too little public activity are left out rather than filled in. Last updated 2026-11-10.

Is this your repo? Ask us to remove this page.