

























Ceph Reef (18.x) marked several features as deprecated in preparation for Squid and future releases. Understanding these deprecations helps you plan migrations before forced removal breaks workloads.
Ceph surfaces deprecation warnings in the health output:
kubectl -n rook-ceph exec -it deploy/rook-ceph-tools -- ceph health detail | grep -i deprecatReview the Ceph log for deprecation notices:
kubectl -n rook-ceph logs deploy/rook-ceph-mon-a | grep -i deprecatedThe old MDS affinity configuration using mds_standby_for_name is deprecated:
# Check if old affinity settings are in use
kubectl -n rook-ceph exec -it deploy/rook-ceph-tools -- ceph config dump | grep mds_standbyMigrate to the new fs set standby_count_wanted approach:
kubectl -n rook-ceph exec -it deploy/rook-ceph-tools -- ceph fs set myfs standby_count_wanted 1The old bootstrap token format for RBD mirroring is deprecated. Generate a new token:
kubectl -n rook-ceph exec -it deploy/rook-ceph-tools -- rbd mirror pool peer bootstrap create \
--site-name primary mypoolThe FileStore OSD backend is deprecated in Reef. All OSDs should use BlueStore:
# Check OSD backend type
kubectl -n rook-ceph exec -it deploy/rook-ceph-tools -- ceph osd metadata | jq '.[].osd_objectstore' | sort | uniq -cIf any OSDs still report filestore, plan a migration to bluestore by reprovisioning those OSDs.
Generate a full config dump to spot deprecated options:
kubectl -n rook-ceph exec -it deploy/rook-ceph-tools -- ceph config dump > config-dump.txt
grep -E "(deprecated|legacy|old)" config-dump.txtRemove deprecated config keys proactively:
kubectl -n rook-ceph exec -it deploy/rook-ceph-tools -- ceph config rm global old_option_nameCheck your CephCluster and related CRDs for deprecated fields in the Rook API:
kubectl get cephcluster rook-ceph -n rook-ceph -o yaml | grep -i deprecatedCompare against the Rook v1.14+ CRD reference to identify removed fields.
Ceph Reef deprecates several legacy features including old MDS standby settings, bootstrap token formats, and the FileStore OSD backend. Use ceph health detail and ceph config dump to audit your cluster, and proactively migrate deprecated configurations before upgrading to Squid to prevent service disruptions.
此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。