Ceph erasure coding design
Choosing k and m for Ceph erasure-coded pools: capacity overhead versus replication, failure domains, overwrites for RBD and CephFS, the FastEC changes in Tentacle, and the recovery performance cost.
Ceph features
Pages on specific Ceph storage features that buyers plan and price separately: erasure coding, replication and multi-site designs, block and file gateways, encryption and immutability.
Ceph RBD mirroring for disaster recovery
Replicating block images to a second Ceph cluster with rbd-mirror: journal versus snapshot mode, one-way or two-way setups, failover and failback steps, and the orchestration RBD leaves to you.
Ceph stretch mode for two data centers
Running one Ceph cluster across two sites with a tiebreaker monitor: size 4 pools, degraded and recovery stretch modes after a site loss, monitor placement, and the limits on erasure coding and device classes.
Ceph NVMe over TCP gateway for VMware and Linux
Exporting RBD images over NVMe/TCP as SAN-style block storage for VMware datastores and Linux hosts: gateway groups, active/standby failover, subsystem limits and network design.
Ceph SMB file shares
Serving Windows and macOS clients from CephFS with the smb manager module: clustered Samba deployment through cephadm, Active Directory membership versus local users, SMB2 and SMB3 only, and per-share settings.
CephFS shared file system and NFS exports
Using CephFS for shared home directories, HPC scratch and application storage: metadata server sizing, native kernel and FUSE clients, and NFS-Ganesha exports for clients that cannot run Ceph clients.
Ceph encryption at rest with encrypted OSDs
Encrypting Ceph data at rest with LUKS-based encrypted OSDs from cephadm specs, TPM2 enrollment from Squid onward, operational caveats for repair tools, and how this pairs with msgr2 encryption in transit.
Ceph S3 Object Lock for immutable storage
Using Object Lock on Ceph RGW buckets to keep backups and records immutable for a retention period, bucket versioning, default retention rules and how it supports ransomware-resistant backup targets.
Scaling CephFS metadata with multiple active MDS daemons
When one metadata server is the bottleneck: setting max_mds, keeping standbys for failover, static and ephemeral directory pinning for home directories and multi-tenant trees, and the dynamic balancer.
CephFS subvolumes for Kubernetes CSI and OpenStack Manila
How the volumes manager module provisions shares: volumes, subvolume groups and subvolumes, quotas as size limits, snapshots and asynchronous clones, and how Manila and Ceph-CSI use them.
CephFS snapshot schedules and retention
Automating CephFS snapshots with the snap_schedule module: intervals, retention specs such as hourly and weekly counts, the 100 snapshots per directory limit and the module's 50 snapshot cap.
CephFS snapshot mirroring to a second cluster
Replicating CephFS directories to a remote cluster for disaster recovery: the cephfs-mirror daemon, peer bootstrap, per-directory mirroring, user capabilities and what asynchronous snapshot replication means for recovery points.
