docs(redis): add Sentinel RDB snapshot cross-cluster migration guide

[chideat]

Adds docs/en/solutions/ecosystem/redis/How_to_Migrate_Redis_Sentinel_With_RDB_Snapshot.md.

What

Documents the offline migration path for a Sentinel-mode Redis instance moving to a new Kubernetes cluster with a different StorageClass, carrying the dataset across as /data/dump.rdb:

• Freeze writes from the client application, record the source keyspace baseline
• BGSAVE, confirm rdb_bgsave_in_progress:0 / rdb_last_bgsave_status:ok, export and checksum
• Create the target with master: 1 / slave: 0 on the RDB parameter template, import, verify the SHA-256 inside the pod
• Force-delete the pod to skip PreStop so the imported snapshot is what gets loaded
• Verify the data, then restore the replica count and switch the application over
• Rollback boundaries: safe until the target accepts its first write

Complements the existing How_to_Migrate_Redis_Across_Clusters.md (RedisShake, continuous replication), which explicitly defers to an RDB export/import path for cases it does not support. Both docs now cross-link.

Notes for review

• ProductsVersion is set to 4.x — the procedure requires redis-operator 4.0+. Existing Redis docs here reference operator 3.10–3.18 and tag 3.x/4.x; correct this if the field tracks platform version independently.
• The doc warns against the AOF parameter template and links to Backup_Restore_Template_Compatibility.md, since with appendonly yes Redis ignores dump.rdb and starts empty — the most likely silent failure for this procedure.
• The stated reason for the forced pod deletion (a graceful shutdown persists the in-memory dataset over the imported file) is inferred from standard Redis shutdown behavior, not verified against the operator's actual PreStop hook. Worth a confirmation.
• No id: in the frontmatter — add_id.sh assigns the KB number in CI. Chinese translation is generated by the translate job on merge.

doom build passes locally.

🤖 Generated with Claude Code