John and I presented the status of Replication in OpenStack Cinder at the Tokyo Summit. This presentation is current for the Liberty release and has some previews of what will be in Mitaka.
FULL ENJOY 🔝 8264348440 🔝 Call Girls in Diplomatic Enclave | Delhi
Cinder - status of replication
1. The status of Replication
John Griffith - john.griffith@solidfire.com
Ed Balduf - ed.balduf@solidfire.com
2. Why are we building replication into Cinder?
▪ What’s it for? What does it buy me?
▪ It’s a DR strategy for non-cloud applications.
▪ Backup (Kind of).
▪ Limited use cases today.
▪ Different vendors
▪ Different terminology for states.
▪ How does failover/failback work?
▪ Crawl-walk-run
▪ Should be aware of Availability Groups (future).
3. Use Cases Considered
▪ NEVER - 2 different vendor backends.
▪ Consider:
▪ 2 backends, same cloud
▪ 2 backends, different cloud
▪ 2 backends, one not in a cloud
▪ Replication one to multiple backends
▪ Automated -vs- Non-Automated failover
▪ Snapshot replication -vs - Continuous replication
4. How we got here
Missteps
▪ Many voices
▪ Lack of agreed upon vision
▪ Difference in interpretation
▪ Unique characteristics that don’t translate well
▪ Lack of testability
▪ Rush at the end of release cycle
5. V1 lessons learned
▪ Lack of community involvement
▪ Only worked for one vendor
▪ Not clearly understood
▪ Documentation was not up to standards
6. V2 Learning from our mistakes
▪ Heavy involvement from multiple vendors
▪ Reviews, reviews, reviews
▪ DON’T RUSH!!!
▪ We will sell no wine, before it’s time
7. How it works and what it does:
Driver must report:
replication_enabled = True
In it’s capabilities.
New config variable:
replication_device = {
device_target_id: <required>,
managed_backend_name: <host@backend_name>,
Vendor Key1: <Vendor Value>,
Vendor Key2: <Vendor Value>,
}
(Note: No trailing comma allowed in replication_device KV pair list)