add: Redis CCR feature #76
Conversation
Deploying aiven-docs with
|
| Latest commit: |
391c499
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://a93a2b3c.aiven-docs.pages.dev |
| Branch Preview URL: | https://harshini-redis-ccr.aiven-docs.pages.dev |
| normal conditions. | ||
|
|
||
| **Handling failures:** In case of any failure of the primary service, Aiven provides | ||
| automatic failover functionality, which promotes the replica service to become |
There was a problem hiding this comment.
it's not automatic afaik. I never implemented this feature and I don't know if it was a pre-existing feature of the integration type.
Manual promotion of the replica service does work, though.
There was a problem hiding this comment.
If the failover process is manual, could you provide more details and steps on how users can promote a replica to primary during a primary cluster failure?
There was a problem hiding this comment.
they need to delete the service integration either via REST calls or with the aiven client (which does the same calls under the hood).
| ## Limitations | ||
|
|
||
| - This feature requires `[Pro Platform](/docs/platform/concepts/pro-platform)`. | ||
| - Aiven for Redis CCR follows an active-passive model, which means that only the data |
There was a problem hiding this comment.
it's actually not possible to write data in the replica service at all while the integration is active.
There was a problem hiding this comment.
Just to confirm, when you say “it’s not possible to write data in the replica service during active integration”, are you suggesting that in Aiven for Redis CCR’s active-passive model, all new data is added to the primary service first and then replicated to the replica service? Could you please confirm or provide more details if necessary?
There was a problem hiding this comment.
how about rewording it to this:
Aiven for Redis CCR uses an active-passive model where data can be written only to the primary service and then replicated to the replica service.
There was a problem hiding this comment.
Yes. And the rewrite looks good.
| ingested in the primary service gets replicated. | ||
| - Aiven for Redis services only support replication within their own service type. | ||
| Cross-service replication to or from Aiven for Dragonfly services is not supported. | ||
| - Enabling up CCR for Aiven for Redis is only possible via API/CLI, with console support |
There was a problem hiding this comment.
Console support is a one day job. I actually implemented it myself just for demo purposes. I assumed when we launch this, we would also have console support, or?
There was a problem hiding this comment.
If the feature support is available on the Aiven console at the time of release, I can include it in the documentation. However, as per the information provided by @hoang-vo, the console updates were expected to be released in the next quarter or so.
There was a problem hiding this comment.
Oh ok, we can confirm with Hoang once he's back.
| following API call: | ||
|
|
||
| ```shell | ||
| curl -X POST https://api.aiven.io/v1/project/PROJECT_NAME/service \ |
There was a problem hiding this comment.
commands looks good but didn't run them myself now. I assume you did and they work?
There was a problem hiding this comment.
I didn't either; I based it on the examples in other services (cassandra, for example).
|
|
||
| ## Manage failover | ||
|
|
||
| If the primary service fails, Aiven automatically promotes the replica service to the |
There was a problem hiding this comment.
same point as above about automatic failover. To manually promote the replica to master, you'd have to delete the integration.
Yes, this is part of the pro platform, which is now called Aiven Enterprise. Keep the docs ready so when the feature is merged, we can publish the docs alongside. |
harshini-rangaswamy
force-pushed
the
harshini-redis-ccr
branch
from
e2490a8 to
ebfe8ee
Compare
Describe your changes
Content for the Redis cross-cluster replication (CCR) feature as part of the Pro Platform.
DOC-853
Checklist
/docs/.