Bring physical cluster replication into GA v24.1 #18538
Conversation
✅ Deploy Preview for cockroachdb-interactivetutorials-docs canceled.
|
✅ Deploy Preview for cockroachdb-api-docs canceled.
|
✅ Netlify Preview
To edit notification comments on pull requests, go to your Netlify site configuration. |
kathancox
force-pushed
the
ga-pcr-updates
branch
3 times, most recently
from
7f867fb
to
ac1e315
Compare
| {{site.data.alerts.end}} | ||
|
|
||
| ## Known limitations | ||
|
|
||
| {% include {{ page.version.version }}/known-limitations/physical-cluster-replication.md %} | ||
|
|
||
| ## Performance |
There was a problem hiding this comment.
@dt Can you check that these performance numbers look OK?
| Cockroach Labs supports PCR up to the following scale: | ||
|
|
||
| - Cluster size: 30TB | ||
| - Nodes: 7 |
There was a problem hiding this comment.
@alicia-l2 I think we can say we support far more than 7 nodes? anything up to ~30 seems like it should behave roughly as expected.
There was a problem hiding this comment.
I'm not sure if we need to mention the nodes at all?
|
|
||
| - Cluster size: 30TB | ||
| - Nodes: 7 | ||
| - Transaction payload: [TPC-C]({% link {{ page.version.version }}/cockroach-workload.md %}#tpcc-workload) |
There was a problem hiding this comment.
@alicia-l2 Should we break this list up into "Cockroach Labs Supports: ..." which just lists cluster size metrics (TB, nodes, WPS) and then "Our testing with the TPC-C has demonstrated that:..." ? Because we support far more than just the tpcc workload; we support any sql workload you want, as long as it doesn't include large schema changes or imports. We just only verified performance with the tpcc workload.
There was a problem hiding this comment.
Yeah agreed, except I think we should take nodes out and keep RPS. We could say "Our testing with TPC-C has demonstrated that we support workloads up to the following scale:"
| - Writes: 10,000 writes per second | ||
| - Reads: 18,000 reads per second | ||
| - [Replication lag]({% link {{ page.version.version }}/physical-cluster-replication-technical-overview.md %}) (steady state, no bulk changes): 20–45 seconds | ||
| - [Replication lag]({% link {{ page.version.version }}/physical-cluster-replication-technical-overview.md %}) (bulk changes / schema changes): Infinite |
There was a problem hiding this comment.
Can we remove this and add a note at the bottom that using PCR with frequent large schema changes or imports may cause significant spike in replication lag?
| @@ -11,15 +11,15 @@ docs_area: deploy | |||
|
|
|||
| {% include_cached new-in.html version="v23.2" %} This page gives an overview of _cluster virtualization_ in CockroachDB {{page.version.version}}. Cluster virtualization allows you to separate a cluster's _control plane_ from its _data plane_. The cluster's control plane manages cluster nodes and node-to-node traffic, while its data plane reads data from and writes data to the cluster's storage. | |||
|
|
|||
| Cluster virtualization is required for [Physical Cluster Replication]({% link {{ page.version.version }}/physical-cluster-replication-overview.md %}). Creating virtual clusters without the intent of using them as either a physical cluster replication source or target is not yet supported. | |||
| Cluster virtualization is required for [**Physical cluster replication (PCR)**]({% link {{ page.version.version }}/physical-cluster-replication-overview.md %}). Creating virtual clusters without the intent of using them as either a PCR primary or standby is not yet supported. | |||
There was a problem hiding this comment.
Technically it is supported but just in preview?
There was a problem hiding this comment.
True, it is supported in the the product, but do we intend to support users who create multiple virtual clusters? I think last release we wanted to avoid users creating multiple virtual clusters other than for PCR. Happy to remove the sentence though, if we don't think it is helpful.
There was a problem hiding this comment.
I have updated the language around this.
| @@ -66,7 +66,7 @@ cockroach sql --url \ | |||
|
|
|||
| This section shows how to connect using the [DB Console]({% link {{ page.version.version }}/ui-overview.md %}) when cluster virtualization is enabled. | |||
|
|
|||
| Unless you specify which virtual cluster to connect to, when you connect using the DB Console, you are logged into the default virtual cluster. When [Physical Cluster Replication]({% link {{ page.version.version }}/physical-cluster-replication-overview.md %}) is enabled, the default virtual cluster is named `application`. | |||
| Unless you specify which virtual cluster to connect to, when you connect using the DB Console, you are logged into the default virtual cluster. When [PCR]({% link {{ page.version.version }}/physical-cluster-replication-overview.md %}) is enabled, the default virtual cluster is named `application`. | |||
There was a problem hiding this comment.
It might be worth specifying that the user has to connect to the standby cluster when trying to see replication lag
There was a problem hiding this comment.
I have added this to v24.1 was not sure if the replication lag db console metric was backported.
Fixes DOC-10255
This PR moves physical cluster replication into GA:
VIRTUAL CLUSTERSQL pages) (v24.1)