-
Notifications
You must be signed in to change notification settings - Fork 458
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Replace the "geo-partitioned replicas" topology pattern with its high-level multi-region equivalent #9265
Milestone
Comments
This was referenced Feb 3, 2021
rmloveland
added a commit
that referenced
this issue
Apr 29, 2021
Fixes #9268 Fixes #9266 Fixes #9265 Fixes #9264 Fixes #10425 Fixes #10319 Addresses #10051 Addresses #10401 Addresses #10461 Addresses #10463 Summary of changes: - Update multi-region "topology patterns" as follows: - Deemphasize "topology" for these patterns, since it really isn't - it's per-table - Add "global tables" as a pattern - Add "regional tables" as a pattern - Remove "duplicate indexes" and "geo-partitioned foos" pages and have them redirect to the above, as appropriate - Fix up ~all links in the docs to point to the new things - Update mentions of partitioning throughout the docs to note that most users should use new v21.1+ multi-region capabilities instead of explicit partitioning - Also removed mention of partitioning from several places as part of deemphasing explicit partitioning generally, since most use cases are covered by MR abstractions - Update `cockroach demo` to deemphasize partitioning and point to new MR things - Update disaster recovery pages to mention most users of new deployments should just use multi-region survival goals - Update various licensing docs to point to the new multi-region things - Update replication reports to mention the new MR things, in lieu of further updates - Renamed a bunch of links that used the phrase "geo-partitioning" to refer to the multi-region latency tutorial, which now goes by the name "Multi-Region Performance" - Note: we explicitly did *not* touch anything related to the multi-region Flask app, since that work is happening via #10394
rmloveland
added a commit
that referenced
this issue
Apr 29, 2021
Fixes #9268 Fixes #9266 Fixes #9265 Fixes #9264 Fixes #10425 Fixes #10319 Addresses #10051 Addresses #10401 Addresses #10461 Addresses #10463 Summary of changes: - Update multi-region "topology patterns" as follows: - Deemphasize "topology" for these patterns, since it really isn't - it's per-table - Add "global tables" as a pattern - Add "regional tables" as a pattern - Remove "duplicate indexes" and "geo-partitioned foos" pages and have them redirect to the above, as appropriate - Fix up ~all links in the docs to point to the new things - Update mentions of partitioning throughout the docs to note that most users should use new v21.1+ multi-region capabilities instead of explicit partitioning - Also removed mention of partitioning from several places as part of deemphasing explicit partitioning generally, since most use cases are covered by MR abstractions - Update `cockroach demo` to deemphasize partitioning and point to new MR things - Update disaster recovery pages to mention most users of new deployments should just use multi-region survival goals - Update various licensing docs to point to the new multi-region things - Update replication reports to mention the new MR things, in lieu of further updates - Renamed a bunch of links that used the phrase "geo-partitioning" to refer to the multi-region latency tutorial, which now goes by the name "Multi-Region Performance" - Note: we explicitly did *not* touch anything related to the multi-region Flask app, since that work is happening via #10394
rmloveland
added a commit
that referenced
this issue
May 3, 2021
Fixes #9268 Fixes #9266 Fixes #9265 Fixes #9264 Fixes #10425 Fixes #10319 Addresses #10051 Addresses #10401 Addresses #10461 Addresses #10463 Summary of changes: - Update multi-region "topology patterns" as follows: - Deemphasize "topology" for these patterns, since it really isn't - it's per-table - Add "global tables" as a pattern - Add "regional tables" as a pattern - Remove "duplicate indexes" and "geo-partitioned foos" pages and have them redirect to the above, as appropriate - Fix up ~all links in the docs to point to the new things - Update mentions of partitioning throughout the docs to note that most users should use new v21.1+ multi-region capabilities instead of explicit partitioning - Also removed mention of partitioning from several places as part of deemphasing explicit partitioning generally, since most use cases are covered by MR abstractions - Update `cockroach demo` to deemphasize partitioning and point to new MR things - Update disaster recovery pages to mention most users of new deployments should just use multi-region survival goals - Update various licensing docs to point to the new multi-region things - Update replication reports to mention the new MR things, in lieu of further updates - Renamed a bunch of links that used the phrase "geo-partitioning" to refer to the multi-region latency tutorial, which now goes by the name "Multi-Region Performance" - Note: we explicitly did *not* touch anything related to the multi-region Flask app, since that work is happening via #10394
rmloveland
added a commit
that referenced
this issue
May 5, 2021
Fixes #9268 Fixes #9266 Fixes #9265 Fixes #9264 Fixes #10425 Fixes #10319 Addresses #10051 Addresses #10401 Addresses #10461 Addresses #10463 Summary of changes: - Update "topology patterns" docs as follows: - Move them as a sub-section under "Multi-Region Capabilities", since they are mostly relevant to multi-region usage - Deemphasize "topology" for these patterns, since these are not network topologies, they are mostly per-table configurable behaviors - Add "global tables" to replace the "duplicate indexes" page - Add "regional tables" to replace the "geo-partitioned XYZ" pages - Fix up ~all links in the docs to point to the new things - Update mentions of partitioning throughout the docs to note that most users should use the new v21.1+ multi-region capabilities instead of explicit partitioning - Also removed mention of partitioning from several places as part of deemphasing explicit partitioning generally, since most use cases are covered by multi-region abstractions - Update the 'Cost-Based Optimizer' page to remove the old skool "nearest index" partitioning stuff, since now we have GLOBAL tables - This required removing links to the above from all the "enterprise features" lists and replacing with links to the new 'Multi-Region Capabilities' page - Update `cockroach demo` to deemphasize partitioning and point to new multi-region things - Update disaster recovery pages to mention most users of new deployments should just use multi-region survival goals - Update various licensing docs to point to the new multi-region things - Update replication reports to mention the new multi-region things, in lieu of further updates - Renamed a bunch of links that used the phrase "geo-partitioning" to refer to the multi-region latency tutorial, which now goes by the name "Multi-Region Performance" - Note: we explicitly did *not* touch anything related to the multi-region Flask app, since that work is happening via #10394
rmloveland
added a commit
that referenced
this issue
May 7, 2021
Fixes #9268 Fixes #9266 Fixes #9265 Fixes #9264 Fixes #10425 Fixes #10319 Addresses #10051 Addresses #10401 Addresses #10461 Addresses #10463 Summary of changes: - Update "topology patterns" docs as follows: - Move them as a sub-section under "Multi-Region Capabilities", since they are mostly relevant to multi-region usage - Deemphasize "topology" for these patterns, since these are not network topologies, they are mostly per-table configurable behaviors - Add "global tables" to replace the "duplicate indexes" page - Add "regional tables" to replace the "geo-partitioned XYZ" pages - Fix up ~all links in the docs to point to the new things - Update mentions of partitioning throughout the docs to note that most users should use the new v21.1+ multi-region capabilities instead of explicit partitioning - Also removed mention of partitioning from several places as part of deemphasing explicit partitioning generally, since most use cases are covered by multi-region abstractions - Update the 'Cost-Based Optimizer' page to remove the old skool "nearest index" partitioning stuff, since now we have GLOBAL tables - This required removing links to the above from all the "enterprise features" lists and replacing with links to the new 'Multi-Region Capabilities' page - Update `cockroach demo` to deemphasize partitioning and point to new multi-region things - Update disaster recovery pages to mention most users of new deployments should just use multi-region survival goals - Update various licensing docs to point to the new multi-region things - Update replication reports to mention the new multi-region things, in lieu of further updates - Renamed a bunch of links that used the phrase "geo-partitioning" to refer to the multi-region latency tutorial, which now goes by the name "Multi-Region Performance" - Updated 'ALTER TABLE ... SET LOCALITY REGIONAL BY ROW' to include info about locality optimized search - Note: we explicitly did *not* touch anything related to the multi-region Flask app, since that work is happening via #10394
rmloveland
added a commit
that referenced
this issue
May 10, 2021
Fixes #9268 Fixes #9266 Fixes #9265 Fixes #9264 Fixes #10425 Fixes #10319 Addresses #10051 Addresses #10401 Addresses #10461 Addresses #10463 Summary of changes: - Update "topology patterns" docs as follows: - Move them as a sub-section under "Multi-Region Capabilities", since they are mostly relevant to multi-region usage - Deemphasize "topology" for these patterns, since these are not network topologies, they are mostly per-table configurable behaviors - Add "global tables" to replace the "duplicate indexes" page - Add "regional tables" to replace the "geo-partitioned XYZ" pages - Fix up ~all links in the docs to point to the new things - Update mentions of partitioning throughout the docs to note that most users should use the new v21.1+ multi-region capabilities instead of explicit partitioning - Also removed mention of partitioning from several places as part of deemphasing explicit partitioning generally, since most use cases are covered by multi-region abstractions - Update the 'Cost-Based Optimizer' page to remove the old skool "nearest index" partitioning stuff, since now we have GLOBAL tables - This required removing links to the above from all the "enterprise features" lists and replacing with links to the new 'Multi-Region Capabilities' page - Update `cockroach demo` to deemphasize partitioning and point to new multi-region things - Update disaster recovery pages to mention most users of new deployments should just use multi-region survival goals - Update various licensing docs to point to the new multi-region things - Update replication reports to mention the new multi-region things, in lieu of further updates - Renamed a bunch of links that used the phrase "geo-partitioning" to refer to the multi-region latency tutorial, which now goes by the name "Multi-Region Performance" - Updated 'ALTER TABLE ... SET LOCALITY REGIONAL BY ROW' to include info about locality optimized search - Note: we explicitly did *not* touch anything related to the multi-region Flask app, since that work is happening via #10394
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The work here is to document how users can achieve the outcomes provided by the geo-partitioned replicas topology pattern using the new multi-region abstractions.
Characteristics:
SQL statements to enable:
ALTER TABLE .. REGIONAL BY ROW
(*)ALTER DATABASE .. SURVIVE AVAILABILITY ZONE FAILURE
(*) Tables altered in this way will have a new region enum added to them, via a schema change.
The text was updated successfully, but these errors were encountered: