You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
TLDR - I'd like to have documentation covering Citus cluster migration scenarios
Our background (the non-TLDR part) is covered in Citus Slack
To successfully migrate a complete Citus cluster the needed steps are somewhat clear:
deploy a new cluster
copy the data over and keep it in sync
switch over the clients to the new cluster
I believe it's possible to replicate a (current) Citus cluster into empty new Citus cluster (with same number of workers) using PUBLICATION + SUBSCRIPTION across each old and new workers, and across the old and new coordinators. By reading the documentation about CDC the idea is clear - just create pubs+subs not for ALL TABLES, but for the non-metadata user tables. But as we're replicating each worker and coordinators separately it's unclear if citus.enable_change_data_capture brings anything on the table (as we want to each target worker / shard to get 1:1 changes from the source)
It would be awesome if this kind of "migrate your Citus into new hardware" type of scenario would be covered in the documentation.
The text was updated successfully, but these errors were encountered:
TLDR - I'd like to have documentation covering Citus cluster migration scenarios
Our background (the non-TLDR part) is covered in Citus Slack
To successfully migrate a complete Citus cluster the needed steps are somewhat clear:
I believe it's possible to replicate a (current) Citus cluster into empty new Citus cluster (with same number of workers) using PUBLICATION + SUBSCRIPTION across each old and new workers, and across the old and new coordinators. By reading the documentation about CDC the idea is clear - just create pubs+subs not for ALL TABLES, but for the non-metadata user tables. But as we're replicating each worker and coordinators separately it's unclear if citus.enable_change_data_capture brings anything on the table (as we want to each target worker / shard to get 1:1 changes from the source)
It would be awesome if this kind of "migrate your Citus into new hardware" type of scenario would be covered in the documentation.
The text was updated successfully, but these errors were encountered: