-
Notifications
You must be signed in to change notification settings - Fork 1.1k
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
[platform] on-premise create universe flow does not accurately place masters in multi-region configuration #3999
Comments
So after conducting a repro with 2.1.2, I find that it picks the first 3 nodes provisioned (n1,n2,n3) regardless of location. |
I've isolated the fault. The problem is that in the on-premise flow we have no subnet information stored in the
We may want to consider a unique key constraint on that field or simply defaulting it to the unique key generated for the availability_zone record. |
Summary: The current master selection logic used subnets instead of zones and regions, which caused issues for on-prem flows, where the subnet might be null. Since subnets and azs are mapped one-to-one anyways, we can use the zones to select master placement. The updated logic now cycles the master selection in each region/zone combination before putting a master in the same region/zone combination again. Test Plan: Added a unit test. Also tested by deploying a multi-region/multi-zone universe and verifying that the masters got sprayed correctly. Reviewers: ram, bogdan Reviewed By: bogdan Subscribers: kannan, jenkins-bot, yugaware Differential Revision: https://phabricator.dev.yugabyte.com/D8164
this issue has been resolved with the commit referenced above. |
Steps to repro:
The text was updated successfully, but these errors were encountered: