Details
-
New Feature
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Won't Do
-
23.02.2
-
None
Description
There should be a easy way to provide exclusion list for xpand nodes when dynamic_node_detection is enabled.
Our customer has 105 nodes in a Xpand cluster. I have been exploring possibility of dedicating some nodes to exclusively serve binlogs and not participate in production workloads. That will help replication keep up to their DR cluster with out any delays. However there is no easy way to do that via maxscale today.
Since they have 105 nodes, its not recommended that the provide manual entry for production nodes and for replication server nodes. They would like dynamic detection to identify all nodes in a cluster. Given that, there is no way today to prevent maxscale scale from sending workloads to certain nodes. With dynamic_node_detection enabled, workloads are sent to all nodes.
Attachments
Activity
| Field | Original Value | New Value |
|---|---|---|
| Description |
There should be a easy way to provide exclusion list for xpand nodes when dynamic_node_detection is enabled.
Our customer PhonePE has 105 nodes in a Xpand cluster. I have been exploring possibility of dedicating some nodes to exclusively serve binlogs and not participate in production workloads. That will help replication keep up to their DR cluster with out any delays. However there is no easy way to do that via maxscale today. Since they have 105 nodes, its not recommended that the provide manual entry for production nodes and for replication server nodes. They would like dynamic detection to identify all nodes in a cluster. Given that, there is no way today to prevent maxscale scale from sending workloads to certain nodes. With dynamic_node_detection enabled, workloads are sent to all nodes. |
There should be a easy way to provide exclusion list for xpand nodes when dynamic_node_detection is enabled.
Our customer has 105 nodes in a Xpand cluster. I have been exploring possibility of dedicating some nodes to exclusively serve binlogs and not participate in production workloads. That will help replication keep up to their DR cluster with out any delays. However there is no easy way to do that via maxscale today. Since they have 105 nodes, its not recommended that the provide manual entry for production nodes and for replication server nodes. They would like dynamic detection to identify all nodes in a cluster. Given that, there is no way today to prevent maxscale scale from sending workloads to certain nodes. With dynamic_node_detection enabled, workloads are sent to all nodes. |
| Rank | Ranked higher |
| Assignee | Todd Stoffel [ toddstoffel ] |
| Fix Version/s | Icebox [ 22648 ] |
| Assignee | Todd Stoffel [ toddstoffel ] | Joe Cotellese [ JIRAUSER54006 ] |
| Component/s | xpandmon [ 14701 ] | |
| Fix Version/s | N/A [ 22001 ] | |
| Fix Version/s | Icebox [ 22648 ] | |
| Resolution | Won't Do [ 10201 ] | |
| Status | Open [ 1 ] | Closed [ 6 ] |