Details
-
Bug
-
Resolution: Fixed
-
Critical
-
Cheshire-Cat
-
Untriaged
-
1
-
No
Description
When the number of pipelines increases, the number of pipeline supervisors also increase linearly. Each supervisor is essentially doing duplicate work of hitting ns_server's pools/nodes endpoint. Too many of these supervisors doing the same thing will cause un-necessary load.
This should be avoided, and the effort should be deduplicated.
Attachments
Issue Links
- backports to
-
MB-47672 [BP 6.6.4] - XDCR hitting pools/nodes endpoint should be coordinated
- Closed
- finishs with
-
MB-23826 XDCR should use 11210/CCCP/Optimized Connection Management for Remote Cluster Topology
- Resolved
- relates to
-
MB-45893 Windows : Unable to add a new 7.0 node to a mixed mode cluster - pre_rebalance_janitor_run_failed
- Closed
-
MB-45387 [System Test]Failed on calling /pools/nodes, err=json: cannot unmarshal array into Go value of type map[string]interface {}, statusCode=500 - error in longevity
- Closed
For Gerrit Dashboard: MB-44823 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
152288,3 | MB-44823 - coordinate contact to ns_server's rest endpoint to pools/default | master | goxdcr | Status: MERGED | +2 | +1 |
160988,3 | MB-48323 - return latest cached value for KV check or memcached address | master | goxdcr | Status: MERGED | +2 | +1 |
164677,1 | MB-47672 - return latest cached value for KV check or memcached address | master | goxdcr | Status: ABANDONED | 0 | 0 |
164679,3 | MB-47672 - return latest cached value for KV check or memcached address | mad-hatter | goxdcr | Status: MERGED | +2 | +1 |