Details
-
Bug
-
Resolution: Unresolved
-
Major
-
7.1.0, 7.1.1
-
None
-
Untriaged
-
1
-
Unknown
Description
Many places in indexer code uses a "selfRestart" method. E.g., monitorKVNodes. If there is any error during the processing of methods like monitorKVNodes, indexer will sleep for 5 seconds and re-spawn the method.
Each time methods like monitorKVNodes start, indexer is doing an entire cluster info cache fetch. This is unnecessary as in most cases, it is sufficient for indexer to just fetch pools or node services info.
The ask of this MB is to:
a. Identify all such places where selfRestart() method is leading to full cluster info cache fetch
b. Avoid doing full cluster info cache fetch in those places
Attachments
Gerrit Reviews
For Gerrit Dashboard: MB-52393 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
175580,2 | MB-52393 Fetch only nodes and services info in MonitorKVNodes | unstable | indexing | Status: ABANDONED | 0 | 0 |