Details
-
Bug
-
Resolution: Fixed
-
Critical
-
4.5.1, 4.6.0, 4.6.1, 4.6.4, 4.6.2, 4.6.3, 4.6.5, 5.0.0, 5.0.1, 5.1.0, 5.1.1, 5.1.2
-
Yes
-
CBSS Sprint 03-16-2018
Description
Background
Every 10 mins GSI indexes writes a snapshot is to disk.
Problem
When snapshotting is happening the CPU usage burst to 100%, there is a a suggestion that this could have an impact on index scan latency and to the other services running on the node (Data, Query, Search, etc)
Workaround
Decrease the number of persistence thread(this workaround is only effective if "number of indexes" < "number of cores" as minimum 1 persistence thread gets assigned to each index)
curl -X POST -u <user:pass> http://localhost:9102/settings --data '{"indexer.settings.moi.persistence_threads": 1}'
|