Details
Description
Description
Description:
This introduces logging for oplog entries that are slow to apply on secondaries. The rule is to log if the op application took longer than "slowMS". This is the same threshold that we use for slow operation logging on primaries.
Note that, unlike slow op logging on primaries, this feature does not come with an option to log all ops you apply. Only oplog entries considered slow to apply (i.e. above the "slowMS" threshold) will be logged on secondaries.
Engineering Ticket Description:
We log slow operations on the primary, it would also be helpful to log slow oplog application to see what's using system resources.
Scope of changes
- source/release-notes/4.2.txt
- relnotes and others mention that the log verbosity level does not affect this logging. As such, for now, will not update on the verbosity logging section that the setting has no effect on secondary slow oplog logging.
- source/core/replica-set-oplog.txt
- source/core/replica-set-secondary.txt
- source/core/replica-set-sync.txt
- source/includes/apiargs-dbcommand-profile-fields.yaml
- source/includes/extracts-4.2-changes.yaml
- source/includes/fact-log-slow-queries.rst
- source/includes/options-mongod.yaml
- source/reference/log-messages.txt
- source/replication.txt
- source/tutorial/manage-the-database-profiler.txt
- source/administration/configuration
- source/administration/monitoring.txt
- source/tutorial/evaluate-operation-performance.txt
- source/tutorial/troubleshoot-replica-sets.txt
Impact to Other Docs
none
MVP (Work and Date)
Resources (Scope or Design Docs, Invision, etc.)
Attachments
Issue Links
- documents
-
SERVER-32146 Log slow oplog entry application
-
- Closed
-