[CONTROLLER-1398] Performance bottleneck Created: 31/Jul/15  Updated: 13/Nov/17  Resolved: 13/Nov/15

Status: Resolved
Project: controller
Component/s: mdsal
Affects Version/s: Helium
Fix Version/s: None

Type: Bug
Reporter: Mina Assignee: Unassigned
Resolution: Won't Do Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All

Attachments: PNG File ODL_MDSAL.png    
External issue ID: 4076

 Description   

Creating 10 services on 2 nodes. Running against the Helium release, I can see so many invocation calls to the md-sal while profiling.

Please see attached for the profiling snapshots.

 Comments   
Comment by Mina [ 31/Jul/15 ]

Attachment ODL_MDSAL.png has been added with description: MD-SAL

Comment by Martin Ciglan [ 03/Aug/15 ]

Hi Mina

Please refer to Lithium build. Thank you very much.

Martin

Comment by Tony Tkacik [ 03/Aug/15 ]

This time is spended creating and invoking change events on listeners, newer faster API is available in Lithium for Data Change Listeners.

25% of CPU time, during ten invocations is not bad, if you are doing profiling,
please monitor controller for longer period and use larger data sets, large burst operations to distinguish normal operations from sub-par performant ones.

Comment by Robert Varga [ 13/Nov/15 ]

The fix is to switch to DataTreeChangeListener – there is nothing we can optimize on DataChangeListeners, unfortunately.

Generated at Wed Feb 07 19:55:27 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.