[OPNFLWPLUG-363] of-flow: while configuring 100k flows karaf went down Created: 18/Feb/15  Updated: 27/Sep/21  Resolved: 27/May/15

Status: Resolved
Project: OpenFlowPlugin
Component/s: General
Affects Version/s: None
Fix Version/s: None

Type: Bug
Reporter: Peter Gubka Assignee: Martin Bobak
Resolution: Cannot Reproduce Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All

External issue ID: 2729

 Description   

odl: https://nexus.opendaylight.org/content/repositories/opendaylight.snapshot/org/opendaylight/integration/distribution-karaf/0.3.0-SNAPSHOT//distribution-karaf-0.3.0-20150217.002825-479.tar.gz

configured features: odl-openflowplugin-all (IMDS used, not clustered DS)

max java heap was set to 4G, perm size to 1G

To running odl i connected 250 switches with mininet and start configuring 100k flows. The task was not finished because karaf died. When i logged on the testing machine, karaf process (java) was not present there.

To configure you can use scripts from
integration/test/tools/odl-mdsal-clustering-tests/clustering-performance-test
e.g. flow_add_delete_test.py --host <host> --threads 5 --flows 20000 --auth

The log file had 3G.

 Comments   
Comment by RichardHill [ 25/Feb/15 ]

Which garbage collector was used? Which JDK? What did the log show?

Comment by Peter Gubka [ 26/Feb/15 ]

(In reply to Hill from comment #1)
> Which garbage collector was used? Which JDK? What did the log show?

garbage collector - default
java java -version
java version "1.7.0_67"
Java(TM) SE Runtime Environment (build 1.7.0_67-b01)
Java HotSpot(TM) 64-Bit Server VM (build 24.65-b04, mixed mode)

logs because of the size (3G) delivered to Martin Bobak directly

Comment by Peter Gubka [ 26/Feb/15 ]

this bug may be related to https://bugs.opendaylight.org/show_bug.cgi?id=2765
the common thing they have is that they connect 250 switches
after that (when flows are being configured) the log file is full of exceptions and errors

Comment by Peter Gubka [ 27/May/15 ]

Att he ent of april his was not a problem any longer.

But, to reproduce exactly the same testcase it will take a lot longer now, as clustered datastore is used now.

And new tests are being created. It is not only important to configure 100k flows, but also collect flow statistics for them. As it is not set the requirement for statistics scale

Closing.

Generated at Wed Feb 07 20:32:16 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.