Performance load test issues on OSP 8 with DPDK environment
Issue
During performance load test with RTP calls and after PIM board takeover. It is possible to find dpdk_watchdog2 BLOCK
in ovs-vswitchd.log
in the compute node. This ends with restarting the openvswitch on the host:
2016-06-14T08:48:56.900Z|01564|ovs_rcu(vhost_thread1)|WARN|blocked 64000 ms waiting for dpdk_watchdog2 to quiesce
2016-06-14T08:48:56.900Z|00010|ovs_rcu(urcu3)|WARN|blocked 64000 ms waiting for dpdk_watchdog2 to quiesce
2016-06-14T08:48:57.051Z|00879|ovs_rcu|WARN|blocked 64000 ms waiting for dpdk_watchdog2 to quiesce
2016-06-14T08:50:00.900Z|00011|ovs_rcu(urcu3)|WARN|blocked 128000 ms waiting for dpdk_watchdog2 to quiesce
2016-06-14T08:50:00.900Z|01565|ovs_rcu(vhost_thread1)|WARN|blocked 128000 ms waiting for dpdk_watchdog2 to quiesce
2016-06-14T08:50:01.051Z|00880|ovs_rcu|WARN|blocked 128000 ms waiting for dpdk_watchdog2 to quiesce
Environment
Red Hat OpenStack Platform 8
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.