Back to bug 1315114

Who When What Removed Added
David Hill 2016-03-06 23:01:13 UTC CC dhill
David Hill 2016-03-06 23:01:30 UTC Summary When restarting neutron-openvswitch-agent on networks nodes with logs of port, server crashes When restarting neutron-openvswitch-agent on networks nodes with lot of ports, server crashes
Benjamin Schmaus 2016-03-06 23:03:43 UTC CC bschmaus
RHEL Program Management 2016-03-06 23:05:03 UTC Keywords ZStream
Red Hat Bugzilla Rules Engine 2016-03-06 23:10:05 UTC Target Release --- 7.0
Jaison Raju 2016-04-01 07:30:35 UTC CC jraju
Chris Fields 2016-04-06 16:07:35 UTC CC chris.fields
Matt Flusche 2016-04-06 20:09:01 UTC CC mflusche
Flags needinfo?(amuller)
Assaf Muller 2016-04-07 13:59:56 UTC Flags needinfo?(amuller) needinfo?(chris.fields)
Chris Fields 2016-04-07 20:05:01 UTC Flags needinfo?(chris.fields)
Chris Lincoln 2016-04-08 20:52:38 UTC CC clincoln
Chris Lincoln 2016-04-08 20:53:37 UTC Flags needinfo?(amuller)
Assaf Muller 2016-04-12 00:08:25 UTC Flags needinfo?(amuller) needinfo?(clincoln)
Assaf Muller 2016-04-12 00:08:39 UTC Flags needinfo?(dhill)
John Skeoch 2016-04-18 07:54:44 UTC CC yeylon srevivo
Perry Myers 2016-04-19 01:07:51 UTC CC pmyers
Chris Lincoln 2016-04-21 21:30:10 UTC Flags needinfo?(amuller)
Chris Lincoln 2016-05-02 13:52:30 UTC Flags needinfo?(clincoln) needinfo?(dhill) needinfo?(amuller) needinfo-
Assaf Muller 2016-05-04 22:00:33 UTC Summary When restarting neutron-openvswitch-agent on networks nodes with lot of ports, server crashes UnixDomainWSGIServer keepalived state change listener in L3 agent has an uncapped number of threads, overloading node
Assaf Muller 2016-05-04 22:01:05 UTC Priority unspecified high
Status NEW ASSIGNED
Assignee amuller vkommadi
Assaf Muller 2016-05-04 22:01:23 UTC Target Milestone --- async
Brandon Perkins 2016-05-04 22:03:36 UTC CC bperkins
anil venkata 2016-05-09 18:29:19 UTC Flags needinfo?(clincoln)
anil venkata 2016-05-18 13:38:03 UTC Link ID Launchpad 1581580
Chris Lincoln 2016-05-19 20:42:13 UTC Flags needinfo?(clincoln)
anil venkata 2016-05-20 13:19:19 UTC Flags needinfo?(clincoln)
Chris Lincoln 2016-05-20 14:21:39 UTC Flags needinfo?(clincoln)
John Skeoch 2016-08-08 01:27:48 UTC CC clincoln dchia
anil venkata 2016-09-27 13:38:29 UTC Link ID OpenStack gerrit 317616
anil venkata 2016-10-04 14:55:27 UTC Blocks 1381618
anil venkata 2016-10-04 14:57:22 UTC Blocks 1381619
anil venkata 2016-10-04 14:59:31 UTC Blocks 1381620
anil venkata 2016-11-21 12:00:25 UTC Fixed In Version openstack-neutron-2015.1.4-10.el7ost
anil venkata 2016-11-21 12:00:42 UTC Status ASSIGNED MODIFIED
errata-xmlrpc 2016-11-28 19:43:57 UTC Status MODIFIED ON_QA
Toni Freger 2016-11-29 09:53:24 UTC QA Contact tfreger ekuris
Eran Kuris 2016-12-01 10:28:47 UTC Status ON_QA VERIFIED
Martin Lopes 2017-01-13 00:50:37 UTC CC mlopes, vkommadi
Flags needinfo?(vkommadi)
anil venkata 2017-01-13 05:39:20 UTC Doc Text Cause:
With wsgi_default_pool_size(=100) concurrent requests, state change server is creating heavy cpu load on l3 agent.

Consequence:

Fix:
A new option ``ha_keepalived_state_change_server_threads`` has been added to configure the number of concurrent threads pawned for keepalived server connection requests. Higher values increase the CPU load on the agent nodes. The default value is half of the number of CPUs present on the node. This allows operators to tune the number of threads to suit their environment. With more threads, simultaneous requests for multiple HA routers state change can be handled faster.

Result:
ha_keepalived_state_change_server_threads can be configured to avoid high load on l3 agents.
Flags needinfo?(vkommadi)
Martin Lopes 2017-01-13 06:16:27 UTC Doc Text Cause:
With wsgi_default_pool_size(=100) concurrent requests, state change server is creating heavy cpu load on l3 agent.

Consequence:

Fix:
A new option ``ha_keepalived_state_change_server_threads`` has been added to configure the number of concurrent threads pawned for keepalived server connection requests. Higher values increase the CPU load on the agent nodes. The default value is half of the number of CPUs present on the node. This allows operators to tune the number of threads to suit their environment. With more threads, simultaneous requests for multiple HA routers state change can be handled faster.

Result:
ha_keepalived_state_change_server_threads can be configured to avoid high load on l3 agents.
Prior to this update, with wsgi_default_pool_size(=100) concurrent requests, the state change server would create a heavy CPU load on the l3 agent.
With this update, a new option `ha_keepalived_state_change_server_threads` has been added to configure the number of concurrent threads spawned for keepalived server connection requests. Higher values increase the CPU load on the agent nodes. The default value is half of the number of CPUs present on the node. This allows operators to tune the number of threads to suit their environment. With more threads, simultaneous requests for multiple HA routers state change can be handled faster.
As a result, ha_keepalived_state_change_server_threads can be configured to avoid high load on l3 agents.
errata-xmlrpc 2017-01-18 01:13:02 UTC Status VERIFIED RELEASE_PENDING
errata-xmlrpc 2017-01-19 13:25:51 UTC Status RELEASE_PENDING CLOSED
Resolution --- ERRATA
Last Closed 2017-01-19 08:25:51 UTC

Back to bug 1315114