Description of problem: Today we skip some tables fully and selective columns of other tables in Open_vSwitch and HWVTEP plugins as part of fixes for bugs 6463, 7201 etc. However there is no way to let user configure which tables and columns to skip. Some use cases require statistics or status information. Also, when doing scale/perf testing it becomes difficult to skip any extra columns or allow few extra tables/columns without access to code.
Proposed the following downstream patch: https://code.engineering.redhat.com/gerrit/#/c/98376/
Hi, Please provide information how to verify.
Before starting Karaf, create a file (etc/org.opendaylight.ovsdb.southbound.cfg) with the following contents. Start the ODL controller. skip-monitoring-manager-status = true After stacking, on one of the compute nodes, enable OVSDB debug logs and look for the Manager update messages. Log file (on Fedora): /var/log/openvswitch/ovsdb-server.log Note: To enable debug in ovsdb, you can do the following. sudo ovs-appctl -t /var/run/openvswitch/ovsdb-server.4233.ctl vlog/set dbg When "status" feild is not skipped (i.e., without the patch), following would be the logs: -------------------------------------------------------------------------------------------- 2017-03-07T07:24:51.264Z|00370|jsonrpc|DBG|tcp:192.168.121.1:6640: send notification, method="update", params=["e33befae-7f91-474e-a872-f754983cbcb8",{"Manager":{"274ee282-10b9-474c-b407-c3a6964c22fb":{"new":{"connection_mode":["set",[]],"is_connected":true,"other_config":["map",[]],"external_ids":["map",[]],"status":["map",[["sec_since_connect","66"],["sec_since_disconnect","67"],["state","ACTIVE"]]],"_uuid":["uuid","274ee282-10b9-474c-b407-c3a6964c22fb"],"target":"tcp:192.168.121.1:6640","max_backoff":["set",[]],"inactivity_probe":["set",[]]},"old":{"status":["map",[["sec_since_connect","61"],["sec_since_disconnect","62"],["state","ACTIVE"]]]}}}}] With the patch: -------------------------------------------------------------------------------------------- 2017-03-07T07:43:51.075Z|00214|jsonrpc|DBG|tcp:192.168.121.1:6640: send notification, method="update", params=["61eeafe9-2721-4cef-99eb-3c133e9bfa8f",{"Manager":{"8a31579f-7401-4484-bbc9-d779cfcb0777":{"new":{"connection_mode":["set",[]],"is_connected":true,"other_config":["map",[]],"external_ids":["map",[]],"_uuid":["uuid","8a31579f-7401-4484-bbc9-d779cfcb0777"],"target":"tcp:192.168.121.1:6640","inactivity_probe":["set",[]],"max_backoff":["set",[]]},"old":{"is_connected":false}}}}]
Similarly when there is a monitor request from the ODL Controller to the OVSDB server, the request would not include the "status" field. 2017-03-07T07:43:47.469Z|00145|jsonrpc|DBG|tcp:192.168.121.1:6640: received request, method="monitor", params=["Open_vSwitch","61eeafe9-2721-4cef-99eb-3c133e9bfa8f", ... <SNIP> "Manager":{"select":{"modify":true,"initial":true,"insert":true,"delete":true}, "columns":["is_connected","other_config","connection_mode","max_backoff","inactivity_probe","_uuid","external_ids","target"]}, <SNIP>
Verified according to Sridhar's comments with opendaylight-5.2.0-6.el7ost.noarch
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2017-0840.html