[jira] [Created] (BROOKLYN-341) Riak: temporary (benign) error polling for sensor values

["Aled Sage (JIRA)" <ji...@apache.org>]

Aled Sage created BROOKLYN-341:
----------------------------------

             Summary: Riak: temporary (benign) error polling for sensor values
                 Key: BROOKLYN-341
                 URL: https://issues.apache.org/jira/browse/BROOKLYN-341
             Project: Brooklyn
          Issue Type: Bug
            Reporter: Aled Sage
            Priority: Minor


Using Brooklyn 0.10.0-SNAPSHOT...

I deployed a single Riak node to CentOS 6.8 on OpenStack (BlueBox singapore). It works ok, but in the log I see every couple of minutes a 500 response from the Riak node when polling for the sensor values. For example:

{noformat}
2016-08-29 13:16:29,906 WARN  o.a.b.c.f.AttributePollHandler [brooklyn-execmanager-lREia443-615]: Read of RiakNodeImpl{id=ovs7drm17d}->Sensor: riak.node.gets.total (java.lang.Integer) gave failure: HttpToolRespons
e{responseCode=500}
2016-08-29 13:16:29,906 TRACE o.a.b.core.sensor.AttributeMap [brooklyn-execmanager-lREia443-615]: setting attribute riak.node.gets.total to -1 (was 0) on RiakNodeImpl{id=ovs7drm17d}
...
2016-08-29 13:16:30,517 INFO  o.a.b.c.f.AttributePollHandler [brooklyn-execmanager-lREia443-615]: Success (following previous problem) reading RiakNodeImpl{id=ovs7drm17d}->Sensor: riak.node.gets.total (java.lang.I
nteger)
2016-08-29 13:16:30,517 TRACE o.a.b.core.sensor.AttributeMap [brooklyn-execmanager-lREia443-615]: setting attribute riak.node.gets.total to 0 (was -1) on RiakNodeImpl{id=ovs7drm17d}
2016-08-29 13:16:31,132 TRACE o.a.b.core.sensor.AttributeMap [brooklyn-execmanager-lREia443-560]: setting attribute riak.node.gets.total to 0 (was 0) on RiakNodeImpl{id=ovs7drm17d}
{noformat}

When we get the failure, it seems to happen for all the sensors that we are polling at that time - e.g. 22 warnings are logged in rapid succession for the different sensors, and then a second later we get 22 log.info messages of "Success (following previous problem)" (one for each sensor).

The worst thing about this is the excessive logging that appears in the info log.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)