[
https://issues.apache.org/jira/browse/METRON-1565?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Nick Allen updated METRON-1565:
-------------------------------
Fix Version/s: 0.5.0
> Metaalerts fix denormalization after moving to active status
> ------------------------------------------------------------
>
> Key: METRON-1565
> URL: https://issues.apache.org/jira/browse/METRON-1565
> Project: Metron
> Issue Type: Bug
> Reporter: Ryan Merriman
> Assignee: Ryan Merriman
> Priority: Major
> Fix For: 0.5.0
>
>
> The process of updating from inactive to active is slightly incomplete. Check
> out
> [https://github.com/apache/metron/pull/970/files#diff-b7359d01c3ffbed48b7fdaa2d32169e7R246].
> Say we have these three steps:
> - Metaalert is updated to inactive
> - (Former) child alert is updated.
> - Metaalert is made active again.
> The update will be missing from the metaalert. We need to update the
> metaalert with the current state of any alerts (which we conveniently have
> because we needed to update them all anyway!). This is a problem with both ES
> and Solr (which shouldn't be surprising since that link is to the abstract
> DAO). Basically, this should just be adding all the alerts in the metaalert
> back into the updated version of the document before passing it to the
> update. It also needs an associated test case.
> This fix should be made against master and pulled into the Solr branch (and
> the metaalerts PR afterwards). See
> [https://github.com/apache/metron/blob/e59059bd9707a6ca46c4137d796b8f2943f06b43/metron-platform/metron-elasticsearch/src/main/java/org/apache/metron/elasticsearch/dao/ElasticsearchMetaAlertDao.java#L351].
> Main thing might just be reordering the class to build the metaalert update
> itself after grabbing all the alerts.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
