[ https://issues.apache.org/jira/browse/SENTRY-1874?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16121153#comment-16121153 ]
Hadoop QA commented on SENTRY-1874: ----------------------------------- Here are the results of testing the latest attachment https://issues.apache.org/jira/secure/attachment/12881148/SENTRY-1874.01.patch against master. {color:red}Overall:{color} -1 due to an error {color:red}ERROR:{color} failed to build with patch (exit code 1) Console output: https://builds.apache.org/job/PreCommit-SENTRY-Build/3137/console This message is automatically generated. > Do not require quiet HMS when taking initial HMS snapshot > --------------------------------------------------------- > > Key: SENTRY-1874 > URL: https://issues.apache.org/jira/browse/SENTRY-1874 > Project: Sentry > Issue Type: Sub-task > Components: Sentry > Affects Versions: 2.0.0 > Reporter: Alexander Kolbasov > Assignee: Alexander Kolbasov > Labels: sentry-ha > Fix For: 2.0.0 > > Attachments: SENTRY-1874.01.patch > > > When Sentry starts initially, it needs to get a full snapshot from HMS and > associate it with a specific notification ID. HMS does not provide any > mechanism to get a consistent snapshot with known notification ID. SO Sentry > does the following: > 1) Read current notification ID_0 > 2) Obtain full snapshot from HMS > 3) Read current notification ID_1 > 4) If (ID_0 != ID_1) goto 1. > This creates a problem that the process may never converge when there is some > constant stream of HMS activity affecting notification IDs. We would like to > be able to use a problem that can converge in the presence of updates. -- This message was sent by Atlassian JIRA (v6.4.14#64029)