[TRAFODION-2663] Simplify HBase config settings in installers #1140
Conversation
Remove settings that are now in per-table trafodion-site.xml config. For Ambari, 2.2 service_advisor over-rides the 2.1 version. This allows possibility of single plug-in to manage 2.1 & 2.2 installs, though this is not tested.
Check Test Started: https://jenkins.esgyn.com/job/Check-PR-master/1876/ |
Test Failed. https://jenkins.esgyn.com/job/Check-PR-master/1876/ |
@selvaganesang notice that hbcheck failed on all the HDP nodes. |
@@ -122,10 +122,6 @@ def mod(self): | |||
} | |||
self.p.post('%s/config_groups' % cluster_url, hbase_config_group) | |||
|
|||
if dbcfgs['secure_hadoop'].upper() == 'Y': |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should these lines stay around for secure hadoop?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
These are not values we are adding, just values we are trying to avoid deleting. Since we are no longer writing value to hbase.coprocessor.region.classes, there is no danger of over-writing these security values.
Can I have access to the Jenkins VM? |
jenkins, retest -- The test volumes are not retained that long, so need to re-run to generate failed VM to debug. |
New Check Test Started: https://jenkins.esgyn.com/job/Check-PR-master/1887/ |
Test Failed. https://jenkins.esgyn.com/job/Check-PR-master/1887/ |
Simplifying the HBase config seems to have changed the timing of HBase start-up on HDP nodes. We are trying to talk to HBase while master is still initializing. I have added re-tries in hbcheck to compensate.
New Check Test Started: https://jenkins.esgyn.com/job/Check-PR-master/1893/ |
Test Failed. https://jenkins.esgyn.com/job/Check-PR-master/1893/ |
This change-set exposed a race condition in installation. Hbase regionserver was dying because pyinstaller is starting regionserver before the new staging dir exists in hdfs. Prior to simplifying the settings, installer won the race of creating hdfs dir before regionserver looked for it. To address this, we should just use HBase default value for. Also remove hbase setting for hbase.client.keyvalue.maxsize from installer. That can be added to trafodion-site.xml if needed.
New Check Test Started: https://jenkins.esgyn.com/job/Check-PR-master/1900/ |
Missed one file in prior commit.
Test Failed. https://jenkins.esgyn.com/job/Check-PR-master/1900/ |
Previous Test Aborted. New Check Test Started: https://jenkins.esgyn.com/job/Check-PR-master/1901/ |
Test Failed. https://jenkins.esgyn.com/job/Check-PR-master/1901/ |
Previous Test Aborted. New Check Test Started: https://jenkins.esgyn.com/job/Check-PR-master/1902/ |
Test Passed. https://jenkins.esgyn.com/job/Check-PR-master/1902/ |
+1 |
Remove settings that are now in per-table trafodion-site.xml config.
For Ambari, 2.2 service_adviser over-rides the 2.1 version. This allows
possibility of single plug-in to manage 2.1 & 2.2 installs, though this is not
tested.
Difference in 2.2 service adviser is just removal of these same hbase properties.