ambari-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alejandro Fernandez" <afernan...@hortonworks.com>
Subject Re: Review Request 36396: RU fails for ZooKeeper only cluster
Date Fri, 10 Jul 2015 17:12:33 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/36396/#review91325
-----------------------------------------------------------

Ship it!


Ship It!

- Alejandro Fernandez


On July 10, 2015, 5:03 p.m., Dmitro Lisnichenko wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/36396/
> -----------------------------------------------------------
> 
> (Updated July 10, 2015, 5:03 p.m.)
> 
> 
> Review request for Ambari, Alejandro Fernandez, Andrew Onischuk, Jonathan Hurley, Nate
Cole, and Tom Beerbower.
> 
> 
> Bugs: AMBARI-12375
>     https://issues.apache.org/jira/browse/AMBARI-12375
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> RU for ZooKeeper only cluster fails on 'Updating configuration hadoop-env' step ('Prepare
Upgrade' stage) with the following error: Server action failed.
> 
> Issue is in added "configure" task in PRE-CLUSTER (details are described at jira)
> I've looked through our upgrade order, and see that we may move this code to Namenode
pre-upgrade sequence. The only difference between executing this code in PRE-CLUSTER and CORE_MASTERS
is that by that time ZOOKEEPER_SERVER, RANGER_ADMIN, RANGER_USERSYNC, JOURNALNODE are restarted.
But these services don't seem to be related to changed properties, so doing that should be
safe.
> Also, I remove support for "configure" commands outside of processing#service#component,
because this approach turns out to be error-prone (each config is owned by a service, so changing
config without checking if service exists at all is an unsafe assumption)
> 
> 
> Diffs
> -----
> 
>   ambari-server/src/main/java/org/apache/ambari/server/state/stack/upgrade/ClusterGrouping.java
5b6bc50 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/after-INSTALL/scripts/params.py
11de040 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/hook.py
1fd36d6 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/params.py
e32f750 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-START/scripts/hook.py
f21e4b1 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-START/scripts/params.py
790769a 
>   ambari-server/src/main/resources/stacks/HDP/2.2/upgrades/upgrade-2.3.xml a202fd7 
> 
> Diff: https://reviews.apache.org/r/36396/diff/
> 
> 
> Testing
> -------
> 
> Performed few manual checks of RU on ZK-only clusters. One final run failed with intermittent
issue described at jira (seems to be a race condition unrelated to ZK), other was successful.
> Now performing manual check of 2.2->2.3 RU on HDFS+YARN+HBASE+ZK+OZIE+RANGER
> 
> 
> [INFO] ------------------------------------------------------------------------
> [INFO] Reactor Summary:
> [INFO] 
> [INFO] Ambari Views ...................................... SUCCESS [4.056s]
> [INFO] Ambari Metrics Common ............................. SUCCESS [2.114s]
> [INFO] Ambari Server ..................................... SUCCESS [57:11.160s]
> [INFO] ------------------------------------------------------------------------
> [INFO] BUILD SUCCESS
> [INFO] ------------------------------------------------------------------------
> [INFO] Total time: 57:17.960s
> [INFO] Finished at: Fri Jul 10 19:28:28 EEST 2015
> [INFO] Final Memory: 40M/595M
> [INFO] ------------------------------------------------------------------------
> 
> 
> Thanks,
> 
> Dmitro Lisnichenko
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message