-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/48266/#review137326
-----------------------------------------------------------



Thanks for providing this patch.

Most of the changes seem straightforward to me, but I'm concerned about 
attaching this check to the Ambari Agent HeartBeat handler.  My concern is that 
this could have performance implications later on.

I'd request that perhaps an alternate approach be considered, since there might 
be ways to implement this as a separate task thread in the TopologyManager, 
that could then add the required logging once the cluster deployment completes.

I think it makes sense to ask Sumit Mohanty to review this as well.  

Thanks.


ambari-server/src/main/java/org/apache/ambari/server/agent/HeartbeatProcessor.java
 (line 191)
<https://reviews.apache.org/r/48266/#comment202486>

    I'm a little concerned that this approach could have performance issues 
down the road.
    
    I'm wondering if it might be worth considering launching a separate task 
thread when the cluster is created by the TopologyManager, that can monitor the 
state of the LogicalRequest as the deployment progresses.
    
    I would recommend asking Sumit to review this as well, since it's not clear 
to me that this is the approach that should be taken.


- Robert Nettleton


On June 11, 2016, 6:43 a.m., Daniel Gergely wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/48266/
> -----------------------------------------------------------
> 
> (Updated June 11, 2016, 6:43 a.m.)
> 
> 
> Review request for Ambari, Laszlo Puskas, Oliver Szabo, Robert Nettleton, 
> Sandor Magyari, Sumit Mohanty, and Sebastian Toader.
> 
> 
> Bugs: AMBARI-17053
>     https://issues.apache.org/jira/browse/AMBARI-17053
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> Add a log message to see when the cluster is ready to use (cluster creation 
> finishes).
> 
> An event after each heartbeat. At that time cluster provision request is 
> checked if it is finished or not. In case of an ambari server restart, the 
> replayed requests are used to determine which one was the provision request. 
> I could not find a nice way to do it, but I could make a workaround.
> 
> 
> Diffs
> -----
> 
>   
> ambari-server/src/main/java/org/apache/ambari/server/agent/HeartbeatProcessor.java
>  c6036c2 
>   
> ambari-server/src/main/java/org/apache/ambari/server/events/AmbariEvent.java 
> 1079806 
>   
> ambari-server/src/main/java/org/apache/ambari/server/events/HeartbeatProcessingFinishedEvent.java
>  PRE-CREATION 
>   
> ambari-server/src/main/java/org/apache/ambari/server/topology/PersistedState.java
>  77419d8 
>   
> ambari-server/src/main/java/org/apache/ambari/server/topology/PersistedStateImpl.java
>  324a397 
>   
> ambari-server/src/main/java/org/apache/ambari/server/topology/TopologyManager.java
>  e3f5b49 
>   
> ambari-server/src/test/java/org/apache/ambari/server/topology/TopologyManagerTest.java
>  fd8653c 
> 
> Diff: https://reviews.apache.org/r/48266/diff/
> 
> 
> Testing
> -------
> 
> Succeeded locally
> 
> 
> Thanks,
> 
> Daniel Gergely
> 
>

Reply via email to