> On Dec. 28, 2015, 3:47 p.m., Apache Ambari wrote:
> > ambari-server/src/main/java/org/apache/ambari/server/stack/StackDirectory.java,
> >  line 487
> > <https://reviews.apache.org/r/41740/diff/1/?file=1176805#file1176805line487>
> >
> >     If a stack doesn't define RCO, where it is inheriting a RCO from?
> >     
> >     I think we should keep the default RCO file, which should be treated as 
> > the fallback in case a stack defines no such file.
> 
> Andrew Onischuk wrote:
>     It's logical that if stack doesn't define it, stack doesn't want to have 
> any dependencies.
>     Why should it inherit HDP stack dependencies?
> 
> Dmitro Lisnichenko wrote:
>     +1, default RCO is probably an insufficient subset of required
> 
> Andrew Onischuk wrote:
>     What I mean is that if someone defines a new stack which is not HDP at 
> all I think ambari-server shouldn't be stick to that specific RCM. If we want 
> to do something default we should probably think of something like 
> common-services but common-stacks and make possible to link stacks to those.
> 
> Andrew Onischuk wrote:
>     Another problem here is that RCO in resources is very outdated and a lot 
> won't work with it. Nobody simply updates it.
> 
> Apache Ambari wrote:
>     I'm suggesting that the default RCO be very minimal since any stack that 
> doesn't define its own should pick it up.
>     Of course, HDP 2.4 should inherit from HDP 2.3.
> 
> Andrew Onischuk wrote:
>     1. minimal? Dependencies are pretty specific. I don't get this part.
>     2. resources shouldn't contain something which stack specific (HDP in our 
> case). Keeping in mind that ambari-server supports creating a new stack 
> without HDFS and any other services we have. Why in that case should that new 
> stack use HDP specific RCO?
> 
> Apache Ambari wrote:
>     Exactly my point Andrew; the default RCO shouldn't be HDP specific and 
> should only have very simple guidelines, e.g., starting NameNode requires 
> first starting DNs, or starting any master in YARN and HDFS requires ZK 
> Server to be started.
>     
>     It's ok to come up with another future Jira to fix the default RCO. I 
> don't think we should remove the file entirely, but rather fix the 
> inheritence model so that HDP 2.4 picks up the same RCO as HDP 2.3
> 
> Andrew Onischuk wrote:
>     What if stack doesn't have ZK Server and Namenodes and Datanode. Like 
> most non-HDP stacks which doesn't use HDFS
> 
> Andrew Onischuk wrote:
>     Why should default ambari-server has references to components which are 
> specific to HDP (you mentioned namenode,datanode, zk server), If the 
> ambari-server itself mean to support any non-HDP stacks with any sets of 
> services and components.
> 
> Andrew Onischuk wrote:
>     If we want to have defaults for stacks which can be reused between 
> mutliple stacks. We should impelement something like common-services, let's 
> say common-stacks, and put role_command_order.json , hooks, stacks-level 
> configs (e.g. cluster-env.xml), upgrade packs there, and make a reference 
> from stacks.
>     
>     The thing you're proposing seems to be a hack in terms of current Ambari 
> architecture. Which is simply making defaults specific to one stack for all 
> stacks.

If a new stack is created without a RCO, will it fail?
If so, then we need to keep the default RCO but leave it nearly empty and file 
a seperate Jira to minimilize it later.
It seems to me that this code is removing the default RCO in order to get HDP 
2.4 to work, which should instead inherit from HDP 2.3 even if a defalt RCO 
exists.


- Alejandro


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/41740/#review111989
-----------------------------------------------------------


On Dec. 28, 2015, 1:52 p.m., Andrew Onischuk wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/41740/
> -----------------------------------------------------------
> 
> (Updated Dec. 28, 2015, 1:52 p.m.)
> 
> 
> Review request for Ambari and Dmitro Lisnichenko.
> 
> 
> Bugs: AMBARI-14504
>     https://issues.apache.org/jira/browse/AMBARI-14504
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> Traceback (most recent call last):
>       File 
> "/var/lib/ambari-agent/cache/common-services/HIVE/0.12.0.2.0/package/scripts/service_check.py",
>  line 106, in <module>
>         HiveServiceCheck().execute()
>       File 
> "/usr/lib/python2.6/site-packages/resource_management/libraries/script/script.py",
>  line 219, in execute
>         method(env)
>       File 
> "/var/lib/ambari-agent/cache/common-services/HIVE/0.12.0.2.0/package/scripts/service_check.py",
>  line 103, in service_check
>         webhcat_service_check()
>       File 
> "/usr/lib/python2.6/site-packages/ambari_commons/os_family_impl.py", line 89, 
> in thunk
>         return fn(*args, **kwargs)
>       File 
> "/var/lib/ambari-agent/cache/common-services/HIVE/0.12.0.2.0/package/scripts/webhcat_service_check.py",
>  line 124, in webhcat_service_check
>         logoutput=True)
>       File 
> "/usr/lib/python2.6/site-packages/resource_management/core/base.py", line 
> 154, in __init__
>         self.env.run()
>       File 
> "/usr/lib/python2.6/site-packages/resource_management/core/environment.py", 
> line 158, in run
>         self.run_action(resource, action)
>       File 
> "/usr/lib/python2.6/site-packages/resource_management/core/environment.py", 
> line 121, in run_action
>         provider_action()
>       File 
> "/usr/lib/python2.6/site-packages/resource_management/core/providers/system.py",
>  line 238, in action_run
>         tries=self.resource.tries, try_sleep=self.resource.try_sleep)
>       File 
> "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", line 
> 70, in inner
>         result = function(command, **kwargs)
>       File 
> "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", line 
> 92, in checked_call
>         tries=tries, try_sleep=try_sleep)
>       File 
> "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", line 
> 140, in _call_wrapper
>         result = _call(command, **kwargs_copy)
>       File 
> "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", line 
> 291, in _call
>         raise Fail(err_msg)
>     resource_management.core.exceptions.Fail: Execution of 
> '/var/lib/ambari-agent/tmp/templetonSmoke.sh 
> os-r7-hdkhfu-devdeploy-3.novalocal ambari-qa 20111 
> idtest.ambari-qa.1451047104.22.pig no_keytab false kinit no_principal' 
> returned 1. Templeton Smoke Test (status cmd): Failed. : http_code <000>
>     
> 
> Probably the reason is because Webhcat should start before Hive check
> 
> Artifacts url:  
> <http://linux-jenkins.qe.hortonworks.com/home/jenkins/qe-artifacts/os-u14
> -bgetmu-ambari-se-21nf-1/ambari-se-21nf-1451273248/artifacts/screenshots/com.h
> w.ambari.ui.tests.installer.InstallHadoop/install/_28_2_6_2_Security_enabling_
> was_FAILED_due_to_exception/>
> 
> 
> Diffs
> -----
> 
>   ambari-server/pom.xml b5a9d49 
>   
> ambari-server/src/main/java/org/apache/ambari/server/stack/StackDirectory.java
>  d417a2f 
>   ambari-server/src/main/resources/role_command_order.json 870a537 
> 
> Diff: https://reviews.apache.org/r/41740/diff/
> 
> 
> Testing
> -------
> 
> mvn clean test
> 
> 
> Thanks,
> 
> Andrew Onischuk
> 
>

Reply via email to