Re: Trouble during deploy

Mon, 04 Mar 2013 22:19:21 -0800 

Hello,

I tried stopping the Ambari server, then resetting, then starting it.
Did everything from scratch and this time, after clicking the Deploy button,
I am redirected to the Install, Start and Test page. Installation proceeds
but after a certain point, I am stuck.

Crawler51 9% Installing JobTracker
Crawler52 11% Installing HDFS Client
Crawler53 16% Installing MapReduce Client

I am getting the following from stdout:

warning: Could not retrieve fact fqdn
warning: Host is missing hostname and/or domain: crawler51
warning: Dynamic lookup of $service_state at 
/var/lib/ambari-agent/puppet/modules/hdp-hadoop/manifests/init.pp:161 is 
deprecated.  Support will be removed in Puppet 2.8.  Use a fully-qualified 
variable name (e.g., $classname::variable) or parameterized classes.
warning: Dynamic lookup of $service_state at 
/var/lib/ambari-agent/puppet/modules/hdp-hadoop/manifests/service.pp:74 is 
deprecated.  Support will be removed in Puppet 2.8.  Use a fully-qualified 
variable name (e.g., $classname::variable) or parameterized classes.
warning: Dynamic lookup of $service_state at 
/var/lib/ambari-agent/puppet/modules/hdp-hadoop/manifests/service.pp:83 is 
deprecated.  Support will be removed in Puppet 2.8.  Use a fully-qualified 
variable name (e.g., $classname::variable) or parameterized classes.
warning: Dynamic lookup of $ambari_db_server_host is deprecated.  Support will 
be removed in Puppet 2.8.  Use a fully-qualified variable name (e.g., 
$classname::variable) or parameterized classes.
notice: 
/Stage[1]/Hdp::Snappy::Package/Hdp::Snappy::Package::Ln[32]/Hdp::Exec[hdp::snappy::package::ln
 32]/Exec[hdp::snappy::package::ln 32]/returns: executed successfully
notice: 
/Stage[2]/Hdp-hadoop::Initialize/Configgenerator::Configfile[core-site]/File[/etc/hadoop/conf/core-site.xml]/content:
 content changed '{md5}aa21ba6ff20cc6766211e37e4f364395' to 
'{md5}4a8180bd03474a5be7e13a3530ab641a'
notice: 
/Stage[2]/Hdp-hadoop::Initialize/Configgenerator::Configfile[mapred-site]/File[/etc/hadoop/conf/mapred-site.xml]/content:
 content changed '{md5}864fa2060a7271cca6769742fdf00b16' to 
'{md5}ae167014591c96734bba8a438f805548'
notice: Finished catalog run in 1.55 seconds
My nodes do not have an FQDN since I have no other IP I can use for the domain. 

Thanks.

Dustine



On 3/5/2013 11:20 AM, Dustine Rene Bernasor wrote:

Hello Yusaku,
When I click the Deploy button,a loader gif appears (sometimes) but I am stuck in the same screen. 
I am not redirected to the Install, Start and Test page.
I will try to do the "ambari-server stop" first then reset then start and see if I still get the same problem. 
If I still get it, I might have to switch to 1.2.1 as you suggested.

By the way, I have attached the ambari-server log.

Thanks.

Dustine

On 3/5/2013 11:01 AM, Yusaku Sako wrote:

Hi Dustine,
What happens after you click on the Deploy button? It just gets stuck on the same screen? Or does it go to the "Install, Start and Test" page with progress bars? If you can post /var/log/ambari-server/ambari-server.log, it would be helpful to troubleshoot. 

Also, it sounds like you are using Ambari 1.2.0?
With 1.2.0, you should "ambari-server stop", followed by "ambari-server reset", then "ambari-server start" if deploy gets stuck. Clear the browser cache and hit http://<ambari-server>:8080. 

BTW, Ambari 1.2.1 handles retrying deploy much better than 1.2.0.
If deploy gets stuck for whatever reason, you can hit refresh on the browser and hit "Deploy" again (no need to do "ambari-server reset", etc). You will not get a message saying you already have a cluster with the same name, etc. I highly recommend trying out 1.2.1, rather than 1.2.0 (if you are not already). In addition to handling retries better, it has 136 fixes over 1.2.0: https://issues.apache.org/jira/issues/?jql=fixVersion%20%3D%20%221.2.1%22%20AND%20project%20%3D%20AMBARI 

Yusaku
On Mon, Mar 4, 2013 at 6:31 PM, Dustine Rene Bernasor <[email protected] <mailto:[email protected]>> wrote: 

    Hello,

    I am trying to deploy a Hadoop cluster with 3 nodes using Ambari.

    This is my set-up:

    HDFS
      NameNode: NodeA
      SecondaryNameNode: NodeA
      DataNodes: 2 hosts

    MapReduce
      JobTracker: NodeA
      TaskTracker: 2 hosts

    Nagios
      Server: NodeA

    Ganglia
      Server: NodeA

    However, after clicking the deploy button, the process seems to
    be stuck.

    I got something like this on the server log:

    
\"component\":\"JOBTRACKER\",\"hostName\":\"Crawler51\",\"serviceId\":\"MAPREDUCE\",\"isInstalled\":false},{\"display_name\":\"Nagios
    
Server\",\"component\":\"NAGIOS_SERVER\",\"hostName\":\"Crawler51\",\"serviceId\":\"NAGIOS\",\"isInstalled\":false},{\"display_name\":\"Ganglia
    
Collector\",\"component\":\"GANGLIA_SERVER\",\"hostName\":\"Crawler51\",\"serviceId\":\"GANGLIA\",\"isInstalled\":false}],\"slaveComponentHosts\":[{\"componentName\":\"DATANODE\",\"displayName\":\"DataNode\",\"hosts\":[{\"hostName\":\"Crawler52\",\"group\":\"Default\",\"isInstalled\":false},{\"hostName\":\"Crawler53\",\"group\":\"Default\",\"isInstalled\":false}]},{\"componentName\":\"TASKTRACKER\",\"displayName\":\"TaskTracker\",\"hosts\":[{\"hostName\":\"Crawler52\",\"group\":\"Default\",\"isInstalled\":false},{\"hostName\":\"Crawler53\",\"group\":\"Default\",\"isInstalled\":false}]},{\"componentName\":\"CLIENT\",\"displayName\":\"client\",\"hosts\":[{\"hostName\":\"Crawler52\",\"group\":\"Default\",\"isInstalled\":false},{\"hostName\":\"Crawler53\",\"group\":\"Default\",\"isInstalled\":false}]}]},\"AddHost\":{},\"AddService\":{}}}"}


    So after waiting for hours and hours, I tried to do it all over
    again. First I did a reset (ambari-server reset) on the Ambari host
    then did everything from scratch. When I reach the Deploy part,
    this time, I get a message that a cluster with the same name
    already exists.

    Here are my questions:
    1. What to do with the stuck deploy?
    2. How to remove the cluster that supposedly exist already? When
    I log in to Ambari, I am redirected to the install wizard.


    Thanks.

    Dustine

Reply via email to