Support Questions
Find answers, ask questions, and share your expertise
Announcements
Alert: Welcome to the Unified Cloudera Community. Former HCC members be sure to read and learn how to activate your account here.

CentOS 7 support for ambari-repo

CentOS 7 support for ambari-repo

Greetings,

 

This is my 1st install of CDH 5.1, I'm using the conf.pseudo. I have all of the hadoop essentials installed, configged to use the conf.pseudo environment and can start/stop gracelfully. The server is running centOS 7, jdl 1.7, python 2.7, all the new version.

 

I'm not sure if I need it, but I attempted to install the ambari-server using the CentOS 6 ambari-repo, since a file isn't available for centOS 7. The install  failed because it requires python 2.6:

 

root:/root->yum install ambari-server
Loaded plugins: fastestmirror, langpacks
Loading mirror speeds from cached hostfile
 * base: centos.chi.host-engine.com
 * extras: repos.dfw.quadranet.com
 * updates: mirror.keystealth.org
Resolving Dependencies
--> Running transaction check
---> Package ambari-server.noarch 0:1.6.1-98 will be installed
--> Processing Dependency: python(abi) = 2.6 for package: ambari-server-1.6.1-98.noarch
--> Finished Dependency Resolution
Error: Package: ambari-server-1.6.1-98.noarch (Updates-ambari-1.6.1)
           Requires: python(abi) = 2.6
           Installed: python-2.7.5-16.el7.x86_64 (@anaconda)
               python(abi) = 2.7
               python(abi) = 2.7
 You could try using --skip-broken to work around the problem
 You could try running: rpm -Va --nofiles --nodigest

I attempted to install python 2.6.9 using source but there were too many errors during make and make test.

 

Is there an update ambari-repo file coming out for CentOS 7 which will support python 2.7?

 

Thanks

 

Kevin

8 REPLIES 8

Re: CentOS 7 support for ambari-repo

Master Collaborator

Kevin,

 

Curious: have you consider using Cloudera Manager instead of Ambari? Things will work a lot more seamlessly that way.

Re: CentOS 7 support for ambari-repo

Sure, how do I install that part? I'm new to hadoop.

Re: CentOS 7 support for ambari-repo

Master Collaborator
It's really quite easy. If you download Cloudera Express (which is free),
you can use the included Cloudera Manager to install and deploy a CDH
cluster with ease:

http://www.cloudera.com/content/support/en/downloads/cloudera_manager/cm-5-1-1.html

There are also two other options:

1. Try Hadoop in the cloud via cloudera.com/live. There is a full CDH
cluster + sample data waiting there for you to use.
2. Download the QuickStart VM (cloudera.com/quickstartvm), which contains a
CDH cluster, Cloudera Manager, and guest OS.

Re: CentOS 7 support for ambari-repo

same error basically as I got for ambari:

Your distribution is not supported. This installer supports: │
                                     │ * Redhat Enterprise Linux 5 or 6                             │
                                     │ * CentOS 5 or 6                                              │
                                     │ * Oracle Enterprise Linux 5                                  │
                                     │ * SUSE Linux Enterprise Server 11 (SP1 or later)             │
                                     │ * Ubuntu 10.04 LTS or 12.04 LTS                              │
                                     │ * Debian 6.0 or 7.0                                          │
                                     │──────────────────────────────────────────────────────────────│


I'll just use what I have until CentOS 7 is supported.


For dev purposes, I'm using a pseudo config, 1 server to mimic a cluster. I now have all but zookeeper and hbase starting automatically.



I'm not getting these 2 failed services to start at all, are they needed? They were part of the CDH 5.1 install packages.

root:/root->hadchk
Checking for running Hadoop Services now
JMX enabled by default
Using config: /etc/zookeeper/conf/zoo.cfg
Mode: standalone
Hadoop namenode is running                                 [  OK  ]
Hadoop datanode is running                                 [  OK  ]
Hadoop secondarynamenode is running                        [  OK  ]
Hadoop resourcemanager is running                          [  OK  ]
Hadoop nodemanager is running                              [  OK  ]
Hadoop httpfs is running                                   [  OK  ]
Hadoop jobtracker is not running                           [FAILED]
Hadoop historyserver is not running                        [FAILED]
Flume NG agent is running                                  [  OK  ]
Spark master is running                                    [  OK  ]
Spark worker is running                                    [  OK  ]
WEBHCat server is running                                  [  OK  ]








Highlighted

Re: CentOS 7 support for ambari-repo

Master Collaborator

True, CentOS 7 is not yet supported (by any distro, AFAIK). Why not just use a VM?

Re: CentOS 7 support for ambari-repo

Well I bit the bullet and downgraded to CentOS 6.5. I'm glad I did because the cloudera manager install ran successfully and except for a couple of config issues which I can't seem to correct, everything is running.?

this config I can't find to change. I don't how or why it would allow an overcommit of memory.



Memory on host cluster-01.kevinsdev is overcommitted. The total memory
allocation is 57.1 GiB bytes but there are only 62.9 GiB bytes of RAM
(12.6 GiB bytes of which are reserved for the system). Visit the
Resources tab on the Host page for allocation details. Reconfigure the
roles on the host to lower the overall memory allocation. Note: Java
maximum heap sizes are multiplied by 1.3 to approximate JVM overhead.

I even configured it to use my local postgres db instead of the embedded one.


I only have 1 server so how do I avoid these other 2 "errors"?

Service hdfs has 1 DataNode. Cloudera suggests at least 3 DataNodes for HDFS.
Service zookeeper has 1 Server. Cloudera suggests at least 3 Servers for ZooKeeper.


btw, I'm not a vmware person so if that's needed for whatever reason, please let me know.


Re: CentOS 7 support for ambari-repo

> Service hdfs has 1 DataNode. Cloudera suggests at least 3 DataNodes for
HDFS.

The default number of replicas for a block in HDFS is 3 for which you'd
need three hosts/datanodes. So this warning will remain unless you manually
lower the replication factor to 1

To do this, you can log on to Cloudera Manager -> HDFS service ->
configuration -> look for "replication factor". Set the value to 1, save,
then restart HDFS.

> Service zookeeper has 1 Server. Cloudera suggests at least 3 Servers for
ZooKeeper.

​To provide high-availability, it is recommended that you have at least
three (an odd) number of zookeeper instances. With a one node cluster, you
can't do that so this warning can be ignored.​


Regards,
Gautam Gopalakrishnan

Re: CentOS 7 support for ambari-repo

Thanks, I'll just leave the warnings as is. I made that change and restarted but the error persists. the rest of them I can ignore, hopefully you guys will add some extra config to raise/lower tolerances to 'bad health warnings'. They're annoying to look at when they are not really a problem.

Not sure if this is a bug but it's annoying as well. When selecting the Web UI link from the various components, for example HBASE, the URL is pre-pended by the clustername, so when clicking the link, it fails until I remove the cluster-01 part. It happens for all the web UI links.


http://cluster-01.kevinsdev:60010/master-status


Not sure yet how much time I'll need to view those pages once development starts.