Installation path A - Cloudera Manager Server failed to start

[#1]

Hi all,

I have started a new local setup of the Cloudera Manager Installer.

It is running on a Oracle Linux Release 6.6 (kernel 2.6.39), inside a VirtualBox.

All goes well, up until the "Starting Cloudera Manager Server".

Errors shown are :

Cloudera Manager Server failed to start.
See /var/log/cloudera-manager-installer/6.start-scm-server.log for details.

Unable to remove cloudera-manager-daemons.
It will need to be removed manually with rpm.
See /var/log/cloudera-manager-installer/11.remove-cloudera-manager-daemons.log for details.

So I checked the logs:

--> 6.start-scm-server.log:
--------------------------------
Starting cloudera-scm-server: ^[[60G[^[[0;31mFAILED^[[0;39m]
~                                                                                                                               
"6.start-scm-server.log" [dos] 1L, 59C

Not much there, checked next log:

--> cloudera-scm-server.log:
----------------------------
usr/sbin/cmf-server: line 30: /usr/share/cmf/bin/cmf-server: No such file or directory
/usr/sbin/cmf-server: line 30: exec: /usr/share/cmf/bin/cmf-server: cannot execute: No such file or directory

This is indeed a missing folder.

Can anyone assist in solving this installation issue ?

I have followed the steps to remove the installation files, and tries reinstalling, with no success.

Regards,

Ken.

[#2 · Accepted solution]

Got this working as follows:

1 clean up old install files and srevices (followed the staps from cloudera document)

2 reboot the machine !

3 reinstall the whole package from sratch

[#3]

Hi All,

I found an extra erro rin the logs:

error: Failed dependencies:
    cloudera-manager-daemons = 5.2.0 is needed by (installed) cloudera-manager-agent-5.2.0-1.cm520.p0.60.el6.x86_64

It seems that the cloudera-manager-daemons cannot be uninstalled, because of a dependency.

The cloudera-manager-agent service is indeed installed, but it is not running and dead.

Any idea on how to clean up the installation and restart installing ?

Regards,

Ken.

[#4]

I've managed to clean the old installation files, and restart a new install, via both installation path A and B it stops at this point:

/usr/share/cmf/bin/initialize_embedded_db.sh: line 360: cd: /var/lib/cloudera-scm-server: No such file or directory
hostname: Unknown host
Creating SCM configuration file: /etc/cloudera-scm-server/db.properties
waiting for server to start...sh: /var/log/cloudera-scm-server/db.log: Permission denied
............................................................could not start server
Could not start database server, giving up

Some tips would be welcome.

Ken.

[#5 · Accepted solution]

Got this working as follows:

1 clean up old install files and srevices (followed the staps from cloudera document)

2 reboot the machine !

3 reinstall the whole package from sratch

[#6]

And now for the bad news (sorry I just saw your thread);  

RHEL/OL/CentOS 6.6 is not yet certified for use with CM/CDH.  OS release level testing is co-ordinated with new releases of Cloudera products, so 6.6 is still in planning to be aligned with future release testing.  

For playing around with / POC you can keep going, but I would not reccomend this in a production environment that you expect to be supported. 

You will have issues with SSL/TLS if you attempt to configure it for CM/Agent interaction, due to a python / openSSL issue that appears to be present in 6.6

We strongly reccomend that users with the intent to use services in a production environment stick to the listed requirements, which OS release is included, from here:

http://www.cloudera.com/content/cloudera/en/documentation/core/latest/topics/installation_reqts.html

There is nothing more painful than getting far down the integration/setup/testing path and discovering a core requirement like OS is not being met, is the cause of a problem you are running into.  

Todd

[#7]

Hello Todd,

Thanks for you reply.

It is a learning environment for me, so no production data :-)

Regards,

Ken.

[#8 · Accepted solution]

Hello,

I reinstalled a new VM, with CentOS this time.

Made sure the firewall is off, and selinux is disabled. Checked the requirements and made sure they where met.

Got up to the point of the installation failing to receive a heartbeat.

Error found in the logs:

ERROR    Heartbeating to 127.0.0.1:7182 failed.
Traceback (most recent call last):
  File "/usr/lib64/cmf/agent/src/cmf/agent.py", line 815, in send_heartbeat
    self.master_port)
  File "/usr/lib64/cmf/agent/build/env/lib/python2.6/site-packages/avro-1.6.3-py2.6.egg/avro/ipc.py", line 464, in __init__
    self.conn.connect()
  File "/usr/lib64/python2.6/httplib.py", line 720, in connect
    self.timeout)
  File "/usr/lib64/python2.6/socket.py", line 567, in create_connection
    raise error, msg
error: [Errno 111] Connection refused

Apparently the hostname and hostname -f returned a diferent name.

So I had to align them.

Updated both /etc/hosts and /etc/sysconfig/network files to have the same hostname (without capitals.)

Restart network, and reboot the VM

Retried the adding of the host, this time with success.

Hope this helps some other troubled users installing !

Regards,

Ken.

[#9]

I'll be attempting the exact same thing tonight.

Ubuntu 14.04 doesn't seem to play particularly nicely. I had strange references to 127.0.1.1 even though I had removed that from the /etc/hosts file. (Just one node).

[#10]

Hi,

Good luck tonight ...

Maybe this can help, the configur

[#11]

Sorry,

My last post got submitted too early 🙂

This is the content of my etc/hosts file:

127.0.0.1   centos.oraclevirtualbox centos localhost localhost.localdomain
::1         centos.oraclevirtualbox centos localhost localhost.localdomain
10.0.2.15   centos.oraclevirtualbox centos localhost.localdomain localhost

And the etc/sysconfog/network file:

NETWORKING=yes
HOSTNAME=centos.oraclevirtualbox

Hope this can help you solve the issue !

Good luck,

Ken.