Installation failed. Failed to receive heartbeat from agent.

[#1]

I am trying to install Cloudrea Manager Standard edition and CHD4 with parcels. This is being installed on three Dell machines running Ubuntu 12.04.2 LTS 64 bit. I am receiving an error on all three machines:

    Ensure that the host's hostname is configured properly.

    Ensure that port 7182 is accessible on the Cloudera Manager server (check firewall rules).

    Ensure that ports 9000 and 9001 are free on the host being added.

    Check agent logs in /var/log/cloudera-scm-agent/ on the host being added (some of the logs can be found in the installation details).

I checked that the hostname is configured.

I checked that Ports 7182, 9000 and 9001 are free (I am guessing that Cloudera is using 9000 & 9001 for python because these ports are in use after the install fails but not before the install).

sudo netstat -tulpn

Active Internet connections (only servers)

Proto Recv-Q Send-Q Local Address           Foreign Address         State       PID/Program name

tcp        0      0 0.0.0.0:6000            0.0.0.0:*               LISTEN      1104/X

tcp        0      0 127.0.0.1:53            0.0.0.0:*               LISTEN      1123/dnsmasq

tcp        0      0 0.0.0.0:22              0.0.0.0:*               LISTEN      655/sshd

tcp        0      0 127.0.0.1:631           0.0.0.0:*               LISTEN      792/cupsd

tcp        0      0 127.0.0.1:6010          0.0.0.0:*               LISTEN      29403/0

tcp6       0      0 :::6000                 :::*                    LISTEN      1104/X

tcp6       0      0 :::22                   :::*                    LISTEN      655/sshd

tcp6       0      0 ::1:631                 :::*                    LISTEN      792/cupsd

tcp6       0      0 ::1:6010                :::*                    LISTEN      29403/0

udp        0      0 127.0.0.1:53            0.0.0.0:*                           1123/dnsmasq

udp        0      0 0.0.0.0:68              0.0.0.0:*                           1073/dhclient

udp        0      0 0.0.0.0:36051           0.0.0.0:*                           790/avahi-daemon: r

udp        0      0 0.0.0.0:5353            0.0.0.0:*                           790/avahi-daemon: r

udp6       0      0 :::50807                :::*                                790/avahi-daemon: r

udp6       0      0 :::5353                 :::*                                790/avahi-daemon: r

And I checked the firewalls and they are open

sudo iptables -L

Chain INPUT (policy ACCEPT)

target     prot opt source               destination

Chain FORWARD (policy ACCEPT)

target     prot opt source               destination

Chain OUTPUT (policy ACCEPT)

target     prot opt source               destination

Lastly I checked the log files in /var/log/cloudera-scm-agent/ and they do not show any errors  and there was only one warning, that the default socket timeout was set to 30.

Can anyone point me to what the possible problem is? We are looking at using Hadoop in one of our solutions and are trying to evaluate it before purchasing the Enterprise version. I cannot use a cloud version because of data restrictions put on my by the data vendor and client so I need to have an internal sandbox to get an idea of what we need to develop and what we will need to support. Thanks!

[#2 · Accepted solution]

@enelso:  Your hostname needs to be tied to an actual IP address on your local network which can send/receive traffic between all the hosts.  The address you have associated your hostname with is the loopback address, which cannot route actual network traffic off the host.

Use "ifconfig -a" to see a listing of your network interfaces and choose one that has an actual IP address. 

[#3]

Clink, I got the same error, and I changed my /etc/hosts file as: 52.60.239.41 ip-172-31-12-19.ca-central-1.compute.internal node1

also, I restart agent on each hosts, 4 of them, I still get the same error.

please help.

Robin

[#4]

I ran into similar problems and finally figured out what was causing the issue.

All my hosts / hostname etc. settings were correct, ports available etc., but the thing that caused the heartbeat to fail.....was that NTPD service has to be running on all hosts in order to "synchronize" the heartbeat.

No sync == no heartbeat.

Can't believe that did not occur to me sooner. Hope this helps someone.

(It may be helpful to add this to the list of possible causes of failing installation, that is shown upon failure)

[#5]

@Arjen,

Thank you for sharing your solution.

Just so there is no confusion, though, the agent does not require ntpd to heartbeat.  By default, however, the agent does run healthchecks that test ntpd or chronyd to ensure that offset is within expected parameters.

While failure of the healthchecks will result in a host being shown in bad health, it is not a functional characteristic of the health check to prevent heartbeats.

If adding ntpd was the only action that allowed the agent to heartbeat, then it could be there was some problem where the agent was not able to progress with the heartbeat due to an unusual condition.  If you or anyone hits a problem like this where the agent cannot heartbeat, let us know again and we will take a closer look.

My reason for explaining this was to make sure everyone was aware that NTP is not required for heartbeating.  The root cause may have been related to NTP, but it was not an intended feature limitation.

Thanks again, for sharing.

Cheers.

Ben

[#6]

Hello!

I bealive I have similar (or identical) problem but I don't fully understand explanation. I'm installing cluster on two virtual machines with Ubuntu 12.04 on them. Both have two network interfaces eth0 and eth1. eth0 is connection with internet through NAT and eth1 is virtual network between them. 

Host 1 name is Hadoop1.local and its IP (eth1) is: 192.168.56.101.

Host 2 name is Hadoop2.local and its IP (eth1) is: 192.168.56.102

I start Cloudera Manager instalator on Hadoop1. In step "" I type both IP addresses (.101 & .102) (Is it OK or should I only start installation on second host- Hadoop2). During "Cluster Installation" step on Hadoop1 I get error:

"Installation failed. Failed to receive heart beat from agent".

I check ports 7182, 9000, 9001 and those are OK (before I start installation those are empty, after installation I have python on them).

In /var/log/cloudera-scm-agent I see no errors.

So I assume there is something wrong with my hosts names.

My /etc/host looks like this (Hadoop1):

127.0.0.1 localhost loopback

192.168.56.102 Hadoop2.local

And second /etc/host like this (Hadoop2):

127.0.0.1 localhost loopback

192.168.56.101 Hadoop1.local

Any help would be appreciate!

Regards

Andrzej

[#7]

Hello!

Any idea how to solve above problem? Any help would be appreciate!

Regards

Andrzej

[#8 · Accepted solution]

Hello,

  I apologize for the delay.  I think there may be a couple of things going on.  For starters, you should add your own hostname and IP address to the /etc/hosts file on each machine.  In other words, both Host1 and Host2 entries should be in the /etc/hosts file on both machines.  Also, have you checked to see if iptables is running?  That is a firewall app that can stop traffic between nodes.  To identify if iptables is running and disable it, do this (as root):

$ sudo chkconfig iptables --list
iptables 0:off 1:off 2:off 3:off 4:off 5:off 6:off

If you see iptables as "on" for any of those values (especially #5), than it's probably getting in your way, you should disable it unless you've got a company policy requiring it be enabled.  To disable it, run this command:

sudo chkconfig iptables off

[#9]

Hi,

Thank You for Your response. I think I had something wrong in my /etc/hosts. Also on one node I had ufw enabled. Following command do the work:

sudo ufw disable

 Thank You for Your support!

Regards

Andrzej

[#10]

I am facing the same exact problem not sure if it is related or not but seems very very close. 

I am trying to install CDH5 on a single node cluster. .. so everything is supposed to be running on the same host ( cloudera manager, and all other hadoop and spark services), more of a sandbox than anything else,

I am running ubuntu 64 bit precise OS.  

I am using cloudera manager admin to install and and cluster installation step it fails: 

 Installation failed. Failed to receive heartbeat from agent.

• Ensure that the host's hostname is configured properly.
• Ensure that port 7182 is accessible on the Cloudera Manager Server (check firewall rules).
• Ensure that ports 9000 and 9001 are free on the host being added.
• Check agent logs in /var/log/cloudera-scm-agent/ on the host being added (some of the logs can be found in the installation details).

Thanks, 

Mukul

[#11]

@clint,  I tried picking the IP from ifconfig -a command and added that to hosts file as below

<IP><FQDN (at that one came in ifconfig -a)><shortname>

I used hostname and hostname -f to find the shortname and FQDN.  I still get the following error during installation 

Installation failed. Failed to receive heartbeat from agent.

• Ensure that the host's hostname is configured properly.
• Ensure that port 7182 is accessible on the Cloudera Manager server (check firewall rules).
• Ensure that ports 9000 and 9001 are free on the host being added.
• Check agent logs in /var/log/cloudera-scm-agent/ on the host being added (some of the logs can be found in the installation details).

I have enabled all TCP ports and ICMP ports.  I am not sure what am I missing.  Could you please help should I be checking something else?

[#12]

Hi,

I'm exactl facing the same issue:

can some one please help me to resolve this issue, I have following information in my /etc/hosts file

127.0.0.1  localhost localhost.localdomain localhost4 localhost4.localdomain4

::1            localhost localhost.localdomain localhost6 localhost6.localdomain6

192.168.42.101 n1.example.com n1

192.168.42.102 n2.example.com n2

192.168.42.103 n3.example.com n3

I can see successful installation message in n1 and above failure message in rest of two nodes.

Please help me