Reply
Explorer
Posts: 21
Registered: ‎03-15-2016

Cloudera manager high availability does not work as expected

Hello

 

We have a CDH 5.13 cluster and want to configure Cloudera manager and management services hih avilability. I worked according to the official guide.

 

I used two server pairs, one for CM itself and one for the management services along with two HAProxy lod balancers, as mentioned in the docs.

 

I folowed the instructions thoroughly but could'nt make it work. The first problem is with this page, section 2-f. <MGMTHostname> is just not reported in the host list, although the agent is running with the right listening_address and the hosts file is properly configured.

 

I tried to ignore it and go on using the actual server names, but then the services are registered with the actual host name and not the load balancer name and if a failover occurs then I have to manually restart the management services so they willl update with the secondary CM host name.

 

It seems like our system is not in sync with the documentation. I tried to google this problem but found nothing valuable.

 

Is this a known issue ? Did I miss something ? How can I make it work ?

 

Thank you

 

Guy

New Contributor
Posts: 4
Registered: ‎09-11-2018

Re: Cloudera manager high availability does not work as expected

[ Edited ]

Meet the same issure on this step 2-c, any update?

When I try ping <MGMTHostname>, I just got mgmt1

Posts: 941
Topics: 1
Kudos: 219
Solutions: 118
Registered: ‎04-22-2014

Re: Cloudera manager high availability does not work as expected

@WitchDocotor,

 

The community needs more information in order to be able to assist.

 

Please share:

 

- what change you made to your /etc/hosts

- what command you ran (ping)

- on which host you were when you ran the command

- the result of the command

 

 

New Contributor
Posts: 4
Registered: ‎09-11-2018

Re: Cloudera manager high availability does not work as expected

on mgmt1 host in the /etc/hosts file and a new line 172.17.1.60 dn1.test.com 172.17.1.60 mgmt1.test.com mgmt1 172.17.1.61 mgmt2.test.com mgmt2 172.17.1.60 dn1.test.com where dn1.test.com is the host I want to use as MGMTHOSTNAME when I ping dn1.test.com on mgmt1 host I got mgmt1:~>ping dn1.test.com PING dn1.test.com (172.17.1.60) 56(84) bytes of data. 64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=1 ttl=64 time=0.040 ms 64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=2 ttl=64 time=0.043 ms 64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=3 ttl=64 time=0.043 ms not the same in the document.
New Contributor
Posts: 4
Registered: ‎09-11-2018

Re: Cloudera manager high availability does not work as expected

on mgmt1 host in the /etc/hosts file and a new line 172.17.1.60 dn1.test.com

 

172.17.1.60 mgmt1.test.com mgmt1
172.17.1.61 mgmt2.test.com mgmt2
172.17.1.60 dn1.test.com

 

where dn1.test.com is the host I want to use as MGMTHOSTNAME

when I ping dn1.test.com on mgmt1 host I got

mgmt1:~>ping dn1.test.com
PING dn1.test.com (172.17.1.60) 56(84) bytes of data.
64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=1 ttl=64 time=0.040 ms
64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=2 ttl=64 time=0.043 ms
64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=3 ttl=64 time=0.043 ms

 

 not the same in the document.

Posts: 941
Topics: 1
Kudos: 219
Solutions: 118
Registered: ‎04-22-2014

Re: Cloudera manager high availability does not work as expected

@WitchDocotor,

 

Please provide information that shows us what you expect to have happen and what is actually happening.

 

If dn1.test.com is what you want to use, the ping test shows the resolution as defined in /etc/hosts.

New Contributor
Posts: 3
Registered: ‎10-23-2017

Re: Cloudera manager high availability does not work as expected

in the document https://www.cloudera.com/documentation/enterprise/5-15-x/topics/admin_cm_ha_service.html#concept_cvw... 2.c it should be [root@MGMT1 ~]# ping MGMTHostname PING MGMTHostname (MGMT1 IP) 56(84) bytes of data. 64 bytes from MGMTHostname (MGMT1 IP): icmp_seq=1 ttl=64 time=0.034 ms 64 bytes from MGMTHostname (MGMT1 IP): icmp_seq=2 ttl=64 time=0.018 ms but In My environment mgmt1:~>ping dn1.test.com PING dn1.test.com (172.17.1.60) 56(84) bytes of data. 64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=1 ttl=64 time=0.040 ms 64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=2 ttl=64 time=0.043 ms 64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=3 ttl=64 time=0.043 ms I expect was mgmt1:~>ping dn1.test.com PING dn1.test.com (172.17.1.60) 56(84) bytes of data. 64 bytes from dn1.test.com (172.17.1.60): icmp_seq=1 ttl=64 time=0.040 ms 64 bytes from dn1.test.com (172.17.1.60): icmp_seq=2 ttl=64 time=0.043 ms 64 bytes from dn1.test.com (172.17.1.60): icmp_seq=3 ttl=64 time=0.043 ms
New Contributor
Posts: 4
Registered: ‎09-11-2018

Re: Cloudera manager high availability does not work as expected

in the document https://www.cloudera.com/documentation/enterprise/5-15-x/topics/admin_cm_ha_service.html#concept_cvw...
2.c
it should be

[root@MGMT1 ~]# ping MGMTHostname
PING MGMTHostname (MGMT1 IP) 56(84) bytes of data.
64 bytes from MGMTHostname (MGMT1 IP): icmp_seq=1 ttl=64 time=0.034 ms
64 bytes from MGMTHostname (MGMT1 IP): icmp_seq=2 ttl=64 time=0.018 ms

but In My environment

mgmt1:~>ping dn1.test.com
PING dn1.test.com (172.17.1.60) 56(84) bytes of data.
64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=1 ttl=64 time=0.040 ms
64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=2 ttl=64 time=0.043 ms
64 bytes from mgmt1.test.com (172.17.1.60): icmp_seq=3 ttl=64 time=0.043 ms

I expect was

mgmt1:~>ping dn1.test.com
PING dn1.test.com (172.17.1.60) 56(84) bytes of data.
64 bytes from dn1.test.com (172.17.1.60): icmp_seq=1 ttl=64 time=0.040 ms
64 bytes from dn1.test.com (172.17.1.60): icmp_seq=2 ttl=64 time=0.043 ms
64 bytes from dn1.test.com (172.17.1.60): icmp_seq=3 ttl=64 time=0.043 ms
Announcements