What is the difference between Apache Ambari Server and Agent?
What is the role\tasks of the server vs Agent?
Ambari server collects informations from all Ambari clients and sends operations to clients (start/stop/restart a service, change configuration for a service, ...).
Ambari client sends informations about machine and services installed on this machine.
You have one Ambari server for your cluster and one Ambari agent per machine on your cluster.
If you need more details, Ambari architecture's is explained here
Related
I am learning about cloudera and came across Ambari agent that resides in each host that is part of a hadoop cluster. So while configuring/creating the cluster does Ambari agent generate the IP addresses for the hosts and send them to DNS or is my understanding completely wrong.
Thanks in advance :)
The agent reports the host information to the Ambari server, it doesn't manipulate anything outside of the Hadoop processes.
The IP & hostname of the nodes would already be assigned prior to installation of the agent
I am using Cloudera Express. The Cloudera Manager version is 5.12.0. I am trying to automate the bring-up of services like hdfs, hbase... I am able to do so by specifying necessary information of each service in host template, and pushing the host template to Cloudera Manager using curl command which uses Cloudera Manager API. Now, I want to automate the bring-up of Cloudera Management Services like host monitor, service monitor, event server, activity monitor and alert publisher. I have tried to do so by adding the corresponding role types and service types of each service in the host template. When I push the host template to Cloudera Manager using curl command, Cloudera Manager shows an error that It could not find service type 'MGMT' with version CDH 5.12.0. As the management services are different from cluster services like hdfs, yarn, hbase..., How should I automate the bring-up of management services? Is their a dedicated API to automate Management Services?
Unfortunately, the host template only applies to clusters not CM. To configure CM look at:
https://cloudera.github.io/cm_api/
https://cloudera.github.io/cm_api/apidocs/v17/index.html (in particular the /cm/service/* APIs)
Two of my drives crashed on the Ambari Server node so I have to re-migrate my Ambari Cluster. No real data was lost (due to a different backup strategy) but the configuration files of the node, including Ambari Server configuration, are gone.
Because two drives crashed, I can not access any files from that node anymore (RAID 5).
I am now in the process of reinstalling the Ambari Server on the same node and would like to have my agents seamlessly reconnect to the "new" Ambari Server.
Is there a way to migrate the existing Cluster settings to the Ambari Server? I am thinking of Cluster settings that were distributed to the agents or similar.
If there is no such way to migrate the cluster, how would I go and install the Ambari Server? Do a fresh install and setup everything again? Will the Ambari agents be able to connect to the "new" Cluster without problems? Note that the Ambari Server will run on the same hostname/ip.
Newbie w/ etcd/zookeeper type services ...
I'm not quite sure how to handle cluster installation for etcd. Should the service be installed on each client or a group of independent servers? I ask because if I'm on a client, how would I query the cluster? Every tutorial I've read shows a curl command running against localhost.
For etcd cluster installation, you can install the service on independent servers and form a cluster. The cluster information can be queried by logging onto one of the machines and running curl or remotely by specifying the IP address of one of the cluster member node.
For more information on how to set it up, follow this article
HDP cluster deployed successfully on AWS EC2. After restart of the HDP cluster nodes, heartbeat lost from ambari server as all Public and Private IP’s and DNS are changed.
Where in ambari server we can configure new IP’s or DNS ??
First, Ambari requires to have FQHN for all your nodes. It is best practice to assign proper hostnames on all your nodes.
A simple word-around for getting back the heartbeat on your Ambari server is to run the following on all your clients nodes:
sudo ambari-agent restart your_ambari.server.hostname.com
It worked for me on Ambari 2.0 and Ubuntu 12. Good luck!