Zookeeper Timeout while Data Fabric Installation from Scratch

[msaidbilgehan]

We are installing Data Fabric from scratch to 6 nodes (Installer node also in it). While installing, all nodes gives ZK timeout. No firewall is enabled in nodes or the cluster network. No application binding any ports. All nodes are Ubuntu 20.04 LTS default installation and nothing more. Here is the SS and logs;

• Logs: https://drive.google.com/open?id=1qZv5Ue4Vv-VjItrFGW3n00tO1D0AF-bT&usp=drive_fs
• Node Naming:
•        node[0-4].cluster
•        installer.cluster
• Cluster Naming: my.cluster1.com
• No External IP
• Secure Cluster (Installer default option)
• SSH Installation passwordless sudo (root and mapr user) 

[Dave Olker]

Looking at the installer log, the nodes in your cluster are:

            "all": [

                "node0.cluster",

                "installer.cluster",

                "node4.cluster",

                "node1.cluster",

                "node3.cluster",

                "node2.cluster"

            ],

            "cldb": [

                "installer.cluster",

                "node0.cluster",

                "node1.cluster",

                "node3.cluster",

                "node4.cluster"

            ],

            "es": [],

            "zk": [

                "installer.cluster",

                "node0.cluster",

                "node1.cluster",

                "node3.cluster",

                "node4.cluster"

            ]

Are those really the fully-qualified domain names for these nodes, and your hostname resolution provider (DNS, NIS, LDAP, etc.) can resolve these hostnames->IP and IP->hostname successfully?  

I'd check for any logfiles specific to the zookeeper under /opt/mapr/zookeeper/zookeeper-3.5.6/logs that you can check for errors.

[msaidbilgehan]

@Dave Olker Those are only local ip-hostname resolutions that only written to hosts file to access IPs over hostname. We didn't add these configurations to DNS servers because still working to get stable installation configurations for cluster. So it is basically a test cluster in our servers that every machine has each others IP and hostname in their hosts and hostname files respectively.