On BDA V4.7/V4.8/V4.9 Cloudera Manager Reports Connectivity Issues Between Nodes with the IPoIB Interface Down- IP Ping Fails / IBPING Succeeds (Doc ID 2303802.1)

Last updated on SEPTEMBER 03, 2017

Applies to:

Big Data Appliance Integrated Software - Version 4.7.0 to 4.9.0 [Release 4.7 to 4.8]
Linux x86-64

Symptoms

After upgrading to BDA V4.7, V4.8 or V4.9 connectivity issues between nodes are intermittently seen in Cloudera Manager(CM). This may result in CM alerts and services going down. It may also result in missing data blocks.

At the same time as the  connectivity issues it is observed that the IPoIB interface is down as observed by ip ping between nodes  failing while ibping is successful.  The external EoIP interface is up. Nodes can self-ping with ip ping but ip ping between nodes fails. The IB fabric is confirmed to be fully healthy.

Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms