AnsweredAssumed Answered

Incorrect name resolution in clush group

Question asked by reedv on Jun 5, 2018
Latest reply on Jun 8, 2018 by deborah

Using clustershell to communicate across our mapr cluster and running into problem where the program seems to be trying to resolve host names that are different from the names explicitly given to the program, eg.

Say I have a group file `/etc/clustershell/groups.d/local.cfg` looking like

[root@mapr001 hadoop-2.7.0]# cat /etc/clustershell/groups.d/local.cfg
all: mapr[001-006],mapr07
clients: mapr[07]


The results of using these groupings are as follows:

[root@mapr001 hadoop-2.7.0]# clush -b -g all date
mapr007: ssh: Could not resolve hostname mapr007: Temporary failure in name resolution
---------------
mapr[001-006] (6)
---------------
Fri May 18 08:45:40 HST 2018
clush: mapr007: exited with exit code 255

[root@mapr001 hadoop-2.7.0]# clush -b -g clients date
---------------
mapr07
---------------


As shown, we see that the all group is turning "mapr07" into "mapr007" which is not what was written in the config file (yet stil able to correctly contact mapr07 when using a group that contains only that node). Does anyone know what is going on here?

 

** I realize this is not a MapR specific question, but clush has been recommend to me on these forums as a useful cluster administration tool, so this seems like a relevant topic to have here and this does really affect the ability to propagate config files across the cluster including client nodes.

Outcomes