Troubleshooting
Problem
When your IBM Smart Analytics System is deployed no NODEGROUP or node list is configured. This is to be done by the System Administrator or person with similar role.
Symptom
DSH command using node list or node group options does not work
Executing commands below as user root on Management node of 7600 / 7700 you will receive messages shown
Example No. 1:
# dsh -a uptime
You receive the following message
dsh: 2617-011 No hosts in node list
Example No. 2:
# dsh -N BCUDATA uptime
You receive the following message
dsh: 2617-018 Cannot open host file /BCUDATA: A file or directory in the path name does not exist.
dsh: 2617-187 BCUDATA is not a valid DSH node group.
dsh: 2617-011 No hosts in node list
Cause
In Example No. 1 the message is because option "-a" is telling "dsh" to execute command "uptime" on all the listed or configured nodes in node list and there are none configured.
In Example No. 2 the message is because option "-N" is telling "dsh" to execute command "uptime" on all the nodes that are register to node group BCUDATA and there is no node group BCUDATA configured.
Environment
This document is applicable to IBM Smart Analytics System (BCU) 7600 and 7700
Resolving The Problem
For you to use node group options with a "dsh" command you first need to configure the node groups or list of nodes for DSH to use.
The DSH that comes with the IBM Smart Analytics System 7600 and 7700 is configured using a set of Operating System environment variables.
In this document we are using only a few of those environment variables to configure nodegroups and execute distributed command across the cluster
Here are the variables we are using:
DSH_NODE_RSH
- The value of this variable is the full path of the remote shell to use for remote command. You can use option "-r" to indicate the remote shell full path, if this option is used it will override the environment variable.
Note: If you do not define DSH_NODE_RSH then you need to use option "-r".
DSH_NODE_LIST
- The value of this variable is the full path of the file containing the list of node target. i.e. When using option "-a". Each line in the file represent a node.
DSH_NODEGROUP_PATH
- The value of this variable is the full path of the directory containing the files that define each NODEGROUP. The name of each file represent a NODEGROUP you can use with option "-N".
Each line in the file contain a hostname or IP address representing a node from that group. File may also contain blank and comment lines.
Note: To define the nodes list you can use both DSH_NODEGROUP_PATH and DSH_NODE_LIST or either one. If both are used "dsh" will merge both list eliminating all duplicates. To define node groups you need to use DSH_NODEGROUP_PATH configuration. Use DSH_NODE_LIST to define nodes that are not defined in any node group.
Note: To make this definition persistent for the "root" on the Management node you may add them to root's ".profile".
See the following example:
Define node groups BCUADMIN, BCUDATA and BCUSTDBY and execute commands using node group option "-N".
1. Login as root to your Management node
2. Define your environment variables
- export DSH_NODE_RSH=/usr/bin/ssh
export DSH_NODEGROUP_PATH=/var/opt/csm/node_group
3. Create/Edit your files to define each node group
- BCUADMIN contains the hostnames for the Admin nodes admin01 and admin02. In this example file we are using comment "#" lines to document the entries.
Edit/create file $DSH_NODEGROUP_PATH/BCUADMIN with the following lines
- BCUDATA contains Data nodes data01 to data08. In this example file we are using comment "#" lines to document the entries.
- BCUSTDBY contains Standby nodes stdby011 and stndby02. In this example file we are using comment "#" lines to document the entries.
# Node gruop BCUADMIN represent all Admins nodes defined in the cluster
# The first Admin node (Catalog) in the cluster
admin01
# Second Admin node (a second cordinator node) in the cluster
admin02
Edit/create file $DSH_NODEGROUP_PATH/BCUDATA with the following lines
# Nodegruop BCUDATA represent all Data nodes defined in the cluster
data01
data02
data03
data04
data05
data06
data07
data08
Edit/create file $DSH_NODEGROUP_PATH/BCUSTDBY with the following lines
# Nodegruop BCUSTDBY represent all the Standby nodes defined in the cluster
# The first Standby node (first HA group) in the cluster
stdby01
# Second Standby node (second HA group) in the cluster
stdby02
i. The following command execute to "uptime" on all the Admin nodes
- # dsh -N BCUADMIN uptime
admin01: 12:55PM up 55 days, 19:08, 1 user, load average: 0.94, 0.99, 0.94
admin02: 12:55PM up 55 days, 19:08, 1 user, load average: 1.12, 1.16, 1.18
- # dsh -a date | dshbak
HOST: admin01
-----------------
Fri Sep 9 13:02:46 CDT 2011
HOST: admin02
-----------------
Fri Sep 9 13:02:50 CDT 2011
HOST: data01
-----------------
Fri Sep 9 13:02:46 CDT 2011
HOST: data02
-----------------
Fri Sep 9 13:02:46 CDT 2011
HOST: data03
-----------------
Fri Sep 9 13:02:50 CDT 2011
HOST: data04
-----------------
Fri Sep 9 13:02:46 CDT 2011
HOST:data05
-----------------
Fri Sep 9 13:02:50 CDT 2011
HOST: data06
-----------------
Fri Sep 9 13:02:46 CDT 2011
HOST: data07
-----------------
Fri Sep 9 13:02:46 CDT 2011
HOST:data08
-----------------
Fri Sep 9 13:02:50 CDT 2011
HOST: stdby01
-----------------
Fri Sep 9 13:02:46 CDT 2011
HOST: stdby02
-----------------
Fri Sep 9 13:02:46 CDT 2011
Was this topic helpful?
Document Information
Modified date:
16 June 2018
UID
swg21515254