topic Re: SAS LASR Monitor Process is failing in SAS Visual Analytics

SAS LASR Monitor Process is failing

sandeep_reddy — Wed, 17 Jun 2020 02:58:54 GMT

Hi Team,

I am getting below error while starting LASRMonitor process from SAS Compute node and the error details are below.

java.io.IOException
at com.sas.grid.broker.monitor.ConnectionManager.startTKGridMon(ConnectionManager.java:247)
at com.sas.grid.broker.core.BrokerCore.main(BrokerCore.java:267)

[ERROR@2020-06-17 11:42:41.618] Monitor thread failed to start due to configuration errors.
[ERROR@2020-06-17 11:42:41.619] No I/O produced by the grid monitor, check your SSH key configuration

I have resolved all SSH related issues and i am able to login to main node without password and vice versa to compute node from main node.

Please help me to resolve this issue or any help to understand more about this issue.

Best Regards,
Sandeep

Re: SAS LASR Monitor Process is failing

AnandVyas — Wed, 17 Jun 2020 03:14:15 GMT

Hi @sandeep_reddy

The error seems to be related to SSH keys. How did you set them up? Its recommended to use HPA console. https://documentation.sas.com/?docsetId=hpaicg&docsetTarget=n03011sashpanalytics0install.htm&docsetVersion=3.9&locale=en

When you say compute, you mean the TkGrid Nodes? You should verify the SSH connectivity from head node (main node where VA is installed) to all the grid nodes and vice versa.

Try this to validate:

Open a command line on the SAS compute server (the system where SAS code is to be executed, such as the Base SAS^® system or the WorkspaceServer system). Change to the user that is encountering errors of the sort described above.
Use SSH to log on to the SAS High-Performance Analytics/TKGrid head node. Do this even if the head node is the same system as the SAS compute server. SAS performs an SSH login regardless of whether TKGrid is co-located with the compute server system.

ssh node0.example.com

You should authenticate without any prompt for a password.

Now, on the SAS High-Performance Analytics/TKGrid head node, run the command below. Edit only the path to TKGrid as necessary in your environment. The simsh command is executed twice and hostname is a command provided to simsh. This runs a nested SSH loop that connects from every TKGrid node to every TKGrid node.

/opt/TKGrid/bin/simsh /opt/TKGrid/bin/simsh hostname

This also needs to complete without any errors or password prompts.

Successful output looks similar to this:

node0: node2: node2.example.com
node0: node3: node3.example.com
node0: node1: node1.example.com
node0: node0: node0.example.com
node1: node2: node2.example.com
node1: node3: node3.example.com
node1: node1: node1.example.com
node1: node0: node0.example.com
node2: node2: node2.example.com
node2: node3: node3.example.com
node2: node1: node1.example.com
node2: node0: node0.example.com
node3: node2: node2.example.com
node3: node3: node3.example.com
node3: node1: node1.example.com
node3: node0: node0.example.com

Re: SAS LASR Monitor Process is failing

sandeep_reddy — Wed, 17 Jun 2020 04:01:03 GMT

Hi Anand,

Thanks for the information.

I have set SSH keys manually and copied to all nodes.

I am getting the expected SSH output without any errors ,after running /opt/TKGrid/bin/simsh /opt/TKGrid/bin/simsh hostname.

Attached is the latest error after adding GRIDSHCOMMAND in sasv9_local.cfg file.

Best Regards,
Sandeep

Re: SAS LASR Monitor Process is failing

AnandVyas — Wed, 17 Jun 2020 04:36:03 GMT

Hi @sandeep_reddy

The error shared by you is clearly highlighted in this KB article. https://support.sas.com/kb/60/126.html

If you are able to validate SSH connectivity from each node to every other node in cluster, I am not sure what's causing the issue. Would suggest to raise a SAS Technical support track by emailing the issue with error snaps and steps taken to support@sas.com