ibm-spectrum-scale-install-infra icon indicating copy to clipboard operation
ibm-spectrum-scale-install-infra copied to clipboard

precheck ssh config (ssh_args) in ansible.cfg - help prevent intermittent hangs

Open mrolyat opened this issue 4 years ago • 0 comments

Some internal testing has revealed intermittent hangs when running playbooks. (currently only observed on RHEL7.6)

Small enhancement request to precheck config during install prechecks to see if ssh_args set to a value proved to help some node instances remove the intermittent hangs.

Am able to debug your hang issue ,  found issue with ssh connectivity .

I found one Ansible link and used this suggestion and able to install successfully on your cluster by changing one paramter in ansible.cfg file.   [https://stackoverflow.com/questions/51675831/ansible-stops-connecting-to-the-host-via-ssh]

Believe the connection may be dropping due to the lack of output from your play. Add the following to your ssh_args in ansible.cfg: -o ServerAliveInterval=50

I am able to create cluster successfully on your cluster environment using latest ansible code.  

GPFS cluster information
========================
  GPFS cluster name:         gpfs1.local
  GPFS cluster id:           6993155638008093391
  GPFS UID domain:           gpfs1.local
  Remote shell command:      /usr/bin/ssh
  Remote file copy command:  /usr/bin/scp
  Repository type:           CCR
GPFS cluster configuration servers:
-----------------------------------
  Primary server:    node-51.localnet.com (not in use)
  Secondary server:  (none)
 Node  Daemon node name       IP address   Admin node name        Designation
------------------------------------------------------------------------------
   1   node-51.localnet.com  10.0.100.51  node-51.localnet.com  quorum-manager-perfmon
   2   node-52.localnet.com  10.0.100.52  node-52.localnet.com  quorum-manager-perfmon
   3   node-53.localnet.com  10.0.100.53  node-53.localnet.com  quorum-manager-perfmon
   4   node-54.localnet.com  10.0.100.54  node-54.localnet.com  quorum-manager-perfmon
 

mrolyat avatar Mar 27 '20 18:03 mrolyat