Bright:Node image management

Node images are all stored in:

  chroot /cm/images/xxx

Edit the list of files not to be synchronised

  cmsh -c "category use default; set excludelistupdate; commit"

Update the Provisioning Nodes

• Only applicable with multiple head nodes

[root@head1 OpenFOAM]# cmsh
[head1]% softwareimage updateprovisioners
Provisioning nodes will be updated in the background.
[head1]% 
Tue May  6 12:47:03 2014 [notice] head1: Provisioning started: sending head1:/cm/images/default-image to head2:/cm/images/default-image

Sync an image without rebooting

[root@head-boston GPU-image]# cmsh
[head-boston]% device
[head-boston->device]% imageupdate -n gpu2 
Performing dry run (use synclog command to review result, then pass -w to perform real update)...
[head-boston->device]% 
Thu Jan 16 16:42:35 2014 [notice] head-boston: Provisioning started: sending head-boston:/cm/images/GPU-image to gpu2:/, mode UPDATE, dry run = yes, no data changes!
Thu Jan 16 16:42:40 2014 [notice] head-boston: Provisioning completed: sent head-boston:/cm/images/GPU-image to gpu2:/, mode UPDATE, dry run = yes, no data changes!
imageupdate -n gpu2 [ COMPLETED ]
[head-boston->device]% imageupdate -n gpu2 -w 
[head-boston->device]% 
Thu Jan 16 16:42:40 2014 [notice] head-boston: Provisioning started: sending head-boston:/cm/images/GPU-image to gpu2:/, mode UPDATE, dry run = no
[head-boston->device]% 
Thu Jan 16 16:42:45 2014 [notice] head-boston: Provisioning completed: sent head-boston:/cm/images/GPU-image to gpu2:/, mode UPDATE, dry run = no
imageupdate -n gpu2 -w [ COMPLETED ]

Sync the whole cluster

  imageupdate -n gpuNode01..gpuNode16 -w