[gridengine users] need to take down master due to root filesystem issues

Nicolas Joly njoly at pasteur.fr
Thu Sep 18 06:38:25 UTC 2014

On Thu, Sep 18, 2014 at 12:01:53AM -0500, Dan Hyatt wrote:


> The reason I am writing is I want to manual fail over to my shadow
> master. I was unable to find a best practices. I know that just
> taking down the master should failover to shadow master first on the
> list. But I would rather do it cleanly. Also, how do I fail back to
> my original master, or does it do it automatically once it is back
> online?

You can migrate the qmaster by hand. Pickup a shadow host and just use
the -migrate option from the startup script :

#       -migrate    shuts down qmaster if it is running
#                   on another host and start the daemons on this host

  /etc/init.d/sgemaster -migrate

We used it many times and it worked flawlessly.

Nicolas Joly

Biology IT Center
Institut Pasteur, Paris.

More information about the users mailing list