[GE users] sge_save_config.sh script and Cloned Cluster Config.
crei at sun.com
Mon Jan 4 09:07:16 GMT 2010
> To group,
> Below is our current SGE environment and what we want to migrate too. We want to retain both the current and new SGE environments. That is, we want them to run concurrently for a period of time. Therefore, both the Solaris and new Linux Qmaster/shadow servers will manage jobs to the same set of Linux cluster nodes.
> Current environment
> - Two Sun Solaris 10 x86 AMD Qmaster and Shadow servers.
> - We are using SGE 6.1u4.
> - All of our SGE exec clients are RedHat Linux systems.
> - This is our current production environment.
> Planned new SGE environment
> - Ultimately replace our Sun servers with two Intel x64 RedHat Linux 5u4 Qmaster and Shadow servers.
> - I have downloaded SGE 6.2u5 binaries for Linux.
> - All of our current SGE exec clients will eventually be managed by new SGE servers.
> 1. I've read the "sge_save_config.sh" script can save my current SGE 6.1u4 configuration so that I can import it into my new SGE 6.2u5 environment. Do you have run this script on the Qmaster or Shadow server or can I run this script on any of the Linux cluster nodes that have access to SGE_ROOT?
> 2. This question follows from the first one. Since I'm going from a Solaris based SGE servers to new Linux based SGE servers, are there any issues with using the "sge_save_config.sh" script when creating a backup of my configuration so that it can be imported in my new ones?
Answer to 1) and 2) :
I propose to run the backup script "sge_save_config.sh" on the qmaster
host. The script is using standard sge_xxxx calls to get the
configuration. As long as you don't get an error message it should also
work on a different administration host. The backup script generates a
directory which contains the configurations in a format which can be
read by the upgrade script when installing the new 62.u5 version. The
stored format is not architecture depended so I assume no problems on
your linux host.
> 3. To confirm, I've been told that I shouldn't have any problems with running our existing 6.1u4 setup concurrently with our new 6.2u5 setup as long as I've defined different TCP port numbers for the qmaster and sge_execd daemons, right? Are there any other things I need to be aware of regarding my upgrade plans?
Answer to 3):
Of course you should also use a different SGE_ROOT directory for the new
> 4. Looking at the documentation, I'm assuming that what I want to do is follow the instructions for installing 6.2u5 using the "Cloned Configuration Method". This will allow me to install 6.2u5 binaries on our new Linux Redhat Qmaster/Shadow servers using different TCP ports and import our current configuration while at the same time still keep our current 6.1u4 servers in production.
> Right? Our SGE binaries are NFS mounted and all SGE servers and clients have access to them.
You should also have an eye on your spooling directory locations. The
cloned cluster should not use any spool directories of the 6.1u4 system.
BTW: What spooling method are you using?
> Sorry for the lengthy e-mail, but I just want to make clear what I am attempting to do. I appreciate any response.
> Kind Regards,
> Wayne Lee
> e-mail: wlee at hess.com
> To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].
Sun Microsystems GmbH Christian Reissmann
Dr.-Leo-Ritter-Str. 7 Software Engineer
D-93049 Regensburg Phone: +49 (0)941 3075 112
Germany Fax: +49 (0)941 3075 222
http://www.sun.de mailto: Christian.Reissmann at sun.com
Sitz der Gesellschaft:
Sun Microsystems GmbH, Sonnenallee 1, D-85551 Kirchheim-Heimstetten
Amtsgericht Muenchen: HRB 161028
Geschaeftsfuehrer: Thomas Schroeder, Wolfgang Engels, Wolf Frenkel
Vorsitzender des Aufsichtsrates: Martin Haering
To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].
More information about the gridengine-users