Thanks Yama and William. Yes, William, makes sense this is really a Rocks question. I'll post there if I have further trouble.
And yes that's right, rocks does have installation scripts which I should be able to modify, haven't looked at that in quite some time. Looking there now, I was having trouble back then getting post-installation steps to run so I'll have to ask for help on th Rocks list. -M On Wed, Nov 18, 2015 at 3:43 PM, Yama Nawabi (Elev) < [email protected]> wrote: > Hi Michael, > > > > (This information is RHEL specific, and I personally use xCAT, so some of > this may need to be interpreted for ROCKS) > > > > Does ROCKS provide a kickstart template that you may be able to edit? In > my xCAT configuration's kickstart template, I add the following after the > selinux and reboot statements: > > > > repo --name=puppetlabs-products --baseurl= > http://yum.puppetlabs.com/el/6/products/x86_64/ > > repo --name=puppetlabs-deps --baseurl= > http://yum.puppetlabs.com/el/6/dependencies/x86_64/ > > services --enabled=puppet > > > > In the packages section, specify the puppet package, and all sge related > packages that need to be installed. You may need to drop the sge packages > in your x86_64/Packages/ directory and regenerate the repo using > createrepo. Find out how ROCKS can be used to drop in a custom puppet.conf > into /etc/puppet and prop up a host that will serve as the puppetmaster. > > > > Upon first boot, you will have a sge installation in /opt/sge and a puppet > configuration. > > > > Use puppet to (I recommend using run-stages): > > 1. Mount whatever NFS mounts that are necessary for your shared cell 2. > Ensure the execd rc file is placed in /etc/init.d 3. Chkconfig the service > and ensure that it's started to ensure sge_execd is running 4. Ensure > whatever ports are necessary are open (although I just used firewall > --disabled in the KS template) > > > > Ensure that you have the hosts already configured on your qmaster. Hope > this helps, let me know if you need more information. > > > > -Yama > > > > Yama > > Systems Engineer > > Institute for Neuroimaging and Informatics, Keck School of Medicine of USC > > > > > > *From:* [email protected] [mailto:[email protected]] > *On Behalf Of *Michael Stauffer > *Sent:* Tuesday, November 17, 2015 1:56 PM > *To:* Gridengine Users Group <[email protected]> > *Subject:* [gridengine users] how best to initialize a rocks node for SGE > on boot? > > > > SoGE 8.1.8 > > Hi, > > What's the proper way to automatically initialize SGE on a rocks-managed > compute node after it's reinstalled? I recently upgraded my cluster from > rocks 6.1 to 6.2, and in the process moved from OGS to SoGE 8.1.8. With > that came a bunch of manual configuration steps for SoGE that rocks used to > handle with its bundled distro of (old) OGS. > > I have a number of steps to perform after a node comes up after a > reinstall. How should I do this to have it happen automatically? In an > /etc/init.d script? How do I get it into the distro in the right place? > Below are the steps I want to perform. Thanks for any advice. > > usermod -u 399 sgeadmin > > groupmod -g 399 sgeadmin > > echo "#manually added" >> /etc/fstab > > echo "<front-end>:/opt/sge /opt/sge nfs > defaults,noatime 0 0" >> /etc/fstab > > mount /opt/sge > > . /etc/profile.d/cfn-sge-env.sh > > cp $SGE_ROOT/default/common/sgeexecd /etc/init.d/sgeexecd.<cluster-name> > > /usr/lib/lsb/install_initd /etc/init.d/sgeexecd.<cluster-name> > > service sgeexecd.<cluster-name> start > > > > -Michael > > >
_______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
