Welcome to the cosicluster server enviorment! Enjoy! and here are a few pointers about what is going on here: CONFIGURATION OF THIS VM: -2 gigs o' space on this ext3 volume -The mac-address is netreged under me (katrakcr) -ssh is restriciting root access, but your reading this, so you already knew that -WE NEED IPTABLES! we should be restricting access to apache (campus only) and locking down access to sql (I think this might require recompiling the current xen kernel with iptables support...) THIS JUST IN: the XenU xen kernel must have iptabels and nfs support compiled in. damnit. -Done. data3.sclabs.clarkson.edu has a new kernel @ /boot/vmlinuz-2.6.11.12-xenU-nfs-iptables -Except, I am having issues with nfs ports and iptables, so iptables is temporarily disabled PACKAGES & INSTALLATION: apache2, php5, mysql-server,mysql-php5 also: build-essential, ant most importantly: lam4-dev, lam4c2 isntalled jdk off suns website jsvc should be compiled from source (you can find more on that in mpiqd's readmes) It seems at this time there is no lammpi runtime package on apt. alkdfjalskfjsalfkjaslkj! compile from source: http://www.lam-mpi.org use : ./configure --without-fc --with-rsh="ssh -x" thus no fortran support nfs! (/jobshare is exported off nfs, restricted to 128.153.144.* readonly) and finally... mpiqd, currently this is on my (katrakcr) svn repository. :/ THE LAB: each lab machine must have a cluster user each cluster user should have identical home directory stuctures aswell as jobshare mounted in there home directories ssh publickey (which the cluster.sclab.clarkson.edu will use when connected) must be present in .ssh/authorized_keys THE MAGIC: apache hosts php upload site php upload site sends jobs to mysql (and verfies users through mysql) php also compiles jobs and throws them in /jobshare (shared via nfs) mpiqd (when turned on) will pop jobs off and run them MPIQD: This is a beast. It was written quickly. Refer to documentation provided with it. May god have mercey on your soul. STUFF: This was all working on the assumtion that there was a reboot script for ITL lab machines (once they are on the cluster build) ALL YOUR FLOPS ARE BELONG TO US!