OpenSuSE issues with a few things for cluster

First off, it appears that the zypper problem is solved. Zypper shares some similer command line bits with yum. This helped with a faster learning curve. Zypper also supports a feature I wish were in yum, but I use grep for. zypper search gcc and yum list | grep gcc are the same in function. Zypper is also much faster than yum. Interpreted languages don’t work so well with large data sets, such as many installation packages, and dense dependency trees they have to construct and traverse. Also, our configuration file almost completely works with minor modifications. I’ll have to fix the package naming differences, but thats about it. What doesn’t work? Funny you should ask.

The compilers. Intel compilers that the customer has installed are incompatible with OpenSuSE 11.1. Looking on intel’s forum support, they provide instructions for a work around … that doesn’t work. The queuing system. SGE, or at least as it is installed right now, doesn’t work on OpenSuSE 11.1 . I have to look into this. So the options for a fully operational OFED 1.4.2 stack are this OpenSuSE 11.1 but without the needed compilers or queueing system, OpenSuSE 10.3 which is nearing EOL, some other Linux. I am thinking about this. Might just give in and go Centos 5.3, load OFED 1.4.2, update the kernel (most distro kernels are not appropriate for HPC). This could conceivably solve the problem. And the Intel compilers will work, as will the queuing system. The end user shouldn’t care which OS is on their cluster as long as it works. The only real upgrade challenge will be the head node. The other nodes can be redone remotely without a problem. Ugh. Not pretty no matter how it is spun.