So, I tracked down at least one of the issues: systemimager did not correctly setup rsync during the OSCAR installation. You might consider adding this check to the system-sanity script (or somewhere else more appropriate).
# checked to see if rsync was running $ rsync localhost:: rsync: failed to connect to localhost: # /etc/xinetd.d/rsync.presis~ existed, but not /etc/xinetd.d/rsync #These steps got rsyncd up and running: $ cp /etc/xinetd.d/rsync.presis~ /etc/xinetd.d/rsync changed disable = yes to disable = no $ cp /etc/systemimager/rsyncd.conf /etc $ /etc/init.d/xinetd restart However, now I get back to the old standard error that I always get after imaging with OSCAR, followed by a reboot of the compute node: "Kernal panic - not syncing". I copied the /etc/modprob.conf file (from a stand alone install on a compute node) to the headnode, but this did not correct the issue. cp modprob.conf.compute-node /var/lib/systemimager/images/oscarimage/etc/modprobe.conf What next? I guess that I will start from scratch again. On Thu, Jun 2, 2011 at 5:39 PM, <ghot...@gmail.com> wrote: > All, > > I just put a fresh install of CENTOS 5.4 x86_64 on a head node and went > through the installation and setup of OSCAR 6.1.1svn06022011 on the head > node. > > Several issues: > 1) During the "Select OSCAR Packages To Install", I had to select > "Experimental" to get most of the desired packages (TORQUE, MPI, etc.). > "Default" only had the Core: base, c3, oda, sc3, sis, switcher, sync-files, > and yum (ganglia, opium, and selinux were "Included", but not checked). What > is the issue here? > > 2) It was very, very slow to build the Client Image. What could be going > wrong? (I have used previous versions of OSCAR and this build took probably > 5 to 6 times longer.) We have a fast gigabit connection. Typical yum updates > download with at least 1 MB/sec. > > 3) The worst of it, the cluster deployment fails during the imaging of the > compute nodes (with the "Have fun! http://systemimager.org"; message). I > tried UYOK and that imaging also failed. > > I have two sets of compute nodes (one older set and one newer set). The > hardware of the newer set of compute nodes is very similar to the hardware > on the head node, except for the hard drives (RAID on head node, one drive > SAS on compute nodes). I might have expected some issues with the slightly > older hardware, but not with the newer hardware (that is very similar to > head node). > > I have previously set up these two sets of hardware with older versions of > OSCAR (the older hardware with OSCAR 4.2 and the newer with OSCAR 5.1rc), > but I wanted to unify them under one head node with the new version of > OSCAR. I don't want to resort to going back to try OSCAR 5.1rc, but I might > have to do so. > > Any advice on getting this to work? It has probably some simple fix. If you > want the log file, a slightly trimmed version is at > http://pastebin.com/KvfqY0ww > > 4) What is the bare-bones list of RPMs to get OSCAR 6.1.1 and CENTOS 5.4 to > work on compute nodes. > > > Thank you in advance for the help. > >
------------------------------------------------------------------------------ Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2
_______________________________________________ Oscar-users mailing list Oscar-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oscar-users
