I upgraded one of my clusters from 2.5.3 to 2.7.0. I see a couple of
minor issues:
My old diskless compute images didn't seem to register startup
completion with the xCAT server (nodestat would show them still
netbooting). I've rebuilt the images and they seem to mostly work.
There seems to be an extraneous "set +x" in initrd-stateless. I
haven't looked at it, but for example the big X (with a cat? on top)
is displayed twice. Once from the echo of the command and once from
the command itself.
initrd-stateless.gz went from about 4M to 23M. This isn't necessarily
a problem, just something I noticed.
There is an extraneous error during xcatdsklspost:
/opt/xcat/xcatdsklspost: line 27: lsvpd: command not found
I can't find anything on the wiki or the mailing list on this. It
looks like 'lsvpd' was added to several to the pkglist files (mostly
ppc related + centos). There is a EPEL 5 lsvpd package, but nothing
in CentOS/EPEL 6.
There is another extraneous error I'm seeing in xcatdsklspost:
grep: /opt/xcat/xcatinfo: No such file or directory
Running postscript: remoteshell
Starting sshd: [ OK ]
Running postscript: syncfiles
I'm guessing it is the grep in download_postscripts() but I didn't
look closely.
I'm running CentOS 5.7 on the xCAT server and the diskless compute
nodes.
I upgraded to 2.7.0 to hopefully be better able to provision diskfull
CentOS 6 nodes. I haven't done anything with that yet.
Stuart Barkley
--
I've never been lost; I was once bewildered for three days, but never lost!
-- Daniel Boone
------------------------------------------------------------------------------
Better than sec? Nothing is better than sec when it comes to
monitoring Big Data applications. Try Boundary one-second
resolution app monitoring today. Free.
http://p.sf.net/sfu/Boundary-dev2dev
_______________________________________________
xCAT-user mailing list
xCAT-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xcat-user
