With xCAT Version 2.8.1 (svn r16213, built Tue May 7 22:55:07 EDT 2013).
Name : xCAT Relocations: /opt/xcat
Version : 2.8.1 Vendor: IBM Corp.
Release : snap201305072259 Build Date: Tue 07 May 2013
10:59:39 PM EDT
Install Date: Tue 11 Jun 2013 11:52:18 PM EDT Build Host:
c97p3l4vp01.ppd.pok.ibm.com
Group : Applications/System Source RPM:
xCAT-2.8.1-snap201305072259.src.rpm
Size : 0 License: EPL
Signature : DSA/SHA1, Tue 07 May 2013 10:59:48 PM EDT, Key ID
20e475a8da736c68
Packager : IBM Corp.
Summary : Meta-package for a common, default xCAT setup
Description :
perl-xCAT.noarch 4:2.8.1-snap201305072255
xCAT.x86_64 2.8.1-snap201305072259
xCAT-client.noarch 4:2.8.1-snap201305072255
xCAT-genesis-base-x86_64.noarch
xCAT-genesis-scripts-x86_64.noarch
xCAT-server.noarch 4:2.8.1-snap201305072255
Please find the procedure as below :
*1. First Scenario*
During Normal Installation of a Stateful Node, with following commands.
nodeset scl-p03-mn osimage=centos6.4-x86_64-install-all
rsetboot scl-p03-mn net
rpower scl-p03-mn boot
Installation of OS completes fine, but default xCAT postscripts -
syslog,remoteshell,syncfiles - do not run. Neither do my custom postscripts
run.
/var/log/httpd/access_log show :
- - [10/Jun/2013:12:00:42 -0400] "GET
/tftpboot/mypostscripts/mypostscript.scl-p03-mn.tmp HTTP/1.0" 404 328 "-"
"Wget/1.12 (linux-gnu)"
- - [10/Jun/2013:12:01:15 -0400] "GET
/tftpboot/mypostscripts/mypostscript.scl-p03-mn.tmp HTTP/1.0" 404 328 "-"
"Wget/1.12 (linux-gnu)"
- - [10/Jun/2013:12:01:46 -0400] "GET
/tftpboot/mypostscripts/mypostscript.scl-p03-mn.tmp HTTP/1.0" 404 328 "-"
"Wget/1.12 (linux-gnu)"
- - [10/Jun/2013:12:02:21 -0400] "GET
/tftpboot/mypostscripts/mypostscript.scl-p03-mn.tmp HTTP/1.0" 404 328 "-"
"Wget/1.12 (linux-gnu)"
- - [10/Jun/2013:12:02:58 -0400] "GET
/tftpboot/mypostscripts/mypostscript.scl-p03-mn.tmp HTTP/1.0" 404 328 "-"
"Wget/1.12 (linux-gnu)"
- - [10/Jun/2013:12:03:29 -0400] "GET
/tftpboot/mypostscripts/mypostscript.scl-p03-mn.tmp HTTP/1.0" 404 328 "-"
"Wget/1.12 (linux-gnu)"
It goes on a loop searching for mypostscript.scl-p03-mn or
mypostscript.scl-p03-mn.tmp, and after 10 retries - it gives up and boots
without installing xCAT default postscripts or my custom postscript -
"sab-basec6-setup"
So when i change my custom postscript name "sab-basec6-setup" to
"mypostscript.scl-p03-mn" and copy it into /tftpboot/mypostscripts/
xCAT detects the script during postinstall, and runs it, as i can find the
log after reboot and the respective changes.
However, xCAT still doesnt run its default xCAT postscripts -
syslog,remoteshell,syncfiles.
*Second Scenario :*
I researched on this topic, and found out if enabling
precreatemypostscripts=1 might solve this problem.
After this step :
If i run : nodeset scl-p03-mn osimage=centos6.4-x86_64-install-all
I see that xCAT default mypostscript gets created inside /tftpboot ->
mypostscript.scl-p03-mn
Now when i run the below commands :
nodeset scl-p03-mn osimage=centos6.4-x86_64-install-all
rsetboot scl-p03-mn net
rpower scl-p03-mn boot
Installation of OS completes fine, but it hangs - when running postscripts.
/var/log/messages shows:
Jun 12 12:07:15 no kernel: imklog 5.8.10, log source = /proc/kmsg started.
Jun 12 12:07:15 no rsyslogd: [origin software="rsyslogd" swVersion="5.8.10"
x-pid="12708" x-info="http://www.rsyslog.com";] start
Jun 12 12:07:15 no xCAT: Install: syslog setup
Jun 12 12:07:15 no xcat: Install: setup /etc/ssh/sshd_config
Jun 12 12:07:15 no xcat: Install: setup root .ssh
Jun 12 12:07:37 no xCAT: Before ssh_dsa_hostkey is checked for error
Jun 12 12:07:37 no xCAT: ssh_dsa_hostkey
After some research - I found out it never comes of the infinite loop
inside "remoteshell" script.
while [ -z "$MYCONT" ]; do
let SLI=$RANDOM%10
let SLI=SLI+10
sleep $SLI
getcredentials.awk ssh_dsa_hostkey | grep -v '<'|sed -e
's/</</' -e 's/>/>/' -e 's/&/&/' -e 's/"/"/' -e "s/'/'/"
> /etc/ssh/ssh_host_dsa_key
MYCONT=`cat /etc/ssh/ssh_host_dsa_key`
done
I believe getcredentials.awk ssh_dsa_hostkey is not working as expected.
After abrupt cancel of installation, and reboot of node - I find that
/tmp/ssh_dsa_hostkey is blank ( which is the one getcredentials.awk is
trying to read)
Also, I observed that .ssh directory is not created inside root, though the
code has passed that stage in remote shell.
if [ -d /xcatpost/_ssh ]
then
logger -p local4.info -t xcat "Install: setup root .ssh"
cd /xcatpost/_ssh
mkdir -p /root/.ssh
cp -f * /root/.ssh
cd - >/dev/null
chmod 700 /root/.ssh
chmod 600 /root/.ssh/*
fi
I see that /var/log/messages mentions : Install: setup root .ssh , but
mkdir -p /root/.ssh
cp -f * /root/.ssh
does not seem to be working.
*Third Scenario : *
I have also tried rinstall, giving custom definitions with
provmethod=install.
rinstall -o centos6.4 -p all -a x86_64 scl-p03-mn
This did not make any difference.
if /tftpboot/mypostscripts/mypostscript.scl-p03-mn exists ( where
mypostscript.scl-p03-mn is default template created by xCAT) - then it
goes on an infinite loop.
-------------------------------------------------------
tcpdump, shows communication between xCAT port 3001 and the node, but i
could make any sense out of the dump.
May i please know, how can we correct this issue, and what is the correct
procedure to run postscripts.
Are we suppose to enable precreatemypostscripts=1 to run default
postscripts? or it doesnt need to be enabled and nodeset is suppose to
create mypostscript inside /tftpboot/mypostscripts ?
Other than default xCAT postscripts - All I do is , add my custom
postscripts to node definition - but they do not run, along with the
default postscripts.
Thanks,
Lohit Valleru
------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:
Build for Windows Store.
http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
xCAT-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/xcat-user
