http://wiki.define-technology.com/mediawiki-1.35.0/index.php?action=history&feed=atom&title=Rocks%3A_Rocks_Solid
Rocks: Rocks Solid - Revision history
2026-05-05T00:07:35Z
Revision history for this page on the wiki
MediaWiki 1.35.0
http://wiki.define-technology.com/mediawiki-1.35.0/index.php?title=Rocks:_Rocks_Solid&diff=2472&oldid=prev
Michael: Created page with "==== Rocks Solid Installation and Configuration ==== RPMs approved and tested: <pre> foundation-pysqlite-2.4.1-1.el5.x86_64.rpm foundation-SQLAlchemy-0.4.8.el5-1...."
2013-05-13T09:46:24Z
<p>Created page with "==== Rocks Solid Installation and Configuration ==== RPMs approved and tested: <pre> foundation-pysqlite-2.4.1-1.el5.x86_64.rpm foundation-SQLAlchemy-0.4.8.el5-1...."</p>
<p><b>New page</b></p><div>==== Rocks Solid Installation and Configuration ====<br />
<br />
RPMs approved and tested:<br />
<br />
<pre><br />
foundation-pysqlite-2.4.1-1.el5.x86_64.rpm <br />
foundation-SQLAlchemy-0.4.8.el5-1.noarch.rpm<br />
foundation-python-elementtree-1.2.6-5.el5.x86_64.rpm<br />
foundation-SQLAlchemy-0.4.8.el5-1.src.rpm<br />
rocks-solid-0.5.0b1-1.el5.noarch.rpm<br />
</pre><br />
<br />
<br />
==== Install at the RPMS ====<br />
<syntaxhighlight><br />
rpm -ivh foundation* rocks-solid*<br />
<br />
* note, we ignored the warning about support group <br />
</syntaxhighlight><br />
<br />
Same for all the compute nodes, put the RPMs in <br />
<syntaxhighlight><br />
/export/rocks/install/contrib/5.3/arch/RPMS<br />
</syntaxhighlight><br />
<br />
Create a new XML configuration file that will extend the current compute.xml configuration file:<br />
<syntaxhighlight><br />
# cd /export/rocks/install/site-profiles/5.1/nodes<br />
# cp skeleton.xml extend-compute.xml<br />
<br />
Inside extend-compute.xml, add the package name by changing the section from:<br />
<br />
<!-- Rocks-Solid --><br />
<package>foundation-pysqlite</package><br />
<package>foundation-python-elementtree</package><br />
<package>foundation-SQLAlchemy</package><br />
<package>rocks-solid</package><br />
</syntaxhighlight><br />
<br />
Save extend-compute and run<br />
<syntaxhighlight> <br />
# cd /export/rocks/install<br />
# rocks create distro<br />
</syntaxhighlight><br />
<br />
==== Naming of the IPMI modules ====<br />
<syntaxhighlight><br />
compute-ipmi-0-0<br />
compute-ipmi-0-1<br />
...<br />
</syntaxhighlight><br />
<br />
==== Edit /etc/rocks-solid.conf ====<br />
<br />
change the following:<br />
<syntaxhighlight><br />
[ipmi]<br />
host_pattern=s/compute/compute-ipmi/<br />
user=ADMIN<br />
passwd=ADMIN<br />
<br />
[temp]<br />
threshold=55<br />
ipmi_attr=.*Sys.* (System Temperature)<br />
<br />
[main]<br />
scheduler=torque<br />
poweron_driver=ipmi (although wol seemed to work!)<br />
</syntaxhighlight><br />
<br />
<br />
Add /etc/rocks-solid.conf to 411 (/var/411/Files.mk) and then run 'rocks sync users'<br />
<br />
The following binaries are installed<br />
<syntaxhighlight><br />
/opt/rocks/bin/check-ignore-host<br />
/opt/rocks/bin/cluster-clean-ps<br />
/opt/rocks/bin/cluster-freehost<br />
/opt/rocks/bin/cluster-ipmi<br />
/opt/rocks/bin/cluster-power<br />
/opt/rocks/bin/cluster-poweron-sched-nodes<br />
/opt/rocks/bin/cluster-powersave<br />
/opt/rocks/bin/cluster-reset-freeze-node<br />
/opt/rocks/bin/cluster-status-acct<br />
/opt/rocks/bin/node-cleanipcs<br />
/opt/rocks/bin/node-envcheck<br />
/opt/rocks/bin/node-term-sge-zombie<br />
/opt/rocks/bin/node-term-user-ps<br />
/opt/rocks/bin/queue-limit-user-cpu<br />
</syntaxhighlight><br />
<br />
Turn nodes off and on depending on the load and queue<br />
<syntaxhighlight><br />
cluster-powersave<br />
cluster-powersave -d -v (verbose and dry run)<br />
</syntaxhighlight><br />
<br />
Check the power status of all the nodes<br />
<syntaxhighlight><br />
cluster-power status<br />
</syntaxhighlight><br />
<br />
Run an ipmi query on all the nodes<br />
<syntaxhighlight><br />
cluster-ipmi ipmicommand<br />
</syntaxhighlight></div>
Michael