Menu
▾
▴
Re: [Openqrm-user] Openqrm use cases
|
From: Ittay D. <it...@ql...> - 2007-04-26 14:21:37
|
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
<title></title>
</head>
<body dir="ltr" bgcolor="#ffffff" text="#000000">
<pre wrap="">Hi Sharninder,
To work with locally deployed images, you need to follow these steps:
1. have a node (not controlled by qrm) that has the software & OS you want installed locally. It should be on one disk.
- the node should be able to boot through PXE from qrm (this means, having a dhcp server, either already set or use the dhcpd plugin so the openqrm server acts as a dhcp server)
- install the openqrm-resource package and configure the qrm ip in its conf file (no need to run any command after installation)
2. have an NFS storage server (either already set, or use the local_nfs plugin to make the openqrm server act as NFS server)
- define, through the GUI, a storage server (under management tools in the navigation menu).
3. use qrm-cli to start the grab process
</pre>
<p style="margin-bottom: 0cm; margin-top: 0pt;">bin]# ./qrm-cli -u qrm
-p qrm filesystem local grab<br>
Prepare server for grabbing an image<br>
<br>
Arguments are: <br>
-t,--type <linux|windows> Type of
image <br>
-m,--mac <xx:xx:xx:xx:xx:xx> Mac of node to be
grabbed <br>
-n,--name <name> Filesystem image
name <br>
-s,--storage_server <name> Storage
server <br>
-i,--identifier <id> Identifier to locate the
image (mount point for nfs)<br>
[-d,--device </dev/hda>] Device file of disk to be
grabbed (linux format) <br>
[--shutdown] Shutdown the node after grabbing
(default is reboot)<br>
[-h,--help] Get
help </p>
<p style="margin-bottom: 0cm; margin-top: 0pt;"><br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;">4. reboot the node via
pxe (changing the bios if needed)</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;"><br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;">5. the node boots into
pxe and our code takes over and grabs the image, putting files on the
storage (via NFS mount) and creating the image definition in qrm.<br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;"><br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;">6. when grabbing is
finished, the node reboots locally (our PXE configuration for the node
is set to tell it to reboot from the local disk again). you probably
want to remove the openqrm-resource package then<br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;"><br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;">after that, you can
create a VE to use the image you grabbed and start it. you'll need to
have openqrm managed nodes. they boot via PXE, our code dumps the image
to the local disk (you need to make sure they have a local disk with
enough space) and reboots them locally. after that, they reboot
locally, until the VE is stopped<br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;"><br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;">hope this helps,</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;">ittay<br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;"><br>
</p>
<p style="margin-bottom: 0cm; margin-top: 0pt;">Matt Rechenburg wrote
on 04/26/07 16:17:<br>
</p>
<blockquote cite="mid:463...@ql..." type="cite">
<pre wrap="">Hi again Sharninder,
Sharninder Khera wrote:
</pre>
<blockquote type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap=""> Do I still need a Virtual environment and pair it with a
kernel-image ?
</pre>
</blockquote>
<pre wrap="">yep, that is the concept, to have a single profile and configuration
option per service -> VE
</pre>
<blockquote type="cite">
<pre wrap=""> I am trying to load a server using this but I keep
getting not enough resources message.
</pre>
</blockquote>
<pre wrap="">oh, please check in the resource overview if your nodes are idle but
still in maintaince mode.
Use the action menu for the resource to uncheck the maintaince mode
and then
the "not enough resources" messages should disappear and your VE should
start.
</pre>
</blockquote>
<pre wrap="">Oh. Thanks Matt. I did that and then openqrm was able to reboot the
server to install the image.
</pre>
</blockquote>
<pre wrap=""><!---->
:) good
</pre>
<blockquote type="cite">
<pre wrap="">I couldn't find this maintainance mode
thing in the documentation anywhere so decided to ask the group.
</pre>
</blockquote>
<pre wrap=""><!---->
yep, we are still progressing to get better with the docu
</pre>
<blockquote type="cite">
<pre wrap="">Anyway, so now what has happened is that the server was rebooted and
since I didn't have default pxe boot set in the BIOS, the resource is
now showing in an error state since it timed out.
My question is that the tests that we run require the server to be
rebooted many times, so we cannot set it boot from pxe everytime. In
that case, using openqrm for our work seems difficult, because in any
case, I will have to manually get the server in pxe mode twice
atleast, first for it to be detected by openqrm and second time for it
to be installed.
</pre>
</blockquote>
<pre wrap=""><!---->
ok, this is true for resources being detected by PXE-booting them.
The normal way to caputure a local image is to integrate the local-installed
and local-booting system via the openqrm-resource package, grab the image,
then set it to PXE-boot to make it available for the resource pool.
After that the node will stay with PXE-booting even if deployed local.
</pre>
<blockquote type="cite">
<pre wrap="">
--
Sharninder
</pre>
</blockquote>
<pre wrap=""><!---->greetz,
Matt
</pre>
</blockquote>
<br>
<div class="moz-signature">-- <br>
<b>Ittay Dror</b>
<br>
Chief Architect,
<br>
R&D, Qlusters Inc.
<br>
Web: <a href="http://www.qlusters.com/">qlusters.com</a>
<br>
Email: <a href="mailto:it...@ql...">it...@ql...</a>
<br>
Phone: +972-3-6081994 <br>
<br>
<a href="www.openqrm.org">openQRM</a> - Data Center Provisioning
</div>
</body>
</html>
|