Note: This is a beta release of Red Hat Bugzilla 5.0. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Also email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback here.
Bug 160376 - GUI unable to handle creation of large numbers of similarly configured diskless nodes that are not all part of the same subnet.
Summary: GUI unable to handle creation of large numbers of similarly configured diskle...
Alias: None
Product: Red Hat Enterprise Linux 4
Classification: Red Hat
Component: system-config-netboot
Version: 4.0
Hardware: All
OS: Linux
Target Milestone: ---
: ---
Assignee: Jaroslav Reznik
QA Contact:
Depends On:
TreeView+ depends on / blocked
Reported: 2005-06-14 19:54 UTC by brian
Modified: 2010-04-06 11:02 UTC (History)
0 users

Fixed In Version:
Doc Type: Enhancement
Doc Text:
Clone Of:
Last Closed: 2010-04-06 11:02:46 UTC
Target Upstream Version:

Attachments (Terms of Use)

Description brian 2005-06-14 19:54:47 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.7.8) Gecko/20050512 Red Hat/1.0.4-1.4.1 Firefox/1.0.4

Description of problem:
This is more of a feature extension request then a full on bug report.  To start off I'm working to use this tool to take over for the in house created nfsroot environment we have been using.  The HPC clusters we are running total over 1000 cpus.  My concern is for sections of my clusters we can use the current ability to create a definition for an entire network subnet and thats very good.  However I have another subnet, the network thats the service network with groups of systems doing different things.  What I would love is some way to manage each group in as similar of a way as I can by defining a network diskless definition instead of one to a given host.  I understand the limitation of pxelinux and the file it uses in pxelinux.cfg.  What I'm looking for here is a way to go into the diskless client definition window and give a range of IPs to use such as and have the gui tool create all of those for me tied to a diskless image and leaving the "snapshot" field blank so that the disklessrc script uses the hostname for the snapshot directory.  Something like that for the creation would be very desireable.  Now after they are created it would be nice to be able to go to the main screen and be able to highlight a number of definitions, ie the 20 I just created, and do a mass edit on them so that I could repoint them to a new image if needed, or mass delete them if that was the right thing.  Another idea would be to from that main screen be allowed to create a tree structure so that you can say make a group "pvfs server nodes" and then the 20 systems above I created would go under that section to give order to a very large list of diskless clients.

I would also add this different point and that is your tool is working well to allow me to specify where my images live and thats very nice.  However I can't adjust the location of /tftpboot/linux-install/*.  I would like to be able to specify somewhere/somehow a root location of this directory and be then able to set the tftp config file to function from that point.  Then under that it would be nice to instead of seeing a mass mess of directories to be able to have that grouping I mentioned in the section above to be reflected here.  So that instead of seeing whats currently in /tftpboot/linux-install/ that I might see my group of "pvfs server nodes" seperated into a directory root of its own inside this root of the tftp server directory.  Maybe /tftpboot/linux-install/pvfs/.  THen inside there a full environment of things linix pxelinux.cfg pxelinux.0 etc....  I know this is a waste of space having so many coppies of pxelinux things but to give sanity to managing lots of things creating groups of things seems valuable.

Sorry for core dumping on you in such an unorganized way but I'm very happy to see this tool in the works and would love to voice some ideas to see if it could even better help my environment.  Thanks.

Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:
Please see the description for a walk through of what I'm suggesting

Additional info:

Comment 1 Jason Vas Dias 2005-06-14 20:41:08 UTC
Thanks for the good ideas and feedback - 
they will be considered for future releases.

We are also developing new cluster configuration GUI tools, and some
of your requirements might be more appropriate as part of a larger
cluster management system. GUI tools are also under development for
DHCP configuration with which s-c-nb would also have to be integrated.
Also bug #159903, covering lack of support for clients of
architectures that don't support PXE (non-i386 or x86_64), is being
addressed, and integration with the Stateless Managed Linux (SML) and
Xen virtualization projects is being investigated. So we are
evaluating all these suggestions for future directions along with
yours, and will be looking at a more feature-full version of s-c-nb
for RHEL-5. The first priority is making it do what is currently 
required of it correctly .



Comment 3 Jaroslav Reznik 2010-04-06 11:02:46 UTC
System-config-netboot is now in maintenance mode only, no further development is planned for RHEL 4.x. Closing as WONTFIX.

Note You need to log in before you can comment on or make changes to this bug.