Hi James,

xCAT does not require BMC access unless you plan on using those 
particular features. For just the simple act of managing the deployment 
of diskul/diskless nodes, there's no real necessity for having BMC 
access. It makes life easier to be able to remotely power on/off the 
nodes and view their consoles, but it's not required.

 > I have noticed I am able to use the manual discovery method 
populating the 'mac' table by hand; Autodiscovery didn't seem to work 
(snmp enabled switch, with switch table port-to-host mapping completed. 
The nodes looped endlessly and never put the static address in 
dhcpd.leases , possibly because of the missing IPMI information causing 
a script to hang..I can't be sure.

I've never been able to get autodiscovery to work either. I've taken the 
road of just populating the mac table manually for new nodes instead of 
battling with discovery :)

 > I *did* manage without any trouble at all to get a stateful Centos 
node to install, but on reboot with the chain set automatically to 
'boot' it stalls after the boot loader has completed; if I cheese it and 
disable tftpd, the node boots as normal.

Sounds like perhaps the BIOS isn't following the request to boot from disk?


 > My question is, are these problems simple configuration issues, or is 
it related to the absence of BMC information?

Simple configuration issues.




On 6/21/2013 10:18 AM, James A.s. Trimbee wrote:
> Is it possible to have a functional XCAT installation with nodes that do not 
> have BMC over ethernet available? I have some inherited HP Proliant servers 
> that have a BMC controller but it has no network capability, either in BIOS 
> or using 'ipmitools' (no LAN channels are available and I cannot enable them 
> in BIOS).
>
> For this reason I have not populated any tables pertinent to hardware 
> management, such as nodehm
>
> I have noticed I am able to use the manual discovery method populating the 
> 'mac' table by hand; Autodiscovery didn't seem to work (snmp enabled switch, 
> with switch table port-to-host mapping completed. The nodes looped endlessly 
> and never put the static address in dhcpd.leases , possibly because of the 
> missing IPMI information causing a script to hang..I can't be sure.
>
> I *did* manage without any trouble at all to get a stateful Centos node to 
> install, but on reboot with the chain set automatically to 'boot' it stalls 
> after the boot loader has completed; if I cheese it and disable tftpd, the 
> node boots as normal.
>
> My question is, are these problems simple configuration issues, or is it 
> related to the absence of BMC information?
> ------------------------------------------------------------------------------
> 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
>


------------------------------------------------------------------------------
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

Reply via email to