[ipxe-devel] ipxe and vmware workstation 8 or virtualbox 4.1.2

[Michael Brown]

On Saturday 17 Sep 2011 03:45:35 Edward Villalovoz wrote:
>               I'm using the latest ipxe and I compiled it in Centos 6.  I'm
> basically using ipxe to provide me with a static IP to pxe boot my ESXi 5
> hosts across a layer 3 network since I do not have DHCP available on the
> other VLANs where my ESXi hosts are located on.  My test environment in my
> office is an ESXi server with multiple subnets trunked into it so that I
>  can have my iPXE host on one VLAN (say 10) and my Centos 6 host on a
>  different VLAN (say 20).  I have this setup this way so I can test layer 3
>  routing with ipxe.  And for the most part, it's working fine except for I
>  have 2 issues.
> 
> 1.       ipxe passes the command line boot arguments in the format of
> ip=<IP>:x:x:x:x etc.  Is there a way to have it pass the arguments like
> ip=xxx netmask=xxx gateway=xxx etc?  It's my understanding that this new
> format was changed some time ago, but I need the older format for the
> project I'm working on.

iPXE doesn't automatically generate either of those formats.  Are you sure 
that information isn't being added by another intermediate bootloader, such as 
pxelinux?

Using iPXE, then you can generate the argument format you want using something 
like:

  kernel vmlinuz ip=${net0/ip} netmask=${net0/netmask} gateway=${net0/gateway}

> 2.       If I use ipxe in a VM on ESXi 4.1 and put a static IP in it
> "ifopen; set net0/ip xxx.xxx.xxx.xx; set net0/netmask xxx.xxx.xxx.xxx; set
> net0/gateway xxx.xxx.xxx.xxx" everything works fine and I can ping the IP
> from another subnet on my network.  But on my home network all I have is
> vmware workstation or virtual box so I'm not testing layer 3, but only
> layer2.  So I'm using the same ipxe.iso file that worked fine in my ESXi
>  4.1 environment at home in my vmware/virtualbox environment.  My issue is
>  after I start pinging the ipxe host's IP from my Centos 6 host, it
>  response for about 5-8 replies, then it just stops responding to a ping
>  from my Centos 6 host.  I don't understand why it works fine on type 1
>  hypervisors, but not on type 2.

Does the "ifstat" command (http://ipxe.org/cmd/ifstat) show anything of 
interest?  (In particular, does the "RX" counter continue to increase even 
after iPXE stops responding to pings, and do you see any RX errors?)

Michael