[ipxe-devel] Native Ceph Connectivity

Brian Rak brak at gameservers.com
Tue May 13 16:02:33 UTC 2014


Ceph is quite complicated, and I suspect you're going to run into 
serious issues trying to get RBD support in iPXE.  You'd really need to 
implement a tiny RBD client, which sounds like it would be pretty 
complicated.

I'd suggest an alternative:  Store the linux kernel and initramfs in 
Ceph Object Storage.  iPXE can already boot from HTTP, so you can load 
your kernel/initrd via object storage, then let the kernel handle 
booting from the RBD.

I can't really help with how to get the XenServer initrd to support 
Ceph, but that seems like a far simpler process then updating iPXE to 
support Ceph.

On 5/13/2014 11:23 AM, Stephen Perkins wrote:
>
> Hi all,
>
> I have a goal in mind and I'm not entirely sure how to reach it.    
> So... with this in mind, I thought I would discuss what I want as an 
> end result and then ask if a certain approach may make sense.
>
> End goal:  Create a highly available (no single point of failure) 
> scale out infrastructure for booting and running diskless XenServer 
> hosts and lots of Guest Operating systems that have live migration 
> capabilities.  I want this with as few systems in place as is possible.
>
> While most people will start down the iSCSI path, I am not entirely 
> sure that this is the approach I would like to look at. The iPXE full 
> iSCSI stack is awesome and provides great capabilities but it requires 
> a lot of clustering work to make it highly available.  This is 
> compounded if you want a cluster with more than 2 nodes.
>
> I am interested in using the ceph clustered storage system.  This 
> system already provides me with a highly available scale out solution 
> and... once configured and working give me the highly available 
> storage I want and integrates well with the Xen Clients.
>
> But... the hard part is that I want to boot diskless XenServers from 
> this ceph store.   This is where:
>
>                 1) iPXE comes in
>
>                 2) My knowledge about ends
>
> My thought is that I would boot a customized iPXE from a tiny USB DOM 
> or a highly available DHCP/TFTP/PXEboot infrastructure.  Once iPXE is 
> running, I would like to boot directly from a ceph cluster volume 
> instead of the more standard iSCSI volume.
>
> So, I wanted to ask thoughts on whether it makes sense to try to 
> develop another backend connectivity option to iPXE.  I would like to 
> look at adding a ceph/RBD option that will allow me to provide a list 
> of IP addresses (and other needed config info) for the ceph cluster 
> and allow me to mount a ceph store to boot from.
>
> Then... I would have to address the problem of how to get an initrd 
> for XenServer that would have the ceph modules available and allow me 
> to boot a root file system from there.
>
> Is iPXE the correct place to look to help provide this?
>
> Is this a hugely monumental project... or just a monumental project 
> (given that the ceph client code exists and is open source)...
>
> I thought I would reach out here before I wrote to the ceph group.  I 
> my approach is embarrassingly wrong, please feel free to let me know!
>
> Thanks,
>
> - Steve
>
>
>
> _______________________________________________
> ipxe-devel mailing list
> ipxe-devel at lists.ipxe.org
> https://lists.ipxe.org/mailman/listinfo.cgi/ipxe-devel

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ipxe.org/pipermail/ipxe-devel/attachments/20140513/10e353f0/attachment.htm>


More information about the ipxe-devel mailing list