[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] xen cloud - shared storage question


  • To: Ahmed Kamal <email.ahmedkamal@xxxxxxxxxxxxxx>
  • From: Peter den Hartog <peterdenhartog@xxxxxxxxx>
  • Date: Mon, 8 Feb 2010 13:21:52 +0100
  • Cc: xen-users@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Mon, 08 Feb 2010 04:23:08 -0800
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc:content-type; b=Dcgg3d40pE8pBASw5g1ROtdxZTV01G6kFlivXyluOMRNEvkFaKp3mHeLKizOaPrZUa dWYLnP39R15bcXFpdoqAaOye9d//BP8zJpLr3tBZQbQMvnXJYYtGOpx+6BGQwRTeVPwq t9yOUXD4ZKJ/UMgarY3K8bLGzAFMM4q31zwng=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>

You can actualy use keepalived for that, it's an IP failover program, that's capaible of realising it become master and adding an VIF to the server, then you can start some services (it can execute stuff on becoming master/slave) So then you can do the mount disking, the nfs sharing with that, the IP failover is allready in keepalived, so then your done i think, i have a bit of experiance with failover and xen so i should be able to work something out :-)

On Mon, Feb 8, 2010 at 12:25 PM, Ahmed Kamal <email.ahmedkamal@xxxxxxxxxxxxxx> wrote:
to me sounds like a plan. Except that assuming the storage box fails, nothing is gonna automate the failover to the second box .. you need to handle that (Mount disk, NFS share, take over IP .. etc) yourself I believe. But I'm no expert

Another potential approach, is to try distributed fault tolerant filesystems a la glusterfs. If you go that route, I'd be interested in your conclusions though.

Regards


On Mon, Feb 8, 2010 at 1:12 PM, Peter den Hartog <peterdenhartog@xxxxxxxxx> wrote:
hehe it's not that i don't like it, the hole setup makes sense but i'm just thinking about how to implement this in my situation :-), so basicly let's say i have 4 servers (physicial) i could take 2 for running the vm's, and 2 for actualy holding the vm's and replicate those last 2 servers with drbd or something like that..


On Mon, Feb 8, 2010 at 12:09 PM, Ahmed Kamal <email.ahmedkamal@xxxxxxxxxxxxxx> wrote:
Yes the VMs will be stored on the NFS sr, and yes that's a SPOF. if you don't like that, replicate the storage box

On Mon, Feb 8, 2010 at 1:06 PM, Peter den Hartog <peterdenhartog@xxxxxxxxx> wrote:
Ok that makes sense, so i would have to have an extra server configured with NFS that he uses to migrate, or are the actual VM's have to be stored on that NFS sr? Because doesn't that mean that you will have a single point of failure, the NFS server? 


On Mon, Feb 8, 2010 at 12:02 PM, Ahmed Kamal <email.ahmedkamal@xxxxxxxxxxxxxx> wrote:
The standard setups needs a third shared storage box exposing either iscsi or NFS, if you don't want to do any hacks

On Mon, Feb 8, 2010 at 12:52 PM, Peter den Hartog <peterdenhartog@xxxxxxxxx> wrote:
I'm pretty succesfull in deploying the xen cloud, but i only got 1 big issue at this moment, when i installed the host software, i just installed all the basic stuff, and now i would like to share the storage of node 1 with node 2, and vica versa so i can migrate (live?) vm's from 1 to 2 and back, i created a storrage pool and this seemed to work.. but while trying to migrate i got an error that there was no shared sr..

is there a way to share the default hd's of the 2 hosts?

Thanks,
Peter

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users




--
Groet // Kind regards,
Peter den Hartog

Sent from Amsterdam, NH, Netherlands




--
Groet // Kind regards,
Peter den Hartog

Sent from Amsterdam, NH, Netherlands




--
Groet // Kind regards,
Peter den Hartog

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.