WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-users

Re: [Xen-users] Cluster (VPC), Eucalyptus and preferred distribution to

To: Xen-user List <xen-users@xxxxxxxxxxxxxxxxxxx>
Subject: Re: [Xen-users] Cluster (VPC), Eucalyptus and preferred distribution to it.
From: Luke S Crawford <lsc@xxxxxxxxx>
Date: 28 Mar 2009 18:57:37 -0400
Delivery-date: Sat, 28 Mar 2009 15:50:21 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <49CDCB51.8030804@xxxxxxxxxx>
List-help: <mailto:xen-users-request@lists.xensource.com?subject=help>
List-id: Xen user discussion <xen-users.lists.xensource.com>
List-post: <mailto:xen-users@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=unsubscribe>
References: <6b7f6eb0903271033i49f55b0apa0aea8b7db1fb020@xxxxxxxxxxxxxx> <49CDA43A.2020406@xxxxxxxxxx> <m3myb66xf9.fsf@xxxxxxxxxxxxxxxxxx> <49CDCB51.8030804@xxxxxxxxxx>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.4
Thomas Goirand <thomas@xxxxxxxxxx> writes:

lsc> How would you like to be able to move your VPS from one provider to 
lsc> another with a few clicks?

> Can't you use rsync for that? Is using rsync THAT hard? I don't think,
> by the way, that a hosting provider will give you in his interface tools
> to ease the move out...

well, that's the problem, see, to use rsync and get a clean image, you really
need the ability to boot into read-only/single user mode.

Is that standard?  I don't know.   I provide a 'boot into a read-only disk'
mechanisim that lets you mount your disk to my customers.  there's an option
in the PVGRUB menu.   (the system is configured to boot from the menu.lst on
the read only partition, which falls through after two seconds to the menu.lst
on the partition writable to the user.  The idea being that no matter how badly
the user messes up their partition, they can boot into the rescue image.)

Of course, you can always just rsync from a unquiesced, running system,
and it works most of the time.  my point is that the provider can help a lot.

As for "that a hosting provider will give you in his interface tools to
ease the move out..."   I think they *should*  -  I should.   
Right now, though, while I do give you the rescue image you can use, I don't
help beyond that.

And GPLhost gives away some of it's management tools.   So GPLhost does seem
like the sort of company that would give you a tool to easily move out or
do backups to your own storage.

Besides, everyone who is cheaper than EC2 (and with the price of ram being
what it is, our ranks are growing) has an interest in making it easy
to switch providers.  With a system like Eucalyptus, you could keep backup 
images on ec2, bring them up once a month for an automated test, but run 
most of the time on a lower-priced service that uses the (largely compatable)
Eucalyptus.   Even a very small percentage of the ec2 userbase would be
incredibly awesome for a small provider like me.  

I'm sure that we are both doing the best we can to serve our customers.
My point is that you are best off if you are in a position where any one 
company can screw up (or raise prices on you)  and you are still OK.  

(further, especially for services with negotiated prices, you are in a 
much stronger negotiation position if you can leave.  VPS pricing, though
is usually not negotiated.)  

lsc> If you single source anything, your vendor has a fiduciary duty to their
lsc> shareholders to screw you.

> Well, at GPLHost, we are fully auto-financed, fully self-owned. Why
> would we have a duty to ourself to screw our customers? I think this is
> quite the opposite, "our job is to help", as I always say. Weather the
> customer is leaving or not... The above might apply to SOME provider, I
> don't think it applies to all.

I'm sure that both of us are doing as well for our customers as we can.
In fact, I admire what gplhost has done.  Open-sourcing your tools says
very good things about your company.   

I'm just saying, in general, you don't want to single source anything.
Companies change hands.  People change.   Companies change.  If nothing
else, if you have 2 providers, you know that a mistake at either one
can't take out all your capacity.   I know I have made some pretty big
mistakes in my time[1] and I would be at least a little suspicious of any
person or company who claimed to have never made a mistake.


[1] the most recent of which involved not setting dom0-min-mem on one of my
servers, resulting in a server crash last weekend.  Immediately, I've
lost a full servers worth of revinue for next month, and probably some
of those customers are not going to renew because of it.  

http://wiki.xen.prgmr.com/xenophilia/2009/03/problems-on-hind-oom-killer-er.html

I plan on setting up a puppet system to make sure I don't forget this sort
of thing, but my point is that it happens.  people do stupid things.  
hardware breaks.  

-- 
Luke S. Crawford
http://prgmr.com/xen/  -   Hosting for the technically adept
                           We don't assume you are stupid.  

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