From: Jeremy Fitzhardinge <jeremy@goop.org>
To: "Radhakrishnan, Prashanth" <Prashanth.Radhakrishnan@netapp.com>
Cc: xen-devel@lists.xensource.com
Subject: Re: Memory-hotplug support for x86_64 domUs?
Date: Tue, 12 Aug 2008 10:46:30 -0700 [thread overview]
Message-ID: <48A1CC76.7090705@goop.org> (raw)
In-Reply-To: <B4A89DAB06701749A8440A7A633FA69AFF4F8F@SACMVEXC2-PRD.hq.netapp.com>
Radhakrishnan, Prashanth wrote:
> Thanks for the info!
>
>
>> But it isn't very useful at present.
>>
>
> Can you please elaborate why. Is it not possible to add memory to VMs
> without rebooting them?
>
Sorry, I was unclear. It's definitely possible in principle to add
memory to VMs on the fly, but it just isn't implemented in the pvops
kernels at the moment. The patch I posted (that you referred to) makes
the balloon driver call into the memory hotplug subsystem to add new
memory to the system. But the new memory is not populated all at once;
it's just a set of unbacked pageframes to start with. The memory
hotplug system automatically puts a file in /sys referring to the newly
added "memory", and allows you to "online" it all. Unfortunately, since
there's no real memory behind the pages, this will almost immediately
cause your domain to crash.
I'd like to make it so that the balloon driver can control onlining
pages on a page-by-page basis, and there should be no usermode interface
for this memory. But so far we haven't worked out mutually acceptible
way to do this.
I could post the straightforward balloon/hotplug patches, but I'm not
comfortable doing so when they leave such an dangerous pit for people to
fall into.
J
prev parent reply other threads:[~2008-08-12 17:46 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <B4A89DAB06701749A8440A7A633FA69AEF672F@SACMVEXC2-PRD.hq.netapp.com>
2008-08-12 16:50 ` Memory-hotplug support for x86_64 domUs? Jeremy Fitzhardinge
[not found] ` <B4A89DAB06701749A8440A7A633FA69AFF4F8F@SACMVEXC2-PRD.hq.netapp.com>
2008-08-12 17:46 ` Jeremy Fitzhardinge [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=48A1CC76.7090705@goop.org \
--to=jeremy@goop.org \
--cc=Prashanth.Radhakrishnan@netapp.com \
--cc=xen-devel@lists.xensource.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.