[sheepdog-users] Sheepdog 0.9 missing live migration feature

Vasiliy Tolstov v.tolstov at selfip.ru
Tue May 12 10:44:03 CEST 2015


12 мая 2015 г. 4:59 пользователь "Liu Yuan" <namei.unix at gmail.com> написал:
>
> On Mon, May 11, 2015 at 01:58:07PM +0200, Walid Moghrabi wrote:
> > Hi,
> >
> > > Sorry for keeping you waiting. I'll backport the patch tonight.
> >
> > You're great :D
> >
> > > Thanks a lot for your help. But I need to say that journaling and
> > > object cache are unstable features. Please don't use them in
> > > production.
> >
> > Too bad :(
> > I was really happy to try this on my setup, I equiped every node with a
separated SSD drive on which I was wanting to store Sheepdog journal and/or
object cache.
> > Why are thse features "unstable" ?
> > What are the risks ? In which conditions shouldn't I use them ?
>
> Journaling won't be useful if you use object cache. Technically, journal
is less
> stable as object cache. Object cache is near mature so it is very
reasonable to
> make good use of object cache if you equip a dedicated SSD.
>
> Just go ahead and stress object cache, if you meet any problem, please
report
> back and we'll try to fix it. In the past, we heavily ran object cache on
our
> prevoius production and no big problem found. But I have to admit, there
might
> be some corner case bugs around need to fix. I'd say that I'll shoot them
down
> if I aim at them :)
>

If  I have only sad disks,cache does not provide any benefits?

> Thanks,
> Yuan
>
> >
> > Unless there is heavy risk, I think I'll still make a try (at least in
my crash tests before moving the cluster to production) because it looks
promising and anyway, Sheepdog is not considered stable until now and I'm
using it with real joy since 0.6 even on production platform so ... ;)
> >
> > Anyway, just for my wn curiosity, here is what I'm planning to do for
my setup, I'd really appreciate any comment on it :
> >
> > 9 nodes with each :
> >   - 2 interfaces, one for cluster communication ("main" network) and
one dedicated to Sheepdog's replication ("storage" network) with fixed IPs,
completely closed and Jumbo frames enabled (mtu 9000)
> >   - 3 600Gb SAS 15k dedicated hard drives that are not part of any RAID
(standalone drives) that I was thinking using in MD mode
> >   - 1 SSD SATA drive (on which the OS resides and that I was thinking
to use for Sheepdog's journl and object cache)
> >
> > So that means 27 hard drives cluster that I wanted to format using
Erasure Code but until now, I don't really now which settings I'll
configure for this ... I'd like to find the good balance between
performances, security and storage space ... any proposition mostly
welcomed.
> >
> >
> > Regards
> >
> >
> > ----- Mail original -----
> > De: "Hitoshi Mitake" <mitake.hitoshi at gmail.com>
> > À: "Walid Moghrabi" <walid.moghrabi at lezard-visuel.com>
> > Cc: "Liu Yuan" <namei.unix at gmail.com>, "sheepdog-users" <
sheepdog-users at lists.wpkg.org>
> > Envoyé: Lundi 11 Mai 2015 12:52:21
> > Objet: Re: [sheepdog-users] Sheepdog 0.9 missing live migration feature
> >
> > Hi Walid,
> >
> > On Mon, May 11, 2015 at 7:08 PM, Walid Moghrabi
> > <walid.moghrabi at lezard-visuel.com> wrote:
> > > Hi,
> > >
> > > Thanks for your information but I'm now building a whole new
production cluster and I really need Live Migration to work.
> > > You said you'll backport the locking patch so that it would re-enable
live migration but the ProxMox "pve-sheepdog" package maintainer (packaged
Sheepdog for the ProxMox distribution) didn't find the commit in the
0.9-stable branch.
> > > I'd like to base my setup on Sheepdog as I really dislike Ceph for
many reasons and this is the only no go point right now.
> > > Would it be possible to push that commit and build a "0.9.2" ?
> > > That would be great.
> >
> > Sorry for keeping you waiting. I'll backport the patch tonight.
> >
> > >
> > > Many thanks in advance for your help.
> > > I'd be happy to provide tests, benchmarks and whatever useful
information you might want from this setup (I'll try to enable as much
features as I can on this 9 nodes setup such ass SSD journaling, SSD object
caching, multi-devices, Erasure Code, dedicated replication network, ...)
> >
> > Thanks a lot for your help. But I need to say that journaling and
> > object cache are unstable features. Please don't use them in
> > production.
> >
> > Thanks,
> > Hitoshi
> >
> > >
> > > Best regards.
> > >
> > >
> > > ----- Mail original -----
> > > De: "Liu Yuan" <namei.unix at gmail.com>
> > > À: "Walid Moghrabi" <walid.moghrabi at lezard-visuel.com>
> > > Cc: "Hitoshi Mitake" <mitake.hitoshi at gmail.com>, "sheepdog-users" <
sheepdog-users at lists.wpkg.org>
> > > Envoyé: Vendredi 6 Mars 2015 04:33:22
> > > Objet: Re: [sheepdog-users] Sheepdog 0.9 missing live migration
feature
> > >
> > > On Thu, Mar 05, 2015 at 06:21:02PM +0100, Walid Moghrabi wrote:
> > >> > Sorry for your inconvenience, Walid. I'll backport it to
stable-0.9 later.
> > >>
> > >>
> > >> Thank you for your kindness, don't worry, I'm not blaming anyone for
this, I was just sad that this great feature was not available anymore but
I'm happy to see there are plans to get it back again.
> > >>
> > >>
> > >> Just a question : I rebuilt my cluster completely when upgrading to
0.9, once this patch is backported to the 0.9 branch, will I have to
re-format again or will it be compatible as is ?
> > >
> > > I'm afraid you have re-format the cluster.
> > >
> > > Yuan
> --
> sheepdog-users mailing lists
> sheepdog-users at lists.wpkg.org
> https://lists.wpkg.org/mailman/listinfo/sheepdog-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.wpkg.org/pipermail/sheepdog-users/attachments/20150512/cb0b1241/attachment-0002.html>


More information about the sheepdog-users mailing list