Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 11 Nov 2016 16:57:35 +0100
From:      Julien Cigar <julien@perdition.city>
To:        Palle Girgensohn <girgen@FreeBSD.org>
Cc:        freebsd-fs@freebsd.org, Julian Akehurst <julian@pingpong.net>
Subject:   Re: Best practice for high availability ZFS pool
Message-ID:  <20161111155735.GM81247@mordor.lan>
In-Reply-To: <A7EF341C-C698-47E2-9EDE-04840A86CD4F@FreeBSD.org>
References:  <5E69742D-D2E0-437F-B4A9-A71508C370F9@FreeBSD.org> <5DA13472-F575-4D3D-80B7-1BE371237CE5@getsomewhere.net> <8E674522-17F0-46AC-B494-F0053D87D2B0@pingpong.net> <5127A334-0805-46B8-9CD9-FD8585CB84F3@chittenden.org> <A7EF341C-C698-47E2-9EDE-04840A86CD4F@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help

--Z/kiM2A+9acXa48/
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Nov 11, 2016 at 04:16:52PM +0100, Palle Girgensohn wrote:
> Hi,
>=20
> Pinging this old thread.
>=20
> We have revisited this question:
>=20
> A simple stable solution for a redundant storage with little or no down t=
ime when a machine breaks. Storage is served using NFS only.
>=20
>=20
> It seems true HA is always complicated. I'd rather go for a simple unders=
tandable solution and accept sub minute downtime rather than a complicated =
solution. For our needs, the pretty solution lined up in the FreeBSD Magazi=
ne seems a bit overly complicated.
>=20
> So here's what we are pondering:
>=20
> - one SAS dual port disk box
>=20
> - connect a master host machine to one port and a slave host machine to t=
he the other port
>=20
> - one host is MASTER, it serves all requests
>=20
> - one host is SLAVE, doing nothing but waiting for the MASTER to fail
>=20
> - fail over would be handled with zpool export / zpool import, or just zp=
ool import -F if the master dies.
>=20
> - MASTER/SLAVE election and avoiding split brain using for example CARP.
>=20
> This is not a real HA solution since zpool import takes about a minute. I=
s this true for a large array?
>=20
> Would this suggestion work?

I'm using someting like this here, a zpool over 2 local disks and 2
iscsi disks and the following failover script:
https://gist.github.com/silenius/cb10171498071bdbf6040e30a0cab5c2

It works like a charm except that I'm having this issue:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D211990=20

(apparently this problem does not appear on 11.0-RELEASE)

>=20
> Are there better ideas out there?
>=20
> Cheers,
> Palle
>=20
>=20
>=20
>=20
>=20
>=20
> > 18 maj 2016 kl. 09:58 skrev Sean Chittenden <sean@chittenden.org>:
> >=20
> > https://www.freebsdfoundation.org/wp-content/uploads/2015/12/vol2_no4_g=
roupon.pdf
> >=20
> > mps(4) was good to us.  What=E2=80=99s your workload?  -sc
> >=20
> > --
> > Sean Chittenden
> > sean@chittenden.org
> >=20
> >=20
> >> On May 18, 2016, at 03:53 , Palle Girgensohn <girgen@pingpong.net> wro=
te:
> >>=20
> >>=20
> >>=20
> >>> 17 maj 2016 kl. 18:13 skrev Joe Love <joe@getsomewhere.net>:
> >>>=20
> >>>=20
> >>>> On May 16, 2016, at 5:08 AM, Palle Girgensohn <girgen@FreeBSD.org> w=
rote:
> >>>>=20
> >>>> Hi,
> >>>>=20
> >>>> We need to set up a ZFS pool with redundance. The main goal is high =
availability - uptime.
> >>>>=20
> >>>> I can see a few of paths to follow.
> >>>>=20
> >>>> 1. HAST + ZFS
> >>>>=20
> >>>> 2. Some sort of shared storage, two machines sharing a JBOD box.
> >>>>=20
> >>>> 3. ZFS replication (zfs snapshot + zfs send | ssh | zfs receive)
> >>>>=20
> >>>> 4. using something else than ZFS, even a different OS if required.
> >>>>=20
> >>>> My main concern with HAST+ZFS is performance. Google offer some insi=
ghts here, I find mainly unsolved problems. Please share any success storie=
s or other experiences.
> >>>>=20
> >>>> Shared storage still has a single point of failure, the JBOD box. Ap=
art from that, is there even any support for the kind of storage PCI cards =
that support dual head for a storage box? I cannot find any.
> >>>>=20
> >>>> We are running with ZFS replication today, but it is just too slow f=
or the amount of data.
> >>>>=20
> >>>> We prefer to keep ZFS as we already have a rather big (~30 TB) pool =
and also tools, scripts, backup all is using ZFS, but if there is no soluti=
on using ZFS, we're open to alternatives. Nexenta springs to mind, but I be=
lieve it is using shared storage for redundance, so it does have single poi=
nts of failure?
> >>>>=20
> >>>> Any other suggestions? Please share your experience. :)
> >>>>=20
> >>>> Palle
> >>>=20
> >>> I don=E2=80=99t know if this falls into the realm of what you want, b=
ut BSDMag just released an issue with an article entitled =E2=80=9CAdding Z=
FS to the FreeBSD dual-controller storage concept.=E2=80=9D
> >>> https://bsdmag.org/download/reusing_openbsd/
> >>>=20
> >>> My understanding in this setup is that the only single point of failu=
re for this model is the backplanes that the drives would connect to.  Depe=
nding on your controller cards, this could be alleviated by simply using mu=
ltiple drive shelves, and only using one drive/shelf as part of a vdev (the=
n stripe or whatnot over your vdevs).
> >>>=20
> >>> It might not be what you=E2=80=99re after, as it=E2=80=99s basically =
two systems with their own controllers, with a shared set of drives.  Some =
expansion from the virtual world to real physical systems will probably nee=
d additional variations.
> >>> I think the TrueNAS system (with HA) is setup similar to this, only w=
ithout the split between the drives being primarily handled by separate con=
trollers, but someone with more in-depth knowledge would need to confirm/de=
ny this.
> >>>=20
> >>> -Jo
> >>=20
> >> Hi,
> >>=20
> >> Do you know any specific controllers that work with dual head?
> >>=20
> >> Thanks.,
> >> Palle
> >>=20
> >>=20
> >> _______________________________________________
> >> freebsd-fs@freebsd.org mailing list
> >> https://lists.freebsd.org/mailman/listinfo/freebsd-fs
> >> To unsubscribe, send any mail to "freebsd-fs-unsubscribe@freebsd.org"
> >=20
>=20
> _______________________________________________
> freebsd-fs@freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-fs
> To unsubscribe, send any mail to "freebsd-fs-unsubscribe@freebsd.org"

--=20
Julien Cigar
Belgian Biodiversity Platform (http://www.biodiversity.be)
PGP fingerprint: EEF9 F697 4B68 D275 7B11  6A25 B2BB 3710 A204 23C0
No trees were killed in the creation of this message.
However, many electrons were terribly inconvenienced.

--Z/kiM2A+9acXa48/
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
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=QYE3
-----END PGP SIGNATURE-----

--Z/kiM2A+9acXa48/--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20161111155735.GM81247>