10 Sep
2019
10 Sep
'19
4:14 a.m.
I have 2 OpenStack environment that I want to integrate to an existing ceph
cluster. I know technically it can be done but has anyone tried this?
- Vlad
ᐧ
10 Sep
10 Sep
4:20 a.m.
on 2019/9/10 17:14, vladimir franciz blando wrote:
I have 2 OpenStack environment that I want to
integrate to an existing
ceph cluster. I know technically it can be done but has anyone tried this?
Sure you can. Ceph could be deployed as separate storage service,
openstack is just its customer. You can have one customer, but also can
have multi-customers for Ceph service.
regards.
4:34 a.m.
We have a single ceph cluster used by 2 openstack installations.
We use different ceph pools for the 2 openstack clusters.
For nova, cinder and glance this is straightforward.
It was a bit more complicated fo radosgw. In this case the setup I used was:
- creating 2 realms (one for each cloud)
- creating one zonegroup for each realm
- creating one zone for each zonegroup
- having 1 ore more rgw instances for each zone
I don't know if there are simpler approaches
Cheers, Massimo
On Tue, Sep 10, 2019 at 11:20 AM Wesley Peng <wesley(a)thepeng.eu> wrote:
on 2019/9/10 17:14, vladimir franciz blando wrote:
Sure you can. Ceph could be deployed as separate storage service,
openstack is just its customer. You can have one customer, but also can
have multi-customers for Ceph service.
regards.
_______________________________________________
ceph-users mailing list -- ceph-users(a)ceph.io
To unsubscribe send an email to ceph-users-leave(a)ceph.io
I have 2 OpenStack environment that I want to
integrate to an existing
ceph cluster. I know technically it can be done but has anyone tried
this?
10:18 p.m.
Thanks for the added info, appreciate it.
- Vlad
On Tue, Sep 10, 2019 at 5:37 PM Massimo Sgaravatto <
massimo.sgaravatto(a)gmail.com> wrote:
We have a single ceph cluster used by 2 openstack
installations.
We use different ceph pools for the 2 openstack clusters.
For nova, cinder and glance this is straightforward.
It was a bit more complicated fo radosgw. In this case the setup I used
was:
- creating 2 realms (one for each cloud)
- creating one zonegroup for each realm
- creating one zone for each zonegroup
- having 1 ore more rgw instances for each zone
I don't know if there are simpler approaches
Cheers, Massimo
On Tue, Sep 10, 2019 at 11:20 AM Wesley Peng <wesley(a)thepeng.eu> wrote:
ᐧ
on 2019/9/10 17:14, vladimir franciz blando wrote:
Sure you can. Ceph could be deployed as separate storage service,
openstack is just its customer. You can have one customer, but also can
have multi-customers for Ceph service.
regards.
_______________________________________________
ceph-users mailing list -- ceph-users(a)ceph.io
To unsubscribe send an email to ceph-users-leave(a)ceph.io
_______________________________________________
ceph-users mailing list -- ceph-users(a)ceph.io
To unsubscribe send an email to ceph-users-leave(a)ceph.io
I have 2 OpenStack environment that I want to
integrate to an existing
ceph cluster. I know technically it can be done but has anyone tried
this?
4:32 a.m.
New subject: 2 OpenStack environment, 1 Ceph cluster [EXT]
On Tue, Sep 10, 2019 at 05:14:34PM +0800, vladimir franciz blando wrote:
I have 2 OpenStack environment that I want to
integrate to an
existing ceph cluster. I know technically it can be done but has
anyone tried this?
Yes, it works fine. You need each OpenStack to have a different client
key, so that they can't trample on each other's pools; and you need some
sort of naming convention, so you can tell which pool belongs to which
OpenStack.
For example, our OpenStack deployments are named after Greek letters, so
we have ceph-ansible create the pools eta-images, eta-vms,
eta-volumes... zeta-images, zeta-vms, zeta-volumes for the two
deployments "eta" and "zeta". ceph-ansible also manages the different
client keys, "client.openstack-eta" and "client.openstack-zeta", with
permissions only for the appropriate pools.
We use Red Hat OpenStack (based on TripleO) so for each deployment there
is a yaml file describing where its Ceph is, e.g. for Eta it looks like
this:
parameter_defaults:
CephClusterFSID: '12341234-1234-1234-1234-123412341234'
CephClientKey: 'asdfasdfasdfasdfasdfasdfasdfasdfasdfas=='
CephExternalMonHost: '1.2.3.4, 5.6.7.8, 9.10.11.12'
NovaRbdPoolName: eta-vms
CinderRbdPoolName: eta-volumes
GlanceRbdPoolName: eta-images
CephClientUserName: openstack-eta
The FSID and mons will be the same for each deployment but the key and
pool names will be different.
Cheers,
Dave
--
** Dave Holland ** Systems Support -- Informatics Systems Group **
** 01223 496923 ** Wellcome Sanger Institute, Hinxton, UK **
--
The Wellcome Sanger Institute is operated by Genome Research
Limited, a charity registered in England with number 1021457 and a
company registered in England with number 2742969, whose registered
office is 215 Euston Road, London, NW1 2BE.
4:46 a.m.
New subject: 2 OpenStack environment, 1 Ceph cluster [EXT]
This is almost inline with how I did it before.. and i was using Red Hat OpenStack as
well.
________________________________
From: Dave Holland <dh3(a)sanger.ac.uk>
Sent: Tuesday, September 10, 2019 5:32 AM
To: vladimir franciz blando <vladimir.blando(a)gmail.com>
Cc: ceph-users(a)ceph.io <ceph-users(a)ceph.io>
Subject: [ceph-users] Re: 2 OpenStack environment, 1 Ceph cluster [EXT]
On Tue, Sep 10, 2019 at 05:14:34PM +0800, vladimir franciz blando wrote:
I have 2 OpenStack environment that I want to
integrate to an
existing ceph cluster. I know technically it can be done but has
anyone tried this?
Yes, it works fine. You need each OpenStack to have a different client
key, so that they can't trample on each other's pools; and you need some
sort of naming convention, so you can tell which pool belongs to which
OpenStack.
For example, our OpenStack deployments are named after Greek letters, so
we have ceph-ansible create the pools eta-images, eta-vms,
eta-volumes... zeta-images, zeta-vms, zeta-volumes for the two
deployments "eta" and "zeta". ceph-ansible also manages the different
client keys, "client.openstack-eta" and "client.openstack-zeta", with
permissions only for the appropriate pools.
We use Red Hat OpenStack (based on TripleO) so for each deployment there
is a yaml file describing where its Ceph is, e.g. for Eta it looks like
this:
parameter_defaults:
CephClusterFSID: '12341234-1234-1234-1234-123412341234'
CephClientKey: 'asdfasdfasdfasdfasdfasdfasdfasdfasdfas=='
CephExternalMonHost: '1.2.3.4, 5.6.7.8, 9.10.11.12'
NovaRbdPoolName: eta-vms
CinderRbdPoolName: eta-volumes
GlanceRbdPoolName: eta-images
CephClientUserName: openstack-eta
The FSID and mons will be the same for each deployment but the key and
pool names will be different.
Cheers,
Dave
--
** Dave Holland ** Systems Support -- Informatics Systems Group **
** 01223 496923 ** Wellcome Sanger Institute, Hinxton, UK **
--
The Wellcome Sanger Institute is operated by Genome Research
Limited, a charity registered in England with number 1021457 and a
company registered in England with number 2742969, whose registered
office is 215 Euston Road, London, NW1 2BE.
_______________________________________________
ceph-users mailing list -- ceph-users(a)ceph.io
To unsubscribe send an email to ceph-users-leave(a)ceph.io
1702
days inactive
1703
days old
5 comments
5 participants
participants (5)
-
Dave Holland -
Massimo Sgaravatto -
Matthew H -
vladimir franciz blando -
Wesley Peng