XOSAN with XO Community edition
-
@olivier when I issue the command
git clone -b master https://github.com/vatesfr/xen-orchestra
orgit pull --ff-only
, isn't that going to pull down the absolute latest source from the repo, including any commits that have occurred since the last "release"?For example, XO 5.16 was released on 2/1 and there have been over 3800 commits since that time. How would I build a new VM with
xo-server 5.16.0
andxo-web 5.16.1
that doesn't include the subsequent commits? Ditto for updating an existing VM. -
@danp This is just because we merged in the mono repo. I don't think there is any issue to pull everything on master.
-
@olivier Technically, it's because you no longer use the concept of a
next-release
branch. :grinning_face_with_smiling_eyes:It just means that everyone could be running a slightly different version unless you are running XOA.
-
Yes, we switched to
master
, but it's been a while now (months?)Also, it's up to you to decide on what "head" (commit/tag/branch) to follow in your own scripts
¯\_(ツ)_/¯
-
@olivier Thanks for the feedback.
Ok... so I see that @julien-f just pushed out new versions of xo-server and xo-web. Anyone care to elaborate on how we could properly script this so that a week from now you could pull the source from this particular point in time?
-
@DustinB3403 StarWinds webpage does not list Citrix, but
StarWind Virtual SAN FREE vs. PAID AUGUST 2018
has a logo for CitrixReady and lists support for KVM and Xen Project VMs."Shared storage for KVM, Citrix XenServer and Xen Project VMs (HA iSCSI & CA SMB 3.1.1)" pg 5.
StarWinds free edition specifically mentions HyperV, Xen, KVM, and
vSphere and even supports Remote Direct Memory Access RDMA for serious disk speed up.Havent convinced management that $6000 for xosan is much less expensive than vmWare. And cannot prove that xosan is faster without much bigger virtual machine allowances in the trial version.
-
@rjt I'm not sure what to look at here. StarWind VSAN has always been available to be used on XenServer, so long as it was deployed on Windows VMs that were on your individual XS hypervisors.
Those VMs would then manage the storage level for you, a costly way to get vSAN, but one that would work for sure.
-
@DustinB3403 i was confirming and adding details to your reply to @Darek-Hamann in the number of hypervisors supported. It can be free, so you mean costly as in windows licensing or performance? If performance, i would hope RDMA would solve that.
-
@rjt said in XOSAN with XO Community edition:
@DustinB3403 i was confirming and adding details to your reply to @Darek-Hamann in the number of hypervisors supported. It can be free, so you mean costly as in windows licensing or performance? If performance, i would hope RDMA would solve that.
Windows licensing.
Microsoft is always expensive, and it's no exception here.
-
FYI we started very interesting discussions with Linbits guys (we could achieve something really powerful by integrating Linstore inside XCP-ng as a new hyperconverged solution). It means really decent perfs (almost same as local storage) and keep it robust and simple.
-
@olivier said in XOSAN with XO Community edition:
FYI we started very interesting discussions with Linbits guys (we could achieve something really powerful by integrating Linstore inside XCP-ng as a new hyperconverged solution). It means really decent perfs (almost same as local storage) and keep it robust and simple.
Linbit does some very fast stuff for sure.
-
Also we could achieve hyperconvergence "the other way" (unlike having a global shared filesystem like Gluster or Ceph) but use fine grained replication (per VM/VM disk). That's really interesting (data locality, tiering, thin pro etc.). Obviously, we'll collaborate to see how to integrate this in our stack