matt4
New Member.
789 views

OES Clustering under Hyper-V

Hello,

I have a site that has a 4-node OES cluster running on physical hardware. The servers and SAN are old and we are getting ready to migrate off them to a virtual environment. The virtual environment is Microsoft Hyper-V on a Nimble SAN. I'm debating whether or not to continue to bother with OES clustering when I setup the new VMs. This cluster is strictly used for file sharing. All the resources are virtual servers/volumes.

Has anyone done OES Clustering under Hyper-V? Does that work/is that even supported?

Any thoughts on if I should even bother with clustering or just rely on the fault tolerance and redundancy provided by the virtual environment?

Thanks.

Matt
Labels (1)
0 Likes
9 Replies
AutomaticReply Absent Member.
Absent Member.

Re: OES Clustering under Hyper-V

matt,

It appears that in the past few days you have not received a response to your
posting. That concerns us, and has triggered this automated reply.

These forums are peer-to-peer, best effort, volunteer run and that if your issue
is urgent or not getting a response, you might try one of the following options:

- Visit https://www.microfocus.com/support-and-services and search the knowledgebase and/or check
all the other self support options and support programs available.
- Open a service request: https://www.microfocus.com/support
- You could also try posting your message again. Make sure it is posted in the
correct newsgroup. (http://forums.microfocus.com)
- You might consider hiring a local partner to assist you.
https://www.partnernetprogram.com/partnerfinder/find.html

Be sure to read the forum FAQ about what to expect in the way of responses:
http://forums.microfocus.com/faq.php

Sometimes this automatic posting will alert someone that can respond.

If this is a reply to a duplicate posting or otherwise posted in error, please
ignore and accept our apologies and rest assured we will issue a stern reprimand
to our posting bot.

Good luck!

Your Micro Focus Forums Team
http://forums.microfocus.com



0 Likes
Knowledge Partner
Knowledge Partner

Re: OES Clustering under Hyper-V

In article <matt.8v9w8n@no-mx.forums.microfocus.com>, Matt wrote:
> Has anyone done OES Clustering under Hyper-V? Does that work/is that
> even supported?


While I've done a bit under vmWare and heard many having done so as
well, I haven't heard of any under Hyper-V yet. If you go that way,
please let us know. The key bit you would need is for Hyper-V to allow
raw simultaneous access LUNs
What version of OES are you running? You would have to make sure you
are matching supported versions of OES to your version of Hyper-V

What drives the requirement for clustering?
All the clustering systems I put in ended up dropping the clustering on
moving to vSphere, even the NetWare cluster that is still NetWare
today. The main reason being that under a proper vSphere system where
all the hardware is redundant, the minute or two time to reboot a vm
was perfectly acceptable vs the cost and complexity of clustering.

Rebooting a physical server too often needed a physical presence where
as a vm can be rebooted from anywhere you can get to your virtual
management console.


Andy of
http://KonecnyConsulting.ca in Toronto
Knowledge Partner
https://forums.novell.com/member.php/75037-konecnya
If you find a post helpful and are logged in the Web interface, please
show your appreciation by clicking on the star below. Thanks!

___
Andy of Konecny Consulting in Toronto
Knowledge Partner Profile
If you find a post helpful, click the Like button below. Thanks!
0 Likes
matt4
New Member.

Re: OES Clustering under Hyper-V

konecnya;2496176 wrote:
In article <matt.8v9w8n@no-mx.forums.microfocus.com>, Matt wrote:
> Has anyone done OES Clustering under Hyper-V? Does that work/is that
> even supported?


While I've done a bit under vmWare and heard many having done so as
well, I haven't heard of any under Hyper-V yet. If you go that way,
please let us know. The key bit you would need is for Hyper-V to allow
raw simultaneous access LUNs
What version of OES are you running? You would have to make sure you
are matching supported versions of OES to your version of Hyper-V

What drives the requirement for clustering?
All the clustering systems I put in ended up dropping the clustering on
moving to vSphere, even the NetWare cluster that is still NetWare
today. The main reason being that under a proper vSphere system where
all the hardware is redundant, the minute or two time to reboot a vm
was perfectly acceptable vs the cost and complexity of clustering.

Rebooting a physical server too often needed a physical presence where
as a vm can be rebooted from anywhere you can get to your virtual
management console.


Andy of
http://KonecnyConsulting.ca in Toronto
Knowledge Partner
https://forums.novell.com/member.php/75037-konecnya
If you find a post helpful and are logged in the Web interface, please
show your appreciation by clicking on the star below. Thanks!



Thanks for the reply.

I'm replacing a PHYSICAL OES 2015 4-node cluster with a virtual OES 2018 SP1 cluster. We are retiring the hardware, including the old SAN (btw, most of the cluster nodes have been up 749 days now, let's see a Windows server do that!).

Honestly, my thoughts mirror yours, there seems to be little point in OES clustering under a virtual environment. I was pretty much set against doing it and just moving to standalone VMs and relying on the fault tolerance and redundancy in the Hyper-V environment. But I figured I'd ask to see if anyone could convince me otherwise (kinda surprised on how few replies!). I've done it under VMWare and then only in lab environments. I've never had a production OES Cluster under VMware, let alone Hyper-V. The fact that no one has done it under Hyper-V from what I can tell is further driving me away from even attempting it.

About the only thing I can possibly figure would be an advantage would be the ability to patch and/or reboot VMs during the day since we could move the volumes around. But that is really a "nit."

I also don't even know if Hyper-V allows for raw simultaneous LUNs, that may be the show stopper right there quite honestly.

Part of me wants to try it in a test/lab just to see if I can make it work under Hyper-V! 🙂

This site already runs lots of SLES 11, 12, and 15 under Hyper-V, including OES 2015 through OES 2018 SP1, so that is all supported.


Matt
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: OES Clustering under Hyper-V

Am 10.02.19 um 23:16 schrieb matt:
(...)
> I have a site that has a 4-node OES cluster running on physical
> hardware. The servers and SAN are old and we are getting ready to
> migrate off them to a virtual environment. The virtual environment is
> Microsoft Hyper-V on a Nimble SAN. I'm debating whether or not to
> continue to bother with OES clustering when I setup the new VMs. This
> cluster is strictly used for file sharing. All the resources are virtual
> servers/volumes.
>
> Has anyone done OES Clustering under Hyper-V? Does that work/is that
> even supported?


I have a site which runs a 2-Node OES Cluster under Hyper-V. At the
first attempt there was a few performance issues. But after update all
(Firmware and HyperV-SW) it runs without problems.

> Any thoughts on if I should even bother with clustering or just rely on
> the fault tolerance and redundancy provided by the virtual environment?

a cluster virtual environment is not the ame as clustered services! If
the host get ill you can move this not functional VM to an other Host at
the HV, but after this the host is ill anywhere.

In case you have clustered services, you can move these service to
another cluster node an can repair the ill node without any headache
about the accessability of the service.

I always set up a NCS. IMHO it is the perfect way to offer services in
the LAN!

Bernd
0 Likes
Knowledge Partner
Knowledge Partner

Re: OES Clustering under Hyper-V

In article <matt.8w7bun@no-mx.forums.microfocus.com>, Matt wrote:
> About the only thing I can possibly figure would be an advantage would
> be the ability to patch and/or reboot VMs during the day since we could
> move the volumes around. But that is really a "nit."

Patching and rebooting remotely in the evening easy enough, relaxing even.
Just make sure you have a secure way in (VPN!) and not via a publically
exposed RDP (even if with a changed port)

> I also don't even know if Hyper-V allows for raw simultaneous LUNs, that
> may be the show stopper right there quite honestly.

I think Bernd answers that one nicely.

> Part of me wants to try it in a test/lab just to see if I can make it
> work under Hyper-V! 🙂

Test/lab is the best way to learn and hone ones skills.

> This site already runs lots of SLES 11, 12, and 15 under Hyper-V,
> including OES 2015 through OES 2018 SP1, so that is all supported.

SLES is the top distro supported by Microsoft for many years so this is
expected. The original formal agreement that allowed for that was met by
much disdain by a number of open source advocates.

One thing I have done at many of my clients is to make sure we don't run
too many services on each server so that for example: if you have to fix
an iprint, then rebooting that one guest doesn't impact file services.
Having a local eDir master that is the primary LDAP lookups is another
segregation that can often be patched and rebooted during light business
hours.


Andy of
http://KonecnyConsulting.ca in Toronto
Knowledge Partner
https://forums.novell.com/member.php/75037-konecnya
If you find a post helpful and are logged in the Web interface, please
show your appreciation by clicking on the star below. Thanks!

___
Andy of Konecny Consulting in Toronto
Knowledge Partner Profile
If you find a post helpful, click the Like button below. Thanks!
0 Likes
Knowledge Partner
Knowledge Partner

Re: OES Clustering under Hyper-V

In article <pY7eE.1305$h_7.510@novprvlin0913.provo.novell.com>, Nntp-user
wrote:
> > Any thoughts on if I should even bother with clustering or just rely on
> > the fault tolerance and redundancy provided by the virtual environment?

> a cluster virtual environment is not the ame as clustered services! If
> the host get ill you can move this not functional VM to an other Host at
> the HV, but after this the host is ill anywhere.
>
> In case you have clustered services, you can move these service to
> another cluster node an can repair the ill node without any headache
> about the accessability of the service.


It is possible to have an ill cluster resource (been there done that, and
you can see in this forum others have as well), though not quite as likely
as an ill guest that stays ill on a reboot. Since patches tend to be the
most likely cause of an ill guest, snap-shot before patching and restore
the snap-shot if that is the case.

> I always set up a NCS. IMHO it is the perfect way to offer services in
> the LAN!

If there are no resource ($$$) limitations, I agree fully. The challenge
is that cost that businesses keep wanting reduced.
Cluster Services does have quite a few costs to consider. Licensing is only
one factor, the skills to manage and troubleshoot aren't that common and
can become quite an issue when things happen when the one person who
understands it is not available ("what do you mean I have to be able to
connect in to fix things while on Vacation|during surgery?")

When P2Ving a NetWare cluster last summer that was a big point of
discussion with the client. We could have easily configure their vSphere
system to support clustering, but when they saw what needed to be set,
their lead vmware guy pointed out that they are used to being able to
migrating live guests and storage between vSphere clusters which would tend
to break a cluster. Because other vmware techs do work on their system
(what do you mean we need to schedule an outage just to move this pair of
guests?), this became a big worry, and since the original cluster was
primarily for hardware redundancy with a few minute reboot not being a
problem, clustering lost out. On principal(and they agreed) I did preserve
all the cluster bits so if they ever changed their minds I could connect
them again.

Uptime: how many 9's can you afford to buy? Each additional 9 costs notably
more than the previous ones.
I suspect that if you really need clustering, then BCC (Business Continuity
Clustering) is also in the equation for you.


Andy of
http://KonecnyConsulting.ca in Toronto
Knowledge Partner
https://forums.novell.com/member.php/75037-konecnya
If you find a post helpful and are logged in the Web interface, please show
your appreciation by clicking on the star below. Thanks!

___
Andy of Konecny Consulting in Toronto
Knowledge Partner Profile
If you find a post helpful, click the Like button below. Thanks!
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: OES Clustering under Hyper-V

Am 04.03.19 um 05:17 schrieb Andy Konecny:
> In article <pY7eE.1305$h_7.510@novprvlin0913.provo.novell.com>, Nntp-user
> wrote:

(...)
> though not quite as likely
> as an ill guest that stays ill on a reboot. Since patches tend to be the
> most likely cause of an ill guest, snap-shot before patching and restore
> the snap-shot if that is the case.

Never do that if the node runs eDir!

>> I always set up a NCS. IMHO it is the perfect way to offer services in
>> the LAN!

> If there are no resource ($$$) limitations, I agree fully. The challenge
> is that cost that businesses keep wanting reduced.

AFAIK is a 2-node cluster license included in OES. (Gloomy voice:
"Because only two nodes are for whimps." 😉

(...)
> When P2Ving a NetWare cluster last summer that was a big point of
> discussion with the client. We could have easily configure their vSphere
> system to support clustering, but when they saw what needed to be set,
> their lead vmware guy pointed out that they are used to being able to
> migrating live guests and storage between vSphere clusters which would tend
> to break a cluster. Because other vmware techs do work on their system
> (what do you mean we need to schedule an outage just to move this pair of
> guests?), this became a big worry, and since the original cluster was
> primarily for hardware redundancy with a few minute reboot not being a
> problem, clustering lost out. On principal(and they agreed) I did preserve
> all the cluster bits so if they ever changed their minds I could connect
> them again.


Huuh ... sorry, my english is not good enought for a extendet
disscusion.So only a few sentences. My expirience is: Fix each
OES-Cluster node to one ESX-Cluster-Node for normal production and
forget anything about 'now we can not move the guests' from ESX guys.

If you need to free one ESX-Node move all resources form the OES
clusternode at this ESX-Node to the other OES cluster node (migrate or
cluster leave) and should down this node. After you do your work at the
ESX node and it is running again, start the OES-node.

If it is neccessary to move a OES cluster node immediately, do that.
There are two possible results:

a) The ESX-Cluster is a good one, than the transfer is quick enought and
nothing happens to the connection.

b) The ESX-Cluster is slow, not perfect configured, has a slow SAN/LAN, ...
Then the OES-Cluster will check that there is one Node missing and
migrate the resources from that OES cluster node to the other node. The
clients will reconnect and ... thats all.

OK, if there are problems with faulty Wkstn/OES-Client-Settings/Odd
Software possibly the reconnect does not smart at the beginning. But
than this should be you playground to correct the environment.

From My POV the OES-Cluster is one of the easiest manageable one.

If it is neccessary to free a ESX node each day than ... 😉

Bernd
0 Likes
Knowledge Partner
Knowledge Partner

Re: OES Clustering under Hyper-V

In article <ck8fE.1317$h_7.608@novprvlin0913.provo.novell.com>, Nntp-user
wrote:
> AFAIK is a 2-node cluster license included in OES. (Gloomy voice:
> "Because only two nodes are for whimps." 😉

That is my understanding as well, thought last done on OES2
The gloomy voice is about right, have done up to 6 nodes, that I got to mirror
even the SBD as part of a migration to a new SAN (because 'best practise'
changed like they do)
The issue with 2 node is that if you have one down for maintenance, and the
other crashes, you have a notable unscheduled outage.

...
> Huuh ... sorry, my english is not good enought for a extended
> disscusion.So only a few sentences. My expirience is: Fix each
> OES-Cluster node to one ESX-Cluster-Node for normal production and
> forget anything about 'now we can not move the guests' from ESX guys.

No need to apologize, your English is way better than my French even though I
grew up having to learn it in Montreal. Though the bit of French I can speak
has the quebecois accent (notably different than Parisian) down pat.
Nailing a guest to a specific host was the problem that particular client had,
just doesn't work with their DR plans or habits.

> If you need to free one ESX-Node move all resources form the OES
> clusternode at this ESX-Node to the other OES cluster node (migrate or
> cluster leave) and should down this node. After you do your work at the
> ESX node and it is running again, start the OES-node.

A very logical approach that I would do as well. Challenge is that it beats up
against vSphere common practise that has to be faced

> From My POV the OES-Cluster is one of the easiest manageable one.

Oh yes, I always found it a joy to work with. Challenge has always been that
the clients don't see that part but the additional complexities to implement
it, especially when vSphere covers their real needs.



Andy of
http://KonecnyConsulting.ca in Toronto
Knowledge Partner
https://forums.novell.com/member.php/75037-konecnya
If you find a post helpful and are logged in the Web interface, please show
your appreciation by clicking on the star below. Thanks!

___
Andy of Konecny Consulting in Toronto
Knowledge Partner Profile
If you find a post helpful, click the Like button below. Thanks!
0 Likes
mmccaffe1 Absent Member.
Absent Member.

Re: OES Clustering under Hyper-V

Hi,

I do not use Hyper-V, but I see no reason why it would not work. I use VMware, Nimble, OES2015 clusters. I use the free 2 node clusters.
I use the clusters to make my patching simpler. I can not patch the one cluster node and then migrate resources to it and let that run for a week or so to make sure there are no patch bugs, if there are I can migrate back to the unpatched node. If all is good I patch the second node in the cluster. This has allowed me to patch during the day and not have to work off hours to do patching.

The cluster also made it easier for me to migrate to OES2018. I just remove node OES2015 node and replace it with a fresh 2018 node and migrate the resources to the 2018 node. Then I can remove the other OES2015 node and replace it with a fresh OES 2018 node. And I can do most of the work during regular hours, has saved me a log of time.

I have also used VMware Raw volumes, as well as vmdk file based disks, and OES iscsi client to access the Nimble SAN without issue.

So if you are fine with OES clusters, I would continue with them. The Virtualization takes care of the rest of high availability.
(BTW, cluster volumes can be loaded on a non-clustered server in an emergency by un-sharing the devices in NNMU.)

Good luck!

-Marty-
0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.