clustering suites for enterprise linux

[ishchattha]

HI everyone,   

I have been looking line and comparing feature of clustering suites for enterprise linux.  And I thought why not ask what everyone is using and get feed back on a couple that I have found.

Found the following  SLURM and clusterlabs (pacemaker and corosync).  Anyone played with them before?   Also was thinking of redhat cluster suite as well.

Let me know or give me some suggestions.

Thanks

Ish

[Mike Dupont]

Howdy,
check out http://mesos.apache.org/ it is very trendy. We will be
having talks about this at the kansaslinuxfest.us on march 21-22 in
Lawrence at the lib.

mike

> --
> You received this message because you are subscribed to the Google Groups
> "kulua-l" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to kulua-l+u...@googlegroups.com.
> To post to this group, send email to kul...@googlegroups.com.
> Visit this group at http://groups.google.com/group/kulua-l.
> For more options, visit https://groups.google.com/d/optout.



--
James Michael DuPont
Member of Free Libre Open Source Software Kosova http://www.flossk.org
Saving Wikipedia(tm) articles from deletion http://SpeedyDeletion.wikia.com

[Seth Galitzer]

I've been trying to use corosync and have always had trouble with it. It
seems to cause more problems than it solves for me.

Seth

> --
> You received this message because you are subscribed to the Google
> Groups "kulua-l" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to kulua-l+u...@googlegroups.com

> <mailto:kulua-l+u...@googlegroups.com>.

> To post to this group, send email to kul...@googlegroups.com

> <mailto:kul...@googlegroups.com>.

> Visit this group at http://groups.google.com/group/kulua-l.
> For more options, visit https://groups.google.com/d/optout.

--

Seth Galitzer
The beatings will continue until morale has improved.

[usererror]

After spinning up a few different configuration s of mesos with vagrant, xen, digital ocean droplets and on metal, I have to say it's easy to set up a development cluster. There are many frameworks for scheduling, messaging, distributed storage and even a script for routing native docker containers with haproxy. You should check it out!

To unsubscribe from this group and stop receiving emails from it, send an email to kulua-l+u...@googlegroups.com.
To post to this group, send email to kul...@googlegroups.com.

[Paul Johnson]

We run ROCKS cluster. It runs fine, scheduler is Maui/Torque
combination. Fine for traditional high performance computing.

Don't know if that's what you mean by "clustering", though.

On 02/09/2015 12:05 AM, ishchattha wrote:
> HI everyone,
> I have been looking line and comparing feature of clustering suites for
> enterprise linux. And I thought why not ask what everyone is using and
> get feed back on a couple that I have found.
>
> Found the following SLURM and clusterlabs (pacemaker and corosync).
> Anyone played with them before? Also was thinking of redhat cluster
> suite as well.
>
> Let me know or give me some suggestions.
> Thanks
> Ish
>
> --
> You received this message because you are subscribed to the Google
> Groups "kulua-l" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to kulua-l+u...@googlegroups.com
> <mailto:kulua-l+u...@googlegroups.com>.
> To post to this group, send email to kul...@googlegroups.com
> <mailto:kul...@googlegroups.com>.
> Visit this group at http://groups.google.com/group/kulua-l.
> For more options, visit https://groups.google.com/d/optout.

--

Paul E. Johnson University of Kansas
Professor Acting Director, Center for
Political Science Research Methods & Data Analysis
http://pj.freefaculty.org http://crmda.ku.edu
paul...@ku.edu pauljoh...@ku.edu
1541 Lilac Lane 1425 Jayhawk Blvd.
Blake Hall, Rm. 504 Watson Library, Rm. 470
Lawrence, Kansas Lawrence, Kansas
66045-3129 66045-7555
Ph: (785) 864-3353

[Billy Crook]

Rocks is HPC. corosync/heartbeat/pacemaker are HA.

How's rocks doing these days? I rather enjoyed the lengths they went
to automate reimaging.

> To unsubscribe from this group and stop receiving emails from it, send an email to kulua-l+u...@googlegroups.com.
> To post to this group, send email to kul...@googlegroups.com.

[usererror]

Ahh, been a while since I have played with rocks. Another cool thing about mesosphere is that it's a kernel for abstracting resources across heterogeneous  os and hardware and / or vm s. The system is capable of distributed stream and batch jobs with spark, storm, hadoop, etc. It's also developed to be elastic and highly available.  There is discussion of snapshots for apps in the future to save states. That's all I'll say about for now, it's got me quite excited.

[usererror]

Oh, and you might want to read the Google Omega white paper, if you haven't already. Cheers!

[Jim Ford]

Since no one has spoken up for SLURM, I suppose I will. I find it to be simple and robust. It is a little esoteric when first looking at it, but it was straightforward enough to set up.

We use it for a small build farm with buildbot scheduling up CIT and builds. It shares with our developer build and test pool as well and handles things seamlessly. Every once-in-a-while we have to reboot a node, but that has more to do with shaking out NFS cobwebs.

There are a ton of options I haven't had a chance to explore and from what I read it can reliably handle 1500+ nodes. For simple and reliable it does the trick for a group of about 80 devs. HTH! Good luck!

Jim

Sent from my iPhone

--

[ishchattha]

Thanks everyone for your inputs,  and other options to look at.   If anyone can think of another high availability clustering let me know.

[David Nicol]

have you looked into "direct return load balancing?"