[openstack-hpc] Looking for practical Openstack + Ceph guidance for shared HPC
Andrew J Younge
ajyounge at indiana.edu
Fri Jul 26 19:56:01 UTC 2013
Hi everybody,
Alright I'm going to jump in on this conversation. First I want to say
this is a great discussion, and I'm excited to see others as interested
in HPC and Openstack as I am. The questions raised regarding the
state-of-art in OpenStack related to high performance capabilities are
spot-on and I have no doubt many people on this list are working hard to
address them today.
Now, I'd actually like to vote against the thought of shoehorning
OpenStack into a typical HPC environment, as suggested in Josh's last
email. As alluded to, doing so would be impractical when considering HPC
scheduling, job reservation times etc, however it would be a downright
nightmare for supporting the software stack necessary for running a
cloud environment. I'd be impressed to find a single HPC system
administrator who'd allow you to install a type1 or even type2
hypervisor in a traditional compute node, let alone the proper higher
order APIs and software packages necessary to make it a reality. This is
where Cloud Iaas solutions provide a true advantage when running
scientific applications over conventional HPC offerings - users have the
freedom to design their environment as they see fit within a VM, a
luxury that's continually a limitation within normal HPC offerings
today. There have been many efforts over the years to layer
higher-order solutions atop traditional HPC and cluster LRMS solutions,
many of which were great but have lost a lot of traction in recent
years. I'll simply just reference the term Grid Computing and say little
more on the subject.
I would strongly argue that bridging the gap between HPC and IaaS should
(must?) be accomplished by providing HPC oriented tools, services, and
hardware within a user-centric IaaS solution like OpenStack, rather than
the other way around. The effort necessary to bring HPC in a Cloud IaaS
is much more tractable this way, as the mechanisms, the capabilities,
and recently the interest are all possible. Actually this is a core
concept of my tentative phd dissertation. There will always be an added
overhead with using virtualization technologies inherent within Cloud
IaaS, however the total overhead is continually decreasing and
capabilities are ever increasing, making IaaS a potential solution for
many mid-tier scientific research projects for the first time.
Regarding Ceph in OpenStack, I do think there is an effort for providing
Ceph within OpenStack already
http://ceph.com/docs/next/rbd/rbd-openstack/. Please note though that
this is distributed block storage for booting and running VMs in
OpenStack, and not distributed shared block storage within a VM itself.
Ideally, it would be great to have a many to 1 relationship with VMs
and volumes (Cinder, nova-volume, EBS), representing a distributed
shared storage among many running VMs simultaneously. There is a new
blueprint for this now
https://blueprints.launchpad.net/cinder/+spec/read-only-volumes however
this is using read-only volumes. My eye is focused on a solution
currently revolving around Lustre, but there's no theoretical reason
similar work couldn't be done or isn't already completed using Ceph.
For scheduling in OpenStack, this seems like a rather large problem with
many potential solutions. If the HPC community has taught us anything
about scheduling, its that its an NP-complete problem with an unlimited
number of potential solutions. There was an original proposal for a
proximity based scheduler
https://wiki.openstack.org/wiki/ProximityScheduler but this is just 1
theoretical opportunity of many. I'd be ecstatic to get involved in a
community effort to push for high performance scheduling in OpenStack
using whatever mechanisms people collectively see fit.
I've hinted at many different things here, so let me know if you'd like
to discuss anything in more detail. Also, if you want more information
on a solution for GPUs and/or InfiniBand in OpenStack as alluded to by
Brian and Nerayan, just let me know. I may have one.
Regards,
Andrew
P.S. Is anybody planning on attending the next OpenStack summit and want
to participate in a discussion on some of these topics?
http://www.openstack.org/summit/openstack-summit-hong-kong-2013/become-a-speaker/
The speaker proposal deadline is in 5 days I believe...
On 7/26/13 12:42 PM, Joshua Dotson wrote:
> All,
>
> Thanks for the great responses. I'm finding this discussion very
> enlightening. I have some mostly idle and uninformed musings about HPC
> node scheduling at the nova-scheduler level:
>
> Efficient handling of typical IaaS loads is much different from what I
> suppose one might call HPCaaS loads. For example, a typical IaaS
> instances is expected to be sometimes idle, awaiting clients and work,
> unless developers really tune their SaaS application stacks, if that's
> what your cloud is running. A good bit of virtualization is
> rarely-elastic "infrastructure", things like OwnCloud or a mirrors
> server. These tasks usually benefit from sharing a physical node.
>
> Beyond that, a lot of SaaS developers (who run on IaaS) stress the
> horizontal scaling before any large vertical scaling, because resiliency
> (especially geo-replication) and cost are very important there. Though
> there are definitely limits, instance granularity (smaller instances)
> and more horizontal than vertical footprints, for SaaS and PaaS stacks
> running in clouds which themselves have a horizontal footprint, are
> industry standard methods of achieving SLAs.
>
> I seldom hear of HPC jobs having such performance profiles, because
> performance, not resiliency, comes first. What you end up with is
> nova-scheduler being tuned out-of-the box to over commit on I/O, memory,
> disk and CPU and having little or no motive to change course to honor
> things like pre-planning the number of hours instances will run, unlike
> HPC resource schedulers -- unless I'm severely mistaken. The API itself
> would need extending to do that, which maybe isn't altogether impossible
> to bring about.
>
> Thus, my experience has been that cloud-centric load balancing is
> essentially sacrilegious to a lot of HPC (and also, importantly, "Big
> Data" -- if for throughput over resiliency/storage) tasks. Along these
> lines, in my view, local-to-node disk space, physical or otherwise, has
> little place in HPC, at scale (e.g. swap off). I'd really like to see
> things like CephFS be battle hardened enough to move us forward toward
> unified storage, of the cloud API variety, which can service HPC-level
> throughput. HPC jobs which are I/O, memory and/or CPU bound might find
> themselves benefiting from over committing, but only because backfilling
> with the opposite variet(ies) of bounding could make sense in such cases
> -- and only then if we're comfortable guessing that the job in question
> isn't going to shift gears and start chomping on memory, etc...
>
> So then, it's almost as if we need to put pop-up Openstack environments
> into an old school PBS or LSF, rather than put HPC(s) into an Openstack
> cloud. I guess what I'm saying is, maybe it would be easier to have a
> "job script" which bootstraps an single-tenant(?) Openstack environment
> at job run time, for those who need IaaS. But no, for many reasons, not
> the least of which is that keeping a perpetual "job" running would be a
> nightmare in old-school HPC land. Beyond that, I don't get the feeling
> old-school HPC job schedulers have any "elastic" or modern REST API
> abilities.
>
> It seems that Nova scheduler needs some form of tiered scheduling, if
> ever HPCaaS were to be really efficient. Nova needs to know how often I
> plan to peg I/O, memory, and other resources. While I suppose some
> machine learning could be introduced to profile each tenant and each of
> their deployments, I do not see us getting much traction with these
> things, until Openstack has a place in their API for a "cluster"
> primitive. Now, I've not been keeping up with details on the Heat
> project
> <http://www.openstack.org/summit/san-diego-2012/openstack-summit-sessions/presentation/heat-a-template-based-orchestration-engine-for-openstack>,
> but maybe that's just what the doctor ordered, as a base for HPCaaS on
> Openstack...
>
> Since Heat is clone of AWS Cloud Formation, this seems relevant:
> https://aws.amazon.com/hpc-applications/
>
> * "Cluster instances can be launched within a /Placement Group/. All
> instances launched within a Placement Group have low latency, full
> bisection 10 Gbps bandwidth between instances. Like many other
> Amazon EC2 resources, /Placement Groups/ are dynamic and can be
> resized if required. You can also connect multiple /Placement
> Groups/ to create very large clusters for massively parallel
> processing."
>
> Thanks again,
> Joshua
>
>
> On Fri, Jul 26, 2013 at 11:24 AM, Narayan Desai <narayan.desai at gmail.com
> <mailto:narayan.desai at gmail.com>> wrote:
>
> I think that workload management is perhaps too vague of a term
> here. There are a few different processes in play.
>
> One is the system level resource management (nova-scheduler, etc).
> This component needs to figure out how to allocate the resources to
> tenants, and pull them back with needed.
>
> Inside of a single tenant's allocation, there is a need for workload
> management. This is a good place for traditional HPC resource
> managers; we've run torque in this capacity, for example.
>
> I think that the model today for nova scheduler is wrong; it only
> supports open-ended leases on compute resources. Another component
> that makes this tricky is the need to support interactive workloads;
> batch is good for well, batch, but interactive is always problematic
> in these environments.
>
> There is some research targeting this problem, in particular Mesos
> and Omega. Mesos looks ok for serial workloads, but has a real
> problem with parallel (non-resizable) ones. Omega looks a lot
> better, but that is an internal google thing. I don't think that
> there is an off the shelf solution today.
> -nld
>
>
> On Fri, Jul 26, 2013 at 1:48 AM, Di Pe <dipeit at gmail.com
> <mailto:dipeit at gmail.com>> wrote:
>
> All,
>
> one issue Joshua touched on was options for workload management.
> While IB and GPU seem to frequently discussed when it comes to
> openstack and HPC they are not so relevant in our HPC
> environment (3000 ish cores, each node connected via 1G,
> scaleout NAS storage, biomedical research, genomics, proteomics,
> statistics) ..... many other midsize shops may be have similar
> setups. We are just starting to look at openstack for a for a
> potential deployment with Ubuntu 14.04. We have good experiences
> using KVM for some of our resources. Some of the things we are
> hoping to get from openstack in the future are:
>
> * flexible partitioning of resources for special sauce
> software (hadoop, interactive HPC software)
> * self service for developers and scientists
> * allow research group that spans multiple research
> organizations (internal / external) controlled access to an
> isolated (virtual) datacenter (potentially with fisma
> compliance)
> * save images that researchers built for later use
> (reproducible in case someone asks how they got to this result)
> * chargeback for HPC resources for internal and external users
> * usage of idle resources for testing in Enterprise IT, VDI ,etc
> * compute fencing (as we are heading to 24 cores per box most
> of our multi threaded code can still only take advantage of
> 4-6 cores. This either leaves stuff idle or users step on
> each other on shared nodes ....cgroups is a bit of a pain to
> maintain.
> * checkpointing and restarting long running jobs (for
> prioritization and better protection against job failures).
> perhaps LXC containers as alternative to KVM (we use BLCR
> today but that community is quite small)
> * standardization of our infrastructure
> * potential participation in futuregrid, XSEDE, etc
>
>
> That's perhaps a lot to ask but we would be looking at a 2-3
> year time frame. What I don't quite unterstand, how would one
> handle workload management. Currently we see people using SGE,
> moab, LSF and some slurm in classic HPC. Concepts like backfill,
> preemption fair share and such things are probably unknown to
> openstack? If so it would perhaps be acceptable to run a
> workload manager on a subset of always-on KVM systems or even
> baremetal for classic HPC. But how does one consolidate the
> reporting, billing, chargeback of 2 separate systems?
>
> Are there any efforts to integrate workload managers directly
> into nova? SGE and Slurm are both open source and would support
> everything we require. Or are folks thinking about writing
> something from scratch in python?
>
> Thanks
> dipe
>
>
>
>
>
> On Thu, Jul 25, 2013 at 6:22 PM, Narayan Desai
> <narayan.desai at gmail.com <mailto:narayan.desai at gmail.com>> wrote:
>
> Brian's right. You will end up doing a lot of work; nova
> isn't ready for this out of the box.
>
> The key problems are:
> - drivers for virtualization (either via SR-IOV or device
> passthrough) for net + gpu
> - IO architecture
>
> There is apparently SR-IOV support for IB, provided that you
> have the right hardware, firmware, and driver, though I
> haven't managed to make it work. This provides a pkey
> isolated multi-tenant environment. THat is basically a
> mellanox only solution.
>
> Like Brian said, Xen is the only way to go for GPU
> virtualization.
>
> You can do some really interesting things with the I/O
> architecture. We've been experimenting with both glusterfs
> and ceph. Both seem to perform decently well; we've managed
> to get the glusterfs setup going at 60 GB/s in aggregate
> with a pile of clients. There isn't good integration of all
> of the capabilities yet in mainline openstack, but this
> looks promising. Ceph looks like the mainline integration is
> better, but we haven't tried those things out yet.
>
> At the end of the day, you need to ask yourself why you want
> to accomplish this. If you're running a workload that is
> well suited to an HPC cluster, you should probably use one.
> If you need multi-tenancy, user control of system software,
> or need to run a workload poorly suited for a traditional
> cluster, then it is worth thinking strongly about. You'll
> end up needing to do a bunch of work though.
>
> In our case, the reason that we pursued this course is
> because we have workloads and developers that benefit from
> the cloud control plane.
>
> hth.
> -nld
>
>
>
> On Thu, Jul 25, 2013 at 3:30 PM, Brian Schott
> <brian.schott at nimbisservices.com
> <mailto:brian.schott at nimbisservices.com>> wrote:
>
> Joshua,
>
> It is something those of us working the gap between HPC
> and cloud have struggled with. We lack a strong project
> team within OpenStack IMHO, although there are several
> small groups doing HPC with OpenStack. Hopefully others
> will chime in on some other topics, such as Infiniband
> support, but we did some testing with a GRID K2 card for
> GPU pass-through with different hypervisors. A talk I
> gave at the OpenStack DC meetup is here:
> http://www.slideshare.net/bfschott/nimbis-schott-openstackgpustatus20130618
>
> The short GPU answer is that it is possible with Xen,
> XenCenter, and XCP to passthrough GPUs today, but
> OpenStack doesn't have support by default in Nova. This
> is still in a roll-your-own mode for deployment.
>
> Brian
>
> -------------------------------------------------
> Brian Schott, CTO
> Nimbis Services, Inc.
> brian.schott at nimbisservices.com
> <mailto:brian.schott at nimbisservices.com>
> ph: 443-274-6064 <tel:443-274-6064> fx: 443-274-6060
> <tel:443-274-6060>
>
>
>
> On Jul 25, 2013, at 4:06 PM, Joshua Dotson
> <josh at knoesis.org <mailto:josh at knoesis.org>> wrote:
>
>> Hello.
>>
>> A contingent of my organization, the Kno.e.sis Center
>> @ Wright State University <http://www.knoesis.org/>,
>> recently received a grant award which we intend to use
>> to support a handful of mid-size HPC-style workloads
>> (MPI <-- definitely, GPGPU <-- if possible/plausible)
>> in addition to many mid-size IaaS-style workloads
>> (MongoDB, Storm, Hadoop, many others). As a third
>> layer, I'm playing with the idea of evaluating an
>> elastic OpenShift Origin atop the same
>> infrastructure. Approximately $400k to $500k will
>> hopefully be available for this deployment, though
>> exact numbers are not yet available to me.
>>
>> While I'm prepared to build a home-grown
>> small-to-mid-size "classical" HPC, using modern
>> hardware, and a smaller silo for home-grown Openstack
>> for the minority stakeholders, I am hoping to find
>> ways of making proponents of both workloads
>> simultaneously happy, or close to it. That is, I
>> would like to give my computer scientist users a
>> friendly method of running their HPC-style jobs on a
>> combined performance-tuned silo of Openstack. Doing
>> so would load-balance the procured hardware and
>> infrastructure with the users who want a Tomcat or a
>> Virtuoso instance.
>>
>> I see a number of serious issues realizing such a
>> goal. For example, the state of Infiniband vs.
>> Openstack seems not quite
>> ready/available/documented/accessible for such use in
>> production, unless I'm just blind to the right blogs.
>> The added myriad abstractions and latency
>> virtualization might impose on an HPC task, not to
>> mention cloud software defined networking (Quantum,
>> especially when sans hardware acceleration), seem
>> likely to really get in the way of practicality,
>> economics and efficiency. That said, most of what we
>> do here isn't HPC, so I believe such trade-offs can be
>> agreed upon, if a reasonable job scheduling and
>> workload management mechanism can be found and agreed
>> upon by all stake holders, grant proposal majority
>> (HPC) and minority (IaaS) alike.
>>
>> I get the impression from my readings that HPC-style
>> deployment (separate from job queuing) against the EC2
>> API should work. I don't have a good feeling that the
>> experience would be particularly friendly, however,
>> without paying for closed source applications. I'm
>> thinking a high-performance Ceph install would help
>> bring up the storage end of things in a modern
>> open-source CoTS way. I've not done specific research
>> on Lustre + Openstack, but no reports of such a setup
>> have presented themselves to me, either.
>>
>> These blue sky ideas matter nil, it seems, if a
>> sufficiently-large high-performance production-quality
>> Openstack deployment is beyond the funds to be
>> allotted, which is something else I'm working on.
>> I've built smallish but useful virt-manager, oVirt and
>> Openstack environments here already, but none of them
>> are enough for the very-important HPC job proposed for
>> this grant. The scientist running the proposed
>> computation gave me following information to clarify
>> what would parity (for his job only) his experience
>> running the computation with an external HPC service
>> provider.
>>
>> * MPI
>> * 20 Gbps Infiniband compute interconnect
>> * 600 cores (those currently used are probably G4
>> Opteron 2.4 Ghz)
>> * 4 GB RAM per core
>> * at lest 2 TB shared storage, though I'm thinking
>> we need much more for use by our general community
>> * unsure of the storage networking topology
>>
>> We're in the shopping phase for this grant award and
>> are still playing with ideas. It seems likely to snap
>> back into an old-school HPC, at this time. I've
>> fielded communication about our needs to a number of
>> Openstack and hardware providers, on the hope that
>> they can bring something helpful to the table.
>>
>> Please let know if you can point me in the right
>> direction(s). I'm up to reading whatever text is
>> thrown at me on this topic. :-)
>>
>> Thanks,
>> Joshua
>> --
>> Joshua M. Dotson
>> Systems Administrator
>> Kno.e.sis Center
>> Wright State University - Dayton, OH
>> _______________________________________________
>> OpenStack-HPC mailing list
>> OpenStack-HPC at lists.openstack.org
>> <mailto:OpenStack-HPC at lists.openstack.org>
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
>
>
> _______________________________________________
> OpenStack-HPC mailing list
> OpenStack-HPC at lists.openstack.org
> <mailto:OpenStack-HPC at lists.openstack.org>
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
>
>
>
> _______________________________________________
> OpenStack-HPC mailing list
> OpenStack-HPC at lists.openstack.org
> <mailto:OpenStack-HPC at lists.openstack.org>
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
>
>
>
> _______________________________________________
> OpenStack-HPC mailing list
> OpenStack-HPC at lists.openstack.org
> <mailto:OpenStack-HPC at lists.openstack.org>
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
>
>
>
> _______________________________________________
> OpenStack-HPC mailing list
> OpenStack-HPC at lists.openstack.org
> <mailto:OpenStack-HPC at lists.openstack.org>
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
>
>
>
>
> --
> Joshua M. Dotson
> Systems Administrator
> Kno.e.sis Center
> Wright State University - Dayton, OH
> 937-985-3246 <tel:937-985-3246>
>
>
> _______________________________________________
> OpenStack-HPC mailing list
> OpenStack-HPC at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
>
--
Andrew J. Younge
Pervasive Tech Institute / School of Informatics & Computing
Indiana University / Bloomington, IN USA
ajyounge at indiana.edu / http://ajyounge.com
More information about the OpenStack-HPC
mailing list