How are you using OpenStack for your HPC workflow? (We're looking at Ironic!)
Hi, everybody!
We at CU-Boulder Research Computing are looking at using OpenStack Ironic for our HPC cluster(s). I’ve got a sysadmin background, so my initial goal is to simply use OpenStack to deploy our traditional queueing system (Slurm), and basically use OpenStack (w/Ironic) as a replacement for Cobbler, xcat, perceus, hand-rolled PXE, or whatever else sysadmin-focused PXE-based node provisioning solution one might use. That said, once we have OpenStack (and have some competency with it), we’d presumably be well positioned to offer actual cloud-y virtualization to our users, or even offer baremetal instances for direct use for some of our (presumably most trusted) users.
But how about the rest of you? Are you using OpenStack in your HPC workloads today? Are you doing “traditional” virtualization? Are you using the nova-compute baremetal driver? Are you using Ironic? Something even more exotic?
And, most importantly, how has it worked for you?
I’ve finally got some nodes to start experimenting with, so I hope to have some real hands-on experience with Ironic soon. (So far, I’ve only been able to do traditional virtualization on some workstations.)
~jonathon
Hi Jonathon,
We are in a similar position at UAB and investigating a solution with openstack managed hw-level provisioning as well. I'm interested in being able to instantiate a cluster via OpenStack on select hardware fabrics using Ironic.
We currently have a traditional ROCKS HPC fabric augmented by Kickstart and Puppet for additional services (NFS servers, monitoring, documentation, project tools etc) spread out over hardware and VMs. We began adding VM services into the cluster to support ancillary tools and newer use cases like web-front ends to HPC job submission for select science domains.
We added an OpenStack fabric in 2013 to simplify such services and make them available to end users to allow greater flexibility and autonomy. it has been invaluable as a way to instantiate new services and other test configurations quickly and easily. This page gives a bit of background and and overview.
https://dev.uabgrid.uab.edu/wiki/OpenStackPlusCeph
My goal is to have a generally available openstack fabric that can serve our research community with Amazon-like functionality, allowing them to build their own scale outs if desired and potentially even help us manage the "HPC engine" (replacing the ROCKS mode with more cloud-oriented provisioning). There's another thread that is looking to Puppet+Foreman to handle the hardware provisioning -- building on our puppet effort. We also have a Chef+Crowbar fabric that provides similar services for our existing OpenStack deploy.
Right now we are planning our migration from Essex to Juno++. This is presenting it's own challenges but gives us a chance to build from a clean foundation. I haven't played with Ironic yet but it is definitely a tool I am interested in.
I'd like to be able to support VM containers for HPC jobs but can't determine yet if these should run inside the OpenStack provisioned cluster as traditional HPC-like jobs or allocate resources to them from raw hardware pool or as VMs (depending on requirements). It's probably a matter of perspective.
~jpr
On 01/13/2015 01:13 PM, Jonathon A Anderson wrote:
Hi, everybody!
We at CU-Boulder Research Computing are looking at using OpenStack Ironic for our HPC cluster(s). I’ve got a sysadmin background, so my initial goal is to simply use OpenStack to deploy our traditional queueing system (Slurm), and basically use OpenStack (w/Ironic) as a replacement for Cobbler, xcat, perceus, hand-rolled PXE, or whatever else sysadmin-focused PXE-based node provisioning solution one might use. That said, once we have OpenStack (and have some competency with it), we’d presumably be well positioned to offer actual cloud-y virtualization to our users, or even offer baremetal instances for direct use for some of our (presumably most trusted) users.
But how about the rest of you? Are you using OpenStack in your HPC workloads today? Are you doing “traditional” virtualization? Are you using the nova-compute baremetal driver? Are you using Ironic? Something even more exotic?
And, most importantly, how has it worked for you?
I’ve finally got some nodes to start experimenting with, so I hope to have some real hands-on experience with Ironic soon. (So far, I’ve only been able to do traditional virtualization on some workstations.)
~jonathon
OpenStack-HPC mailing list OpenStack-HPC@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
At CERN, we have taken two approaches for allocating compute resources on a single cloud
- Run classic high throughput computing using a batch system (LSF in our case) using virtualised resources. There are some losses from the memory needs of the hypervisor (so you have a slot or so less per machine) and some local I/O impacts but other than that, there are major benefits of being able to automate recovery/recycling of VMs such as for security of hardware issues.
- Run cloud services ala Amazon. Here users have one of the standard images such as CentOS or their own if they wish with cloud-init.
Compute resources at CERN are allocated using a pledge system, i.e. the experiments request and justify their needs for the year, resources are purchased and then allocated out according to these pledges. There is no charging as such.
The biggest challenges we've faced are
- Elasticity of cloud - the perception from Amazon is that you can scale up and down. Within a standard private cloud on-premise, the elasticity can only come from removing other work (since we aim to be using the resources to the full). We use short queue, opportunistic batch work to fill in so that we drain and re-instantiate the high throughput computing batch workers to accommodate some elasticity but it is limited. Spot market functionality would be interesting but we've not seen something in OpenStack yet.
- Scheduling - We have more work to do that there are resources. The cloud itself has no 'queue' or fair share. The experiment workflows thus have to place their workload into the cloud within their quota and maintain a queue or work on their side. INFN are working on some enhancements to OpenStack with Blazar or Nova queues which is worth following.
- Quota - Given the fixed resources, the quota controls are vital to avoid overcommitting. Currently, quotas are flat so the cloud administrators are asked to adjust the quotas to balance the user priorities within their overall pledges. The developments in Nested Projects coming along with Kilo will be a major help here and we're working with BARC to deliver this in Nova so an experiment resource co-ordinator can be given the power to manage quotas for their subgroups.
- VM Lifecycle - We have around 200 arrivals and departures a month. Without a credit card, it would be easy for their compute resources to remain running after they have left. We have an automated engine which ensures that VMs of departing staff are quiesced and deleted following a standard lifecycle.
- Accounting - The cloud accounting systems are based on vCPUs. There is no concept in OpenStack of a relative performance measure so you could be allocated a VM on 3 year old hardware or the latest and the vCPU metric is the same. In the high throughput use cases, there should be a relative unit which is used to scale with regards to the amount of work that could be done. With over 20 different hardware configurations running (competitive public procurement cycles over 4 years), we can't define 100 flavors with different accounting rates and expect our users to guess which ones have capacity.
We're starting to have a look at bare metal and containers too. Using OpenStack as an overall resource allocation system to ensure all compute usage is accounted and managed is of great interest. The highlighted items above will remain but hopefully we are work with others in the community to address them (as we are with nested projects)
Tim
-----Original Message----- From: Jonathon A Anderson [mailto:Jonathon.Anderson@Colorado.EDU] Sent: 13 January 2015 20:14 To: openstack-hpc@lists.openstack.org Subject: [openstack-hpc] How are you using OpenStack for your HPC workflow? (We're looking at Ironic!)
Hi, everybody!
We at CU-Boulder Research Computing are looking at using OpenStack Ironic for our HPC cluster(s). I’ve got a sysadmin background, so my initial goal is to simply use OpenStack to deploy our traditional queueing system (Slurm), and basically use OpenStack (w/Ironic) as a replacement for Cobbler, xcat, perceus, hand-rolled PXE, or whatever else sysadmin-focused PXE-based node provisioning solution one might use. That said, once we have OpenStack (and have some competency with it), we’d presumably be well positioned to offer actual cloud-y virtualization to our users, or even offer baremetal instances for direct use for some of our (presumably most trusted) users.
But how about the rest of you? Are you using OpenStack in your HPC workloads today? Are you doing “traditional” virtualization? Are you using the nova- compute baremetal driver? Are you using Ironic? Something even more exotic?
And, most importantly, how has it worked for you?
I’ve finally got some nodes to start experimenting with, so I hope to have some real hands-on experience with Ironic soon. (So far, I’ve only been able to do traditional virtualization on some workstations.)
~jonathon
OpenStack-HPC mailing list OpenStack-HPC@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
Tim Bell wrote:
- Accounting - The cloud accounting systems are based on vCPUs. There is no concept in OpenStack of a relative performance measure so you could be allocated a VM on 3 year old hardware or the latest and the vCPU metric is the same. In the high throughput use cases, there should be a relative unit which is used to scale with regards to the amount of work that could be done. With over 20 different hardware configurations running (competitive public procurement cycles over 4 years), we can't define 100 flavors with different accounting rates and expect our users to guess which ones have capacity.
This is quite interesting. I guess that CERN do not use the VUP (Vax Unit of Power) any more! Could you not use some sort of virtual currency, translating time occupied on a server X HEPSpec rating of that server into a virtual amount of money or units?
On 14 January 2015 at 06:24, Tim Bell Tim.Bell@cern.ch wrote:
At CERN, we have taken two approaches for allocating compute resources on a single cloud
- Run classic high throughput computing using a batch system (LSF in our
case) using virtualised resources. There are some losses from the memory needs of the hypervisor (so you have a slot or so less per machine) and some local I/O impacts but other than that, there are major benefits of being able to automate recovery/recycling of VMs such as for security of hardware issues.
- Run cloud services ala Amazon. Here users have one of the standard
images such as CentOS or their own if they wish with cloud-init.
Compute resources at CERN are allocated using a pledge system, i.e. the experiments request and justify their needs for the year, resources are purchased and then allocated out according to these pledges. There is no charging as such.
The biggest challenges we've faced are
- Elasticity of cloud - the perception from Amazon is that you can scale
up and down. Within a standard private cloud on-premise, the elasticity can only come from removing other work (since we aim to be using the resources to the full). We use short queue, opportunistic batch work to fill in so that we drain and re-instantiate the high throughput computing batch workers to accommodate some elasticity but it is limited. Spot market functionality would be interesting but we've not seen something in OpenStack yet.
- Scheduling - We have more work to do that there are resources. The cloud
itself has no 'queue' or fair share. The experiment workflows thus have to place their workload into the cloud within their quota and maintain a queue or work on their side. INFN are working on some enhancements to OpenStack with Blazar or Nova queues which is worth following.
- Quota - Given the fixed resources, the quota controls are vital to avoid
overcommitting. Currently, quotas are flat so the cloud administrators are asked to adjust the quotas to balance the user priorities within their overall pledges. The developments in Nested Projects coming along with Kilo will be a major help here and we're working with BARC to deliver this in Nova so an experiment resource co-ordinator can be given the power to manage quotas for their subgroups.
- VM Lifecycle - We have around 200 arrivals and departures a month.
Without a credit card, it would be easy for their compute resources to remain running after they have left. We have an automated engine which ensures that VMs of departing staff are quiesced and deleted following a standard lifecycle.
- Accounting - The cloud accounting systems are based on vCPUs. There is
no concept in OpenStack of a relative performance measure so you could be allocated a VM on 3 year old hardware or the latest and the vCPU metric is the same. In the high throughput use cases, there should be a relative unit which is used to scale with regards to the amount of work that could be done. With over 20 different hardware configurations running (competitive public procurement cycles over 4 years), we can't define 100 flavors with different accounting rates and expect our users to guess which ones have capacity.
We're starting to have a look at bare metal and containers too. Using OpenStack as an overall resource allocation system to ensure all compute usage is accounted and managed is of great interest. The highlighted items above will remain but hopefully we are work with others in the community to address them (as we are with nested projects)
Tim
-----Original Message----- From: Jonathon A Anderson [mailto:Jonathon.Anderson@Colorado.EDU] Sent: 13 January 2015 20:14 To: openstack-hpc@lists.openstack.org Subject: [openstack-hpc] How are you using OpenStack for your HPC
workflow?
(We're looking at Ironic!)
Hi, everybody!
We at CU-Boulder Research Computing are looking at using OpenStack Ironic for our HPC cluster(s). I’ve got a sysadmin background, so my initial
goal is to
simply use OpenStack to deploy our traditional queueing system (Slurm),
and
basically use OpenStack (w/Ironic) as a replacement for Cobbler, xcat,
perceus,
hand-rolled PXE, or whatever else sysadmin-focused PXE-based node provisioning solution one might use. That said, once we have OpenStack
(and
have some competency with it), we’d presumably be well positioned to
offer
actual cloud-y virtualization to our users, or even offer baremetal
instances for
direct use for some of our (presumably most trusted) users.
But how about the rest of you? Are you using OpenStack in your HPC
workloads
today? Are you doing “traditional” virtualization? Are you using the
nova-
compute baremetal driver? Are you using Ironic? Something even more
exotic?
And, most importantly, how has it worked for you?
I’ve finally got some nodes to start experimenting with, so I hope to
have some
real hands-on experience with Ironic soon. (So far, I’ve only been able
to do
traditional virtualization on some workstations.)
~jonathon
OpenStack-HPC mailing list OpenStack-HPC@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
OpenStack-HPC mailing list OpenStack-HPC@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
As you mention, our current unit is HEP-Spec06 which is based on a subset of the Spec2006 benchmark suite.
From the cloud accounting, we just get given the wall clock and the CPU time used. However, depending on the performance rating of the machine, the value of this time slot to the physicist can differ significantly.
Finding out which hypervisor the guest is running during that time window is also not available (AFAIK) in the accounting record. So, finding the underlying hardware is not possible from the pure accounting data (so an external mapping is needed such as the nova db)
We had thought about having some way that ceilometer would be informed about the rating of the machine and then report this as part of the record to the accounting system which could then choose whether to scale the values as part of the accounting process.
Things get really complicated when you have potentially overcommitted or undercommitted hypervisors. In the case of overcommit, clearly the vCPU is worth less than the reference benchmark. In the case of undercommit, the most recent CPUs can do various frequency scaling tricks.
Not an easy problem to solve, even in the simple case.
Tim
From: John Hearns [mailto:hearnsj@googlemail.com] Sent: 14 January 2015 12:37 To: openstack-hpc@lists.openstack.org Subject: Re: [openstack-hpc] How are you using OpenStack for your HPC workflow? (We're looking at Ironic!)
Tim Bell wrote:
- Accounting - The cloud accounting systems are based on vCPUs. There is no concept in OpenStack of a relative performance measure so you could be allocated a VM on 3 year old hardware or the latest and the vCPU metric is the same. In the high throughput use cases, there should be a relative unit which is used to scale with regards to the amount of work that could be done. With over 20 different hardware configurations running (competitive public procurement cycles over 4 years), we can't define 100 flavors with different accounting rates and expect our users to guess which ones have capacity.
This is quite interesting. I guess that CERN do not use the VUP (Vax Unit of Power) any more! Could you not use some sort of virtual currency, translating time occupied on a server X HEPSpec rating of that server into a virtual amount of money or units?
On 14 January 2015 at 06:24, Tim Bell <Tim.Bell@cern.chmailto:Tim.Bell@cern.ch> wrote:
At CERN, we have taken two approaches for allocating compute resources on a single cloud
- Run classic high throughput computing using a batch system (LSF in our case) using virtualised resources. There are some losses from the memory needs of the hypervisor (so you have a slot or so less per machine) and some local I/O impacts but other than that, there are major benefits of being able to automate recovery/recycling of VMs such as for security of hardware issues.
- Run cloud services ala Amazon. Here users have one of the standard images such as CentOS or their own if they wish with cloud-init.
Compute resources at CERN are allocated using a pledge system, i.e. the experiments request and justify their needs for the year, resources are purchased and then allocated out according to these pledges. There is no charging as such.
The biggest challenges we've faced are
- Elasticity of cloud - the perception from Amazon is that you can scale up and down. Within a standard private cloud on-premise, the elasticity can only come from removing other work (since we aim to be using the resources to the full). We use short queue, opportunistic batch work to fill in so that we drain and re-instantiate the high throughput computing batch workers to accommodate some elasticity but it is limited. Spot market functionality would be interesting but we've not seen something in OpenStack yet.
- Scheduling - We have more work to do that there are resources. The cloud itself has no 'queue' or fair share. The experiment workflows thus have to place their workload into the cloud within their quota and maintain a queue or work on their side. INFN are working on some enhancements to OpenStack with Blazar or Nova queues which is worth following.
- Quota - Given the fixed resources, the quota controls are vital to avoid overcommitting. Currently, quotas are flat so the cloud administrators are asked to adjust the quotas to balance the user priorities within their overall pledges. The developments in Nested Projects coming along with Kilo will be a major help here and we're working with BARC to deliver this in Nova so an experiment resource co-ordinator can be given the power to manage quotas for their subgroups.
- VM Lifecycle - We have around 200 arrivals and departures a month. Without a credit card, it would be easy for their compute resources to remain running after they have left. We have an automated engine which ensures that VMs of departing staff are quiesced and deleted following a standard lifecycle.
- Accounting - The cloud accounting systems are based on vCPUs. There is no concept in OpenStack of a relative performance measure so you could be allocated a VM on 3 year old hardware or the latest and the vCPU metric is the same. In the high throughput use cases, there should be a relative unit which is used to scale with regards to the amount of work that could be done. With over 20 different hardware configurations running (competitive public procurement cycles over 4 years), we can't define 100 flavors with different accounting rates and expect our users to guess which ones have capacity.
We're starting to have a look at bare metal and containers too. Using OpenStack as an overall resource allocation system to ensure all compute usage is accounted and managed is of great interest. The highlighted items above will remain but hopefully we are work with others in the community to address them (as we are with nested projects)
Tim
-----Original Message----- From: Jonathon A Anderson [mailto:Jonathon.Anderson@Colorado.EDUmailto:Jonathon.Anderson@Colorado.EDU] Sent: 13 January 2015 20:14 To: openstack-hpc@lists.openstack.orgmailto:openstack-hpc@lists.openstack.org Subject: [openstack-hpc] How are you using OpenStack for your HPC workflow? (We're looking at Ironic!)
Hi, everybody!
We at CU-Boulder Research Computing are looking at using OpenStack Ironic for our HPC cluster(s). I’ve got a sysadmin background, so my initial goal is to simply use OpenStack to deploy our traditional queueing system (Slurm), and basically use OpenStack (w/Ironic) as a replacement for Cobbler, xcat, perceus, hand-rolled PXE, or whatever else sysadmin-focused PXE-based node provisioning solution one might use. That said, once we have OpenStack (and have some competency with it), we’d presumably be well positioned to offer actual cloud-y virtualization to our users, or even offer baremetal instances for direct use for some of our (presumably most trusted) users.
But how about the rest of you? Are you using OpenStack in your HPC workloads today? Are you doing “traditional” virtualization? Are you using the nova- compute baremetal driver? Are you using Ironic? Something even more exotic?
And, most importantly, how has it worked for you?
I’ve finally got some nodes to start experimenting with, so I hope to have some real hands-on experience with Ironic soon. (So far, I’ve only been able to do traditional virtualization on some workstations.)
~jonathon
OpenStack-HPC mailing list OpenStack-HPC@lists.openstack.orgmailto:OpenStack-HPC@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
_______________________________________________ OpenStack-HPC mailing list OpenStack-HPC@lists.openstack.orgmailto:OpenStack-HPC@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
On Wed, 14 Jan 2015 06:24:31 +0000 Tim Bell Tim.Bell@cern.ch wrote:
We're starting to have a look at bare metal and containers too. Using OpenStack as an overall resource allocation system to ensure all compute usage is accounted and managed is of great interest. The highlighted items above will remain but hopefully we are work with others in the community to address them (as we are with nested projects)
Take a look at Mesos (mesos.apache.org, mesosphere.com). I did some thinking on openstack+hpc topic but didn't like the conclusions I arrived at so I went shopping for alternatives. Mesos+docker looks like the best one so far.
https://jure.pecar.org/2014-12-22/spirals
Hi Jonathon,
At France Grilles, we are using SlipStream to deploy Torque or Hadoop Cluster on our Cloud instances. It is very convenient for users, and can also be used as Cloud Broker.
This solution works pretty well. We have launched a 'Cloud Challenge' call last year to choose a scientific project that could use all our Cloud Computing resources (thousands of VMs spread over 5 datacenters, not all are running OpenStack). A project has been selected end of December and will start this month. It will use SlipStream to manage the VMs. If people are interested, I can give a feedback once the contest is over.
Cheers,
Jerome Pansanel
Le 13/01/2015 20:13, Jonathon A Anderson a écrit :
Hi, everybody!
We at CU-Boulder Research Computing are looking at using OpenStack Ironic for our HPC cluster(s). I’ve got a sysadmin background, so my initial goal is to simply use OpenStack to deploy our traditional queueing system (Slurm), and basically use OpenStack (w/Ironic) as a replacement for Cobbler, xcat, perceus, hand-rolled PXE, or whatever else sysadmin-focused PXE-based node provisioning solution one might use. That said, once we have OpenStack (and have some competency with it), we’d presumably be well positioned to offer actual cloud-y virtualization to our users, or even offer baremetal instances for direct use for some of our (presumably most trusted) users.
But how about the rest of you? Are you using OpenStack in your HPC workloads today? Are you doing “traditional” virtualization? Are you using the nova-compute baremetal driver? Are you using Ironic? Something even more exotic?
And, most importantly, how has it worked for you?
I’ve finally got some nodes to start experimenting with, so I hope to have some real hands-on experience with Ironic soon. (So far, I’ve only been able to do traditional virtualization on some workstations.)
~jonathon
OpenStack-HPC mailing list OpenStack-HPC@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
Jonathon,
We're looking to utilize star cluster toolbox to provision fully automated grid environments. Star cluster is using SGE as batch queuing system. We gain good results running star cluster on Amazon web services (AWS)
Sent from my iPad
On Jan 14, 2015, at 2:49 AM, Jerome Pansanel jerome.pansanel@iphc.cnrs.fr wrote:
Hi Jonathon,
At France Grilles, we are using SlipStream to deploy Torque or Hadoop Cluster on our Cloud instances. It is very convenient for users, and can also be used as Cloud Broker.
This solution works pretty well. We have launched a 'Cloud Challenge' call last year to choose a scientific project that could use all our Cloud Computing resources (thousands of VMs spread over 5 datacenters, not all are running OpenStack). A project has been selected end of December and will start this month. It will use SlipStream to manage the VMs. If people are interested, I can give a feedback once the contest is over.
Cheers,
Jerome Pansanel
Le 13/01/2015 20:13, Jonathon A Anderson a écrit :
Hi, everybody!
We at CU-Boulder Research Computing are looking at using OpenStack Ironic for our HPC cluster(s). I’ve got a sysadmin background, so my initial goal is to simply use OpenStack to deploy our traditional queueing system (Slurm), and basically use OpenStack (w/Ironic) as a replacement for Cobbler, xcat, perceus, hand-rolled PXE, or whatever else sysadmin-focused PXE-based node provisioning solution one might use. That said, once we have OpenStack (and have some competency with it), we’d presumably be well positioned to offer actual cloud-y virtualization to our users, or even offer baremetal instances for direct use for some of our (presumably most trusted) users.
But how about the rest of you? Are you using OpenStack in your HPC workloads today? Are you doing “traditional” virtualization? Are you using the nova-compute baremetal driver? Are you using Ironic? Something even more exotic?
And, most importantly, how has it worked for you?
I’ve finally got some nodes to start experimenting with, so I hope to have some real hands-on experience with Ironic soon. (So far, I’ve only been able to do traditional virtualization on some workstations.)
~jonathon
OpenStack-HPC mailing list OpenStack-HPC@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
-- Jerome Pansanel Technical Director at France Grilles Grid & Cloud Computing Operations Manager at IPHC IPHC || GSM: +33 (0)6 25 19 24 43 23 rue du Loess, BP 28 || Tel: +33 (0)3 88 10 66 24 F-67037 STRASBOURG Cedex 2 || Fax: +33 (0)3 88 10 62 34
OpenStack-HPC mailing list OpenStack-HPC@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-hpc
participants (7)
-
Andrei Vakhnin
-
Jerome Pansanel
-
John Hearns
-
John-Paul Robinson
-
Jonathon A Anderson
-
Jure Pečar
-
Tim Bell