Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
That is awesome, happy to finally see it enabled! On Mon, Feb 15, 2016 at 9:32 PM, Anastasia Urlapovawrote: > Aleksey, great news! > > On Mon, Feb 15, 2016 at 7:36 PM, Alexey Shtokolov > wrote: > >> Fuelers, >> >> Task based deployment engine has been enabled in master (Fuel 9.0) by >> default [0] >> >> [0] - https://review.openstack.org/#/c/273693/ >> >> WBR, Alexey Shtokolov >> >> 2016-02-09 21:57 GMT+03:00 Vladimir Kuklin : >> >>> Folks >>> >>> It seems that docker removal spoilt our celebration a bit. Here is a bug >>> link https://bugs.launchpad.net/fuel/+bug/1543720 . Fix is trivial, but >>> will postpone swarm run for another day. Nevertheless, it seems to be the >>> only issue affecting our ability to use TBD. >>> >>> Stay tuned! >>> >>> On Tue, Feb 9, 2016 at 2:26 PM, Igor Kalnitsky >>> wrote: >>> > I've run BVT more than 100 times, it works, You run it some time ago. There were a lot of opportunities to introduce regression in both Nailgun and tasks of Fuel Library. ;) > We are going to run a swarm test today against the ISO with enabled task-based deployment So there will be a custom ISO, correct? If so, it works for me and I'll wait for its result. On Tue, Feb 9, 2016 at 1:17 PM, Alexey Shtokolov wrote: > Igor, > > We are going to run a swarm test today against the ISO with enabled > task-based deployment, than check results and merge changes tomorrow. > I've run BVT more than 100 times, it works, but I would like to check more > deployment cases. > And I guess it should be easy to troubleshoot if docker-related and > task-based related changes will be separated by a few days. > > 2016-02-09 13:39 GMT+03:00 Igor Kalnitsky : >> >> Well, I'm going to build a new ISO and run BVT. As soon as they are >> green, I'm going to approve the change. >> >> On Tue, Feb 9, 2016 at 12:32 PM, Bogdan Dobrelya < bdobre...@mirantis.com> >> wrote: >> > On 08.02.2016 17:05, Igor Kalnitsky wrote: >> >> Hey Fuelers, >> >> >> >> When we are going to enable it? I think since HCF is passed for >> >> stable/8.0, it's time to enable task-based deployment for master >> >> branch. >> >> >> >> Opinion? >> > >> > This must be done for the 9.0, IMHO. >> > >> >> >> >> - Igor >> >> >> >> On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya >> >> wrote: >> >>> On 02.02.2016 17:35, Alexey Shtokolov wrote: >> Hi Fuelers! >> >> As you may be aware, since [0] Fuel has implemented a new >> orchestration >> engine [1] >> We switched the deployment paradigm from role-based (aka granular) to >> task-based and now Fuel can deploy all nodes simultaneously using >> cross-node dependencies between deployment tasks. >> >>> >> >>> That is great news! Please do not forget about docs updates as well. >> >>> Those docs are always forgotten like poor orphans... I submitted a >> >>> patch >> >>> [0] to MOS docs, please review and add more details, if possible, for >> >>> plugins impact as well. >> >>> >> >>> [0] https://review.fuel-infra.org/#/c/16509/ >> >>> >> >> This feature is experimental in Fuel 8.0 and will be enabled by >> default >> for Fuel 9.0 >> >> Allow me to show you the results. We made some benchmarks on our bare >> metal lab [2] >> >> Case #1. 3 controllers + 7 computes w/ ceph. >> Task-based deployment takes *~38* minutes vs *~1h15m* for granular >> (*~2* >> times faster) >> Here and below the deployment time is average time for 10 runs >> >> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >> Task-based deployment takes *~41* minutes vs *~1h32m* for granular >> (*~2.24* times faster) >> >> >> >> Also we took measurements for Fuel CI test cases. Standard BVT >> (Master >> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one >> host) >> >> Fuel CI slaves with *4 *cores *~1.1* times faster >> In case of 4 cores for 7 VMs they are fighting for CPU resources and >> it >> marginalizes the gain of task-based deployment >> >> Fuel CI slaves with *6* cores *~1.6* times faster >> >> Fuel CI slaves with *12* cores *~1.7* times faster >> >>> >> >>> These are really outstanding results! >> >>>
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
Aleksey, great news! On Mon, Feb 15, 2016 at 7:36 PM, Alexey Shtokolovwrote: > Fuelers, > > Task based deployment engine has been enabled in master (Fuel 9.0) by > default [0] > > [0] - https://review.openstack.org/#/c/273693/ > > WBR, Alexey Shtokolov > > 2016-02-09 21:57 GMT+03:00 Vladimir Kuklin : > >> Folks >> >> It seems that docker removal spoilt our celebration a bit. Here is a bug >> link https://bugs.launchpad.net/fuel/+bug/1543720 . Fix is trivial, but >> will postpone swarm run for another day. Nevertheless, it seems to be the >> only issue affecting our ability to use TBD. >> >> Stay tuned! >> >> On Tue, Feb 9, 2016 at 2:26 PM, Igor Kalnitsky >> wrote: >> >>> > I've run BVT more than 100 times, it works, >>> >>> You run it some time ago. There were a lot of opportunities to >>> introduce regression in both Nailgun and tasks of Fuel Library. ;) >>> >>> > We are going to run a swarm test today against the ISO with enabled >>> task-based deployment >>> >>> So there will be a custom ISO, correct? If so, it works for me and >>> I'll wait for its result. >>> >>> On Tue, Feb 9, 2016 at 1:17 PM, Alexey Shtokolov >>> wrote: >>> > Igor, >>> > >>> > We are going to run a swarm test today against the ISO with enabled >>> > task-based deployment, than check results and merge changes tomorrow. >>> > I've run BVT more than 100 times, it works, but I would like to check >>> more >>> > deployment cases. >>> > And I guess it should be easy to troubleshoot if docker-related and >>> > task-based related changes will be separated by a few days. >>> > >>> > 2016-02-09 13:39 GMT+03:00 Igor Kalnitsky : >>> >> >>> >> Well, I'm going to build a new ISO and run BVT. As soon as they are >>> >> green, I'm going to approve the change. >>> >> >>> >> On Tue, Feb 9, 2016 at 12:32 PM, Bogdan Dobrelya < >>> bdobre...@mirantis.com> >>> >> wrote: >>> >> > On 08.02.2016 17:05, Igor Kalnitsky wrote: >>> >> >> Hey Fuelers, >>> >> >> >>> >> >> When we are going to enable it? I think since HCF is passed for >>> >> >> stable/8.0, it's time to enable task-based deployment for master >>> >> >> branch. >>> >> >> >>> >> >> Opinion? >>> >> > >>> >> > This must be done for the 9.0, IMHO. >>> >> > >>> >> >> >>> >> >> - Igor >>> >> >> >>> >> >> On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya >>> >> >> wrote: >>> >> >>> On 02.02.2016 17:35, Alexey Shtokolov wrote: >>> >> Hi Fuelers! >>> >> >>> >> As you may be aware, since [0] Fuel has implemented a new >>> >> orchestration >>> >> engine [1] >>> >> We switched the deployment paradigm from role-based (aka >>> granular) to >>> >> task-based and now Fuel can deploy all nodes simultaneously using >>> >> cross-node dependencies between deployment tasks. >>> >> >>> >>> >> >>> That is great news! Please do not forget about docs updates as >>> well. >>> >> >>> Those docs are always forgotten like poor orphans... I submitted a >>> >> >>> patch >>> >> >>> [0] to MOS docs, please review and add more details, if possible, >>> for >>> >> >>> plugins impact as well. >>> >> >>> >>> >> >>> [0] https://review.fuel-infra.org/#/c/16509/ >>> >> >>> >>> >> >>> >> This feature is experimental in Fuel 8.0 and will be enabled by >>> >> default >>> >> for Fuel 9.0 >>> >> >>> >> Allow me to show you the results. We made some benchmarks on our >>> bare >>> >> metal lab [2] >>> >> >>> >> Case #1. 3 controllers + 7 computes w/ ceph. >>> >> Task-based deployment takes *~38* minutes vs *~1h15m* for >>> granular >>> >> (*~2* >>> >> times faster) >>> >> Here and below the deployment time is average time for 10 runs >>> >> >>> >> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >>> >> Task-based deployment takes *~41* minutes vs *~1h32m* for >>> granular >>> >> (*~2.24* times faster) >>> >> >>> >> >>> >> >>> >> Also we took measurements for Fuel CI test cases. Standard BVT >>> >> (Master >>> >> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs >>> on one >>> >> host) >>> >> >>> >> Fuel CI slaves with *4 *cores *~1.1* times faster >>> >> In case of 4 cores for 7 VMs they are fighting for CPU resources >>> and >>> >> it >>> >> marginalizes the gain of task-based deployment >>> >> >>> >> Fuel CI slaves with *6* cores *~1.6* times faster >>> >> >>> >> Fuel CI slaves with *12* cores *~1.7* times faster >>> >> >>> >>> >> >>> These are really outstanding results! >>> >> >>> (tl;dr) >>> >> >>> I believe the next step may be to leverage the "external install >>> & svc >>> >> >>> management" feature (example [1]) of the Liberty release (7.0.0) >>> of >>> >> >>> Puppet-Openstack (PO) modules. So we could use separate concurrent >>> >> >>> cross-depends based tasks
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
Fuelers, Task based deployment engine has been enabled in master (Fuel 9.0) by default [0] [0] - https://review.openstack.org/#/c/273693/ WBR, Alexey Shtokolov 2016-02-09 21:57 GMT+03:00 Vladimir Kuklin: > Folks > > It seems that docker removal spoilt our celebration a bit. Here is a bug > link https://bugs.launchpad.net/fuel/+bug/1543720 . Fix is trivial, but > will postpone swarm run for another day. Nevertheless, it seems to be the > only issue affecting our ability to use TBD. > > Stay tuned! > > On Tue, Feb 9, 2016 at 2:26 PM, Igor Kalnitsky > wrote: > >> > I've run BVT more than 100 times, it works, >> >> You run it some time ago. There were a lot of opportunities to >> introduce regression in both Nailgun and tasks of Fuel Library. ;) >> >> > We are going to run a swarm test today against the ISO with enabled >> task-based deployment >> >> So there will be a custom ISO, correct? If so, it works for me and >> I'll wait for its result. >> >> On Tue, Feb 9, 2016 at 1:17 PM, Alexey Shtokolov >> wrote: >> > Igor, >> > >> > We are going to run a swarm test today against the ISO with enabled >> > task-based deployment, than check results and merge changes tomorrow. >> > I've run BVT more than 100 times, it works, but I would like to check >> more >> > deployment cases. >> > And I guess it should be easy to troubleshoot if docker-related and >> > task-based related changes will be separated by a few days. >> > >> > 2016-02-09 13:39 GMT+03:00 Igor Kalnitsky : >> >> >> >> Well, I'm going to build a new ISO and run BVT. As soon as they are >> >> green, I'm going to approve the change. >> >> >> >> On Tue, Feb 9, 2016 at 12:32 PM, Bogdan Dobrelya < >> bdobre...@mirantis.com> >> >> wrote: >> >> > On 08.02.2016 17:05, Igor Kalnitsky wrote: >> >> >> Hey Fuelers, >> >> >> >> >> >> When we are going to enable it? I think since HCF is passed for >> >> >> stable/8.0, it's time to enable task-based deployment for master >> >> >> branch. >> >> >> >> >> >> Opinion? >> >> > >> >> > This must be done for the 9.0, IMHO. >> >> > >> >> >> >> >> >> - Igor >> >> >> >> >> >> On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya >> >> >> wrote: >> >> >>> On 02.02.2016 17:35, Alexey Shtokolov wrote: >> >> Hi Fuelers! >> >> >> >> As you may be aware, since [0] Fuel has implemented a new >> >> orchestration >> >> engine [1] >> >> We switched the deployment paradigm from role-based (aka >> granular) to >> >> task-based and now Fuel can deploy all nodes simultaneously using >> >> cross-node dependencies between deployment tasks. >> >> >>> >> >> >>> That is great news! Please do not forget about docs updates as >> well. >> >> >>> Those docs are always forgotten like poor orphans... I submitted a >> >> >>> patch >> >> >>> [0] to MOS docs, please review and add more details, if possible, >> for >> >> >>> plugins impact as well. >> >> >>> >> >> >>> [0] https://review.fuel-infra.org/#/c/16509/ >> >> >>> >> >> >> >> This feature is experimental in Fuel 8.0 and will be enabled by >> >> default >> >> for Fuel 9.0 >> >> >> >> Allow me to show you the results. We made some benchmarks on our >> bare >> >> metal lab [2] >> >> >> >> Case #1. 3 controllers + 7 computes w/ ceph. >> >> Task-based deployment takes *~38* minutes vs *~1h15m* for granular >> >> (*~2* >> >> times faster) >> >> Here and below the deployment time is average time for 10 runs >> >> >> >> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >> >> Task-based deployment takes *~41* minutes vs *~1h32m* for granular >> >> (*~2.24* times faster) >> >> >> >> >> >> >> >> Also we took measurements for Fuel CI test cases. Standard BVT >> >> (Master >> >> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on >> one >> >> host) >> >> >> >> Fuel CI slaves with *4 *cores *~1.1* times faster >> >> In case of 4 cores for 7 VMs they are fighting for CPU resources >> and >> >> it >> >> marginalizes the gain of task-based deployment >> >> >> >> Fuel CI slaves with *6* cores *~1.6* times faster >> >> >> >> Fuel CI slaves with *12* cores *~1.7* times faster >> >> >>> >> >> >>> These are really outstanding results! >> >> >>> (tl;dr) >> >> >>> I believe the next step may be to leverage the "external install & >> svc >> >> >>> management" feature (example [1]) of the Liberty release (7.0.0) of >> >> >>> Puppet-Openstack (PO) modules. So we could use separate concurrent >> >> >>> cross-depends based tasks *within a single node* as well, like: >> >> >>> - task: install_all_packages - a singleton task for a node, >> >> >>> - task: [configure_x, for each x] - concurrent for a node, >> >> >>> - task: [manage_service_x, for each x] - some may be concurrent >> for a
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
+1 On Tue, Feb 9, 2016 at 11:51 AM, Evgeniy Lwrote: > +1 > > On Mon, Feb 8, 2016 at 7:58 PM, Vladimir Kozhukalov < > vkozhuka...@mirantis.com> wrote: > >> +1 to enable it ASAP. >> >> It will also affect our deployment tests (~1 hour vs. ~2.5 hours). >> >> Vladimir Kozhukalov >> >> On Mon, Feb 8, 2016 at 7:35 PM, Bulat Gaifullin >> wrote: >> >>> +1. >>> >>> Regards, >>> Bulat Gaifullin >>> Mirantis Inc. >>> >>> >>> >>> > On 08 Feb 2016, at 19:05, Igor Kalnitsky >>> wrote: >>> > >>> > Hey Fuelers, >>> > >>> > When we are going to enable it? I think since HCF is passed for >>> > stable/8.0, it's time to enable task-based deployment for master >>> > branch. >>> > >>> > Opinion? >>> > >>> > - Igor >>> > >>> > On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya < >>> bdobre...@mirantis.com> wrote: >>> >> On 02.02.2016 17:35, Alexey Shtokolov wrote: >>> >>> Hi Fuelers! >>> >>> >>> >>> As you may be aware, since [0] Fuel has implemented a new >>> orchestration >>> >>> engine [1] >>> >>> We switched the deployment paradigm from role-based (aka granular) to >>> >>> task-based and now Fuel can deploy all nodes simultaneously using >>> >>> cross-node dependencies between deployment tasks. >>> >> >>> >> That is great news! Please do not forget about docs updates as well. >>> >> Those docs are always forgotten like poor orphans... I submitted a >>> patch >>> >> [0] to MOS docs, please review and add more details, if possible, for >>> >> plugins impact as well. >>> >> >>> >> [0] https://review.fuel-infra.org/#/c/16509/ >>> >> >>> >>> >>> >>> This feature is experimental in Fuel 8.0 and will be enabled by >>> default >>> >>> for Fuel 9.0 >>> >>> >>> >>> Allow me to show you the results. We made some benchmarks on our bare >>> >>> metal lab [2] >>> >>> >>> >>> Case #1. 3 controllers + 7 computes w/ ceph. >>> >>> Task-based deployment takes *~38* minutes vs *~1h15m* for granular >>> (*~2* >>> >>> times faster) >>> >>> Here and below the deployment time is average time for 10 runs >>> >>> >>> >>> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >>> >>> Task-based deployment takes *~41* minutes vs *~1h32m* for granular >>> >>> (*~2.24* times faster) >>> >>> >>> >>> >>> >>> >>> >>> Also we took measurements for Fuel CI test cases. Standard BVT >>> (Master >>> >>> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on >>> one host) >>> >>> >>> >>> Fuel CI slaves with *4 *cores *~1.1* times faster >>> >>> In case of 4 cores for 7 VMs they are fighting for CPU resources and >>> it >>> >>> marginalizes the gain of task-based deployment >>> >>> >>> >>> Fuel CI slaves with *6* cores *~1.6* times faster >>> >>> >>> >>> Fuel CI slaves with *12* cores *~1.7* times faster >>> >> >>> >> These are really outstanding results! >>> >> (tl;dr) >>> >> I believe the next step may be to leverage the "external install & svc >>> >> management" feature (example [1]) of the Liberty release (7.0.0) of >>> >> Puppet-Openstack (PO) modules. So we could use separate concurrent >>> >> cross-depends based tasks *within a single node* as well, like: >>> >> - task: install_all_packages - a singleton task for a node, >>> >> - task: [configure_x, for each x] - concurrent for a node, >>> >> - task: [manage_service_x, for each x] - some may be concurrent for a >>> >> node, while another shall be serialized. >>> >> >>> >> So, one might use the "--tags" separator for concurrent puppet runs to >>> >> make things go even faster, for example: >>> >> >>> >> # cat test.pp >>> >> notify >>> >> {"A": tag => "a" } >>> >> notify >>> >> {"B": tag => "b" } >>> >> >>> >> # puppet apply test.pp >>> >> Notice: A >>> >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >>> >> Notice: B >>> >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >>> >> >>> >> # puppet apply test.pp --tags a >>> >> Notice: A >>> >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >>> >> >>> >> # puppet apply test.pp --tags a & puppet apply test.pp --tags b >>> >> Notice: B >>> >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >>> >> Notice: A >>> >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >>> >> >>> >> Which is supposed to be faster, although not for this example. >>> >> >>> >> [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp >>> >> >>> >>> >>> >>> You can see additional information and charts in the presentation >>> [3]. >>> >>> >>> >>> [0] >>> >>> - >>> http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html >>> >>> [1] >>> >>> - >>> https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html >>> >>> [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x >>> HP >>> >>> ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) >>> >>> [3] - >>> >>> >>>
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
On 08.02.2016 17:05, Igor Kalnitsky wrote: > Hey Fuelers, > > When we are going to enable it? I think since HCF is passed for > stable/8.0, it's time to enable task-based deployment for master > branch. > > Opinion? This must be done for the 9.0, IMHO. > > - Igor > > On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya> wrote: >> On 02.02.2016 17:35, Alexey Shtokolov wrote: >>> Hi Fuelers! >>> >>> As you may be aware, since [0] Fuel has implemented a new orchestration >>> engine [1] >>> We switched the deployment paradigm from role-based (aka granular) to >>> task-based and now Fuel can deploy all nodes simultaneously using >>> cross-node dependencies between deployment tasks. >> >> That is great news! Please do not forget about docs updates as well. >> Those docs are always forgotten like poor orphans... I submitted a patch >> [0] to MOS docs, please review and add more details, if possible, for >> plugins impact as well. >> >> [0] https://review.fuel-infra.org/#/c/16509/ >> >>> >>> This feature is experimental in Fuel 8.0 and will be enabled by default >>> for Fuel 9.0 >>> >>> Allow me to show you the results. We made some benchmarks on our bare >>> metal lab [2] >>> >>> Case #1. 3 controllers + 7 computes w/ ceph. >>> Task-based deployment takes *~38* minutes vs *~1h15m* for granular (*~2* >>> times faster) >>> Here and below the deployment time is average time for 10 runs >>> >>> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >>> Task-based deployment takes *~41* minutes vs *~1h32m* for granular >>> (*~2.24* times faster) >>> >>> >>> >>> Also we took measurements for Fuel CI test cases. Standard BVT (Master >>> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one host) >>> >>> Fuel CI slaves with *4 *cores *~1.1* times faster >>> In case of 4 cores for 7 VMs they are fighting for CPU resources and it >>> marginalizes the gain of task-based deployment >>> >>> Fuel CI slaves with *6* cores *~1.6* times faster >>> >>> Fuel CI slaves with *12* cores *~1.7* times faster >> >> These are really outstanding results! >> (tl;dr) >> I believe the next step may be to leverage the "external install & svc >> management" feature (example [1]) of the Liberty release (7.0.0) of >> Puppet-Openstack (PO) modules. So we could use separate concurrent >> cross-depends based tasks *within a single node* as well, like: >> - task: install_all_packages - a singleton task for a node, >> - task: [configure_x, for each x] - concurrent for a node, >> - task: [manage_service_x, for each x] - some may be concurrent for a >> node, while another shall be serialized. >> >> So, one might use the "--tags" separator for concurrent puppet runs to >> make things go even faster, for example: >> >> # cat test.pp >> notify >> {"A": tag => "a" } >> notify >> {"B": tag => "b" } >> >> # puppet apply test.pp >> Notice: A >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> Notice: B >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> >> # puppet apply test.pp --tags a >> Notice: A >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >> # puppet apply test.pp --tags a & puppet apply test.pp --tags b >> Notice: B >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> Notice: A >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >> Which is supposed to be faster, although not for this example. >> >> [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp >> >>> >>> You can see additional information and charts in the presentation [3]. >>> >>> [0] >>> - >>> http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html >>> [1] >>> - >>> https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html >>> [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x HP >>> ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) >>> [3] - >>> https://docs.google.com/presentation/d/1jZCFZlXHs_VhjtVYS2VuWgdxge5Q6sOMLz4bRLuw7YE >>> >>> --- >>> WBR, Alexey Shtokolov >>> >>> >>> __ >>> OpenStack Development Mailing List (not for usage questions) >>> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe >>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev >>> >> >> >> -- >> Best regards, >> Bogdan Dobrelya, >> Irc #bogdando >> >> __ >> OpenStack Development Mailing List (not for usage questions) >> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe >> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > > __ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe >
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
+1 On Mon, Feb 8, 2016 at 7:58 PM, Vladimir Kozhukalov < vkozhuka...@mirantis.com> wrote: > +1 to enable it ASAP. > > It will also affect our deployment tests (~1 hour vs. ~2.5 hours). > > Vladimir Kozhukalov > > On Mon, Feb 8, 2016 at 7:35 PM, Bulat Gaifullin> wrote: > >> +1. >> >> Regards, >> Bulat Gaifullin >> Mirantis Inc. >> >> >> >> > On 08 Feb 2016, at 19:05, Igor Kalnitsky >> wrote: >> > >> > Hey Fuelers, >> > >> > When we are going to enable it? I think since HCF is passed for >> > stable/8.0, it's time to enable task-based deployment for master >> > branch. >> > >> > Opinion? >> > >> > - Igor >> > >> > On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya < >> bdobre...@mirantis.com> wrote: >> >> On 02.02.2016 17:35, Alexey Shtokolov wrote: >> >>> Hi Fuelers! >> >>> >> >>> As you may be aware, since [0] Fuel has implemented a new >> orchestration >> >>> engine [1] >> >>> We switched the deployment paradigm from role-based (aka granular) to >> >>> task-based and now Fuel can deploy all nodes simultaneously using >> >>> cross-node dependencies between deployment tasks. >> >> >> >> That is great news! Please do not forget about docs updates as well. >> >> Those docs are always forgotten like poor orphans... I submitted a >> patch >> >> [0] to MOS docs, please review and add more details, if possible, for >> >> plugins impact as well. >> >> >> >> [0] https://review.fuel-infra.org/#/c/16509/ >> >> >> >>> >> >>> This feature is experimental in Fuel 8.0 and will be enabled by >> default >> >>> for Fuel 9.0 >> >>> >> >>> Allow me to show you the results. We made some benchmarks on our bare >> >>> metal lab [2] >> >>> >> >>> Case #1. 3 controllers + 7 computes w/ ceph. >> >>> Task-based deployment takes *~38* minutes vs *~1h15m* for granular >> (*~2* >> >>> times faster) >> >>> Here and below the deployment time is average time for 10 runs >> >>> >> >>> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >> >>> Task-based deployment takes *~41* minutes vs *~1h32m* for granular >> >>> (*~2.24* times faster) >> >>> >> >>> >> >>> >> >>> Also we took measurements for Fuel CI test cases. Standard BVT (Master >> >>> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one >> host) >> >>> >> >>> Fuel CI slaves with *4 *cores *~1.1* times faster >> >>> In case of 4 cores for 7 VMs they are fighting for CPU resources and >> it >> >>> marginalizes the gain of task-based deployment >> >>> >> >>> Fuel CI slaves with *6* cores *~1.6* times faster >> >>> >> >>> Fuel CI slaves with *12* cores *~1.7* times faster >> >> >> >> These are really outstanding results! >> >> (tl;dr) >> >> I believe the next step may be to leverage the "external install & svc >> >> management" feature (example [1]) of the Liberty release (7.0.0) of >> >> Puppet-Openstack (PO) modules. So we could use separate concurrent >> >> cross-depends based tasks *within a single node* as well, like: >> >> - task: install_all_packages - a singleton task for a node, >> >> - task: [configure_x, for each x] - concurrent for a node, >> >> - task: [manage_service_x, for each x] - some may be concurrent for a >> >> node, while another shall be serialized. >> >> >> >> So, one might use the "--tags" separator for concurrent puppet runs to >> >> make things go even faster, for example: >> >> >> >> # cat test.pp >> >> notify >> >> {"A": tag => "a" } >> >> notify >> >> {"B": tag => "b" } >> >> >> >> # puppet apply test.pp >> >> Notice: A >> >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >> Notice: B >> >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> >> >> >> # puppet apply test.pp --tags a >> >> Notice: A >> >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >> >> >> # puppet apply test.pp --tags a & puppet apply test.pp --tags b >> >> Notice: B >> >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> >> Notice: A >> >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >> >> >> Which is supposed to be faster, although not for this example. >> >> >> >> [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp >> >> >> >>> >> >>> You can see additional information and charts in the presentation [3]. >> >>> >> >>> [0] >> >>> - >> http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html >> >>> [1] >> >>> - >> https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html >> >>> [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x HP >> >>> ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) >> >>> [3] - >> >>> >> https://docs.google.com/presentation/d/1jZCFZlXHs_VhjtVYS2VuWgdxge5Q6sOMLz4bRLuw7YE >> >>> >> >>> --- >> >>> WBR, Alexey Shtokolov >> >>> >> >>> >> >>> >> __ >> >>> OpenStack Development Mailing List (not for usage questions) >> >>>
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
+1 On Tue, Feb 9, 2016 at 12:28 PM, Anastasia Urlapovawrote: > +1 > > On Tue, Feb 9, 2016 at 11:51 AM, Evgeniy L wrote: > >> +1 >> >> On Mon, Feb 8, 2016 at 7:58 PM, Vladimir Kozhukalov < >> vkozhuka...@mirantis.com> wrote: >> >>> +1 to enable it ASAP. >>> >>> It will also affect our deployment tests (~1 hour vs. ~2.5 hours). >>> >>> Vladimir Kozhukalov >>> >>> On Mon, Feb 8, 2016 at 7:35 PM, Bulat Gaifullin >> > wrote: >>> +1. Regards, Bulat Gaifullin Mirantis Inc. > On 08 Feb 2016, at 19:05, Igor Kalnitsky wrote: > > Hey Fuelers, > > When we are going to enable it? I think since HCF is passed for > stable/8.0, it's time to enable task-based deployment for master > branch. > > Opinion? > > - Igor > > On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya < bdobre...@mirantis.com> wrote: >> On 02.02.2016 17:35, Alexey Shtokolov wrote: >>> Hi Fuelers! >>> >>> As you may be aware, since [0] Fuel has implemented a new orchestration >>> engine [1] >>> We switched the deployment paradigm from role-based (aka granular) to >>> task-based and now Fuel can deploy all nodes simultaneously using >>> cross-node dependencies between deployment tasks. >> >> That is great news! Please do not forget about docs updates as well. >> Those docs are always forgotten like poor orphans... I submitted a patch >> [0] to MOS docs, please review and add more details, if possible, for >> plugins impact as well. >> >> [0] https://review.fuel-infra.org/#/c/16509/ >> >>> >>> This feature is experimental in Fuel 8.0 and will be enabled by default >>> for Fuel 9.0 >>> >>> Allow me to show you the results. We made some benchmarks on our bare >>> metal lab [2] >>> >>> Case #1. 3 controllers + 7 computes w/ ceph. >>> Task-based deployment takes *~38* minutes vs *~1h15m* for granular (*~2* >>> times faster) >>> Here and below the deployment time is average time for 10 runs >>> >>> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >>> Task-based deployment takes *~41* minutes vs *~1h32m* for granular >>> (*~2.24* times faster) >>> >>> >>> >>> Also we took measurements for Fuel CI test cases. Standard BVT (Master >>> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one host) >>> >>> Fuel CI slaves with *4 *cores *~1.1* times faster >>> In case of 4 cores for 7 VMs they are fighting for CPU resources and it >>> marginalizes the gain of task-based deployment >>> >>> Fuel CI slaves with *6* cores *~1.6* times faster >>> >>> Fuel CI slaves with *12* cores *~1.7* times faster >> >> These are really outstanding results! >> (tl;dr) >> I believe the next step may be to leverage the "external install & svc >> management" feature (example [1]) of the Liberty release (7.0.0) of >> Puppet-Openstack (PO) modules. So we could use separate concurrent >> cross-depends based tasks *within a single node* as well, like: >> - task: install_all_packages - a singleton task for a node, >> - task: [configure_x, for each x] - concurrent for a node, >> - task: [manage_service_x, for each x] - some may be concurrent for a >> node, while another shall be serialized. >> >> So, one might use the "--tags" separator for concurrent puppet runs to >> make things go even faster, for example: >> >> # cat test.pp >> notify >> {"A": tag => "a" } >> notify >> {"B": tag => "b" } >> >> # puppet apply test.pp >> Notice: A >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> Notice: B >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> >> # puppet apply test.pp --tags a >> Notice: A >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >> # puppet apply test.pp --tags a & puppet apply test.pp --tags b >> Notice: B >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> Notice: A >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >> Which is supposed to be faster, although not for this example. >> >> [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp >> >>> >>> You can see additional information and charts in the presentation [3]. >>> >>> [0] >>> - http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html >>> [1] >>> -
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
Well, I'm going to build a new ISO and run BVT. As soon as they are green, I'm going to approve the change. On Tue, Feb 9, 2016 at 12:32 PM, Bogdan Dobrelyawrote: > On 08.02.2016 17:05, Igor Kalnitsky wrote: >> Hey Fuelers, >> >> When we are going to enable it? I think since HCF is passed for >> stable/8.0, it's time to enable task-based deployment for master >> branch. >> >> Opinion? > > This must be done for the 9.0, IMHO. > >> >> - Igor >> >> On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya >> wrote: >>> On 02.02.2016 17:35, Alexey Shtokolov wrote: Hi Fuelers! As you may be aware, since [0] Fuel has implemented a new orchestration engine [1] We switched the deployment paradigm from role-based (aka granular) to task-based and now Fuel can deploy all nodes simultaneously using cross-node dependencies between deployment tasks. >>> >>> That is great news! Please do not forget about docs updates as well. >>> Those docs are always forgotten like poor orphans... I submitted a patch >>> [0] to MOS docs, please review and add more details, if possible, for >>> plugins impact as well. >>> >>> [0] https://review.fuel-infra.org/#/c/16509/ >>> This feature is experimental in Fuel 8.0 and will be enabled by default for Fuel 9.0 Allow me to show you the results. We made some benchmarks on our bare metal lab [2] Case #1. 3 controllers + 7 computes w/ ceph. Task-based deployment takes *~38* minutes vs *~1h15m* for granular (*~2* times faster) Here and below the deployment time is average time for 10 runs Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. Task-based deployment takes *~41* minutes vs *~1h32m* for granular (*~2.24* times faster) Also we took measurements for Fuel CI test cases. Standard BVT (Master node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one host) Fuel CI slaves with *4 *cores *~1.1* times faster In case of 4 cores for 7 VMs they are fighting for CPU resources and it marginalizes the gain of task-based deployment Fuel CI slaves with *6* cores *~1.6* times faster Fuel CI slaves with *12* cores *~1.7* times faster >>> >>> These are really outstanding results! >>> (tl;dr) >>> I believe the next step may be to leverage the "external install & svc >>> management" feature (example [1]) of the Liberty release (7.0.0) of >>> Puppet-Openstack (PO) modules. So we could use separate concurrent >>> cross-depends based tasks *within a single node* as well, like: >>> - task: install_all_packages - a singleton task for a node, >>> - task: [configure_x, for each x] - concurrent for a node, >>> - task: [manage_service_x, for each x] - some may be concurrent for a >>> node, while another shall be serialized. >>> >>> So, one might use the "--tags" separator for concurrent puppet runs to >>> make things go even faster, for example: >>> >>> # cat test.pp >>> notify >>> {"A": tag => "a" } >>> notify >>> {"B": tag => "b" } >>> >>> # puppet apply test.pp >>> Notice: A >>> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >>> Notice: B >>> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >>> >>> # puppet apply test.pp --tags a >>> Notice: A >>> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >>> >>> # puppet apply test.pp --tags a & puppet apply test.pp --tags b >>> Notice: B >>> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >>> Notice: A >>> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >>> >>> Which is supposed to be faster, although not for this example. >>> >>> [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp >>> You can see additional information and charts in the presentation [3]. [0] - http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html [1] - https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x HP ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) [3] - https://docs.google.com/presentation/d/1jZCFZlXHs_VhjtVYS2VuWgdxge5Q6sOMLz4bRLuw7YE --- WBR, Alexey Shtokolov __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev >>> >>> >>> -- >>> Best regards, >>> Bogdan Dobrelya, >>> Irc #bogdando >>> >>> __ >>> OpenStack Development Mailing List (not for usage questions) >>> Unsubscribe:
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
Igor, We are going to run a swarm test today against the ISO with enabled task-based deployment, than check results and merge changes tomorrow. I've run BVT more than 100 times, it works, but I would like to check more deployment cases. And I guess it should be easy to troubleshoot if docker-related and task-based related changes will be separated by a few days. 2016-02-09 13:39 GMT+03:00 Igor Kalnitsky: > Well, I'm going to build a new ISO and run BVT. As soon as they are > green, I'm going to approve the change. > > On Tue, Feb 9, 2016 at 12:32 PM, Bogdan Dobrelya > wrote: > > On 08.02.2016 17:05, Igor Kalnitsky wrote: > >> Hey Fuelers, > >> > >> When we are going to enable it? I think since HCF is passed for > >> stable/8.0, it's time to enable task-based deployment for master > >> branch. > >> > >> Opinion? > > > > This must be done for the 9.0, IMHO. > > > >> > >> - Igor > >> > >> On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya < > bdobre...@mirantis.com> wrote: > >>> On 02.02.2016 17:35, Alexey Shtokolov wrote: > Hi Fuelers! > > As you may be aware, since [0] Fuel has implemented a new > orchestration > engine [1] > We switched the deployment paradigm from role-based (aka granular) to > task-based and now Fuel can deploy all nodes simultaneously using > cross-node dependencies between deployment tasks. > >>> > >>> That is great news! Please do not forget about docs updates as well. > >>> Those docs are always forgotten like poor orphans... I submitted a > patch > >>> [0] to MOS docs, please review and add more details, if possible, for > >>> plugins impact as well. > >>> > >>> [0] https://review.fuel-infra.org/#/c/16509/ > >>> > > This feature is experimental in Fuel 8.0 and will be enabled by > default > for Fuel 9.0 > > Allow me to show you the results. We made some benchmarks on our bare > metal lab [2] > > Case #1. 3 controllers + 7 computes w/ ceph. > Task-based deployment takes *~38* minutes vs *~1h15m* for granular > (*~2* > times faster) > Here and below the deployment time is average time for 10 runs > > Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. > Task-based deployment takes *~41* minutes vs *~1h32m* for granular > (*~2.24* times faster) > > > > Also we took measurements for Fuel CI test cases. Standard BVT (Master > node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one > host) > > Fuel CI slaves with *4 *cores *~1.1* times faster > In case of 4 cores for 7 VMs they are fighting for CPU resources and > it > marginalizes the gain of task-based deployment > > Fuel CI slaves with *6* cores *~1.6* times faster > > Fuel CI slaves with *12* cores *~1.7* times faster > >>> > >>> These are really outstanding results! > >>> (tl;dr) > >>> I believe the next step may be to leverage the "external install & svc > >>> management" feature (example [1]) of the Liberty release (7.0.0) of > >>> Puppet-Openstack (PO) modules. So we could use separate concurrent > >>> cross-depends based tasks *within a single node* as well, like: > >>> - task: install_all_packages - a singleton task for a node, > >>> - task: [configure_x, for each x] - concurrent for a node, > >>> - task: [manage_service_x, for each x] - some may be concurrent for a > >>> node, while another shall be serialized. > >>> > >>> So, one might use the "--tags" separator for concurrent puppet runs to > >>> make things go even faster, for example: > >>> > >>> # cat test.pp > >>> notify > >>> {"A": tag => "a" } > >>> notify > >>> {"B": tag => "b" } > >>> > >>> # puppet apply test.pp > >>> Notice: A > >>> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' > >>> Notice: B > >>> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' > >>> > >>> # puppet apply test.pp --tags a > >>> Notice: A > >>> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' > >>> > >>> # puppet apply test.pp --tags a & puppet apply test.pp --tags b > >>> Notice: B > >>> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' > >>> Notice: A > >>> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' > >>> > >>> Which is supposed to be faster, although not for this example. > >>> > >>> [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp > >>> > > You can see additional information and charts in the presentation [3]. > > [0] > - > http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html > [1] > - > https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html > [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x HP > ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) > [3] - > >
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
> I've run BVT more than 100 times, it works, You run it some time ago. There were a lot of opportunities to introduce regression in both Nailgun and tasks of Fuel Library. ;) > We are going to run a swarm test today against the ISO with enabled > task-based deployment So there will be a custom ISO, correct? If so, it works for me and I'll wait for its result. On Tue, Feb 9, 2016 at 1:17 PM, Alexey Shtokolovwrote: > Igor, > > We are going to run a swarm test today against the ISO with enabled > task-based deployment, than check results and merge changes tomorrow. > I've run BVT more than 100 times, it works, but I would like to check more > deployment cases. > And I guess it should be easy to troubleshoot if docker-related and > task-based related changes will be separated by a few days. > > 2016-02-09 13:39 GMT+03:00 Igor Kalnitsky : >> >> Well, I'm going to build a new ISO and run BVT. As soon as they are >> green, I'm going to approve the change. >> >> On Tue, Feb 9, 2016 at 12:32 PM, Bogdan Dobrelya >> wrote: >> > On 08.02.2016 17:05, Igor Kalnitsky wrote: >> >> Hey Fuelers, >> >> >> >> When we are going to enable it? I think since HCF is passed for >> >> stable/8.0, it's time to enable task-based deployment for master >> >> branch. >> >> >> >> Opinion? >> > >> > This must be done for the 9.0, IMHO. >> > >> >> >> >> - Igor >> >> >> >> On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya >> >> wrote: >> >>> On 02.02.2016 17:35, Alexey Shtokolov wrote: >> Hi Fuelers! >> >> As you may be aware, since [0] Fuel has implemented a new >> orchestration >> engine [1] >> We switched the deployment paradigm from role-based (aka granular) to >> task-based and now Fuel can deploy all nodes simultaneously using >> cross-node dependencies between deployment tasks. >> >>> >> >>> That is great news! Please do not forget about docs updates as well. >> >>> Those docs are always forgotten like poor orphans... I submitted a >> >>> patch >> >>> [0] to MOS docs, please review and add more details, if possible, for >> >>> plugins impact as well. >> >>> >> >>> [0] https://review.fuel-infra.org/#/c/16509/ >> >>> >> >> This feature is experimental in Fuel 8.0 and will be enabled by >> default >> for Fuel 9.0 >> >> Allow me to show you the results. We made some benchmarks on our bare >> metal lab [2] >> >> Case #1. 3 controllers + 7 computes w/ ceph. >> Task-based deployment takes *~38* minutes vs *~1h15m* for granular >> (*~2* >> times faster) >> Here and below the deployment time is average time for 10 runs >> >> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >> Task-based deployment takes *~41* minutes vs *~1h32m* for granular >> (*~2.24* times faster) >> >> >> >> Also we took measurements for Fuel CI test cases. Standard BVT >> (Master >> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one >> host) >> >> Fuel CI slaves with *4 *cores *~1.1* times faster >> In case of 4 cores for 7 VMs they are fighting for CPU resources and >> it >> marginalizes the gain of task-based deployment >> >> Fuel CI slaves with *6* cores *~1.6* times faster >> >> Fuel CI slaves with *12* cores *~1.7* times faster >> >>> >> >>> These are really outstanding results! >> >>> (tl;dr) >> >>> I believe the next step may be to leverage the "external install & svc >> >>> management" feature (example [1]) of the Liberty release (7.0.0) of >> >>> Puppet-Openstack (PO) modules. So we could use separate concurrent >> >>> cross-depends based tasks *within a single node* as well, like: >> >>> - task: install_all_packages - a singleton task for a node, >> >>> - task: [configure_x, for each x] - concurrent for a node, >> >>> - task: [manage_service_x, for each x] - some may be concurrent for a >> >>> node, while another shall be serialized. >> >>> >> >>> So, one might use the "--tags" separator for concurrent puppet runs to >> >>> make things go even faster, for example: >> >>> >> >>> # cat test.pp >> >>> notify >> >>> {"A": tag => "a" } >> >>> notify >> >>> {"B": tag => "b" } >> >>> >> >>> # puppet apply test.pp >> >>> Notice: A >> >>> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >>> Notice: B >> >>> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> >>> >> >>> # puppet apply test.pp --tags a >> >>> Notice: A >> >>> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >>> >> >>> # puppet apply test.pp --tags a & puppet apply test.pp --tags b >> >>> Notice: B >> >>> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> >>> Notice: A >> >>> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >>> >> >>> Which is supposed to be
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
Folks It seems that docker removal spoilt our celebration a bit. Here is a bug link https://bugs.launchpad.net/fuel/+bug/1543720 . Fix is trivial, but will postpone swarm run for another day. Nevertheless, it seems to be the only issue affecting our ability to use TBD. Stay tuned! On Tue, Feb 9, 2016 at 2:26 PM, Igor Kalnitskywrote: > > I've run BVT more than 100 times, it works, > > You run it some time ago. There were a lot of opportunities to > introduce regression in both Nailgun and tasks of Fuel Library. ;) > > > We are going to run a swarm test today against the ISO with enabled > task-based deployment > > So there will be a custom ISO, correct? If so, it works for me and > I'll wait for its result. > > On Tue, Feb 9, 2016 at 1:17 PM, Alexey Shtokolov > wrote: > > Igor, > > > > We are going to run a swarm test today against the ISO with enabled > > task-based deployment, than check results and merge changes tomorrow. > > I've run BVT more than 100 times, it works, but I would like to check > more > > deployment cases. > > And I guess it should be easy to troubleshoot if docker-related and > > task-based related changes will be separated by a few days. > > > > 2016-02-09 13:39 GMT+03:00 Igor Kalnitsky : > >> > >> Well, I'm going to build a new ISO and run BVT. As soon as they are > >> green, I'm going to approve the change. > >> > >> On Tue, Feb 9, 2016 at 12:32 PM, Bogdan Dobrelya < > bdobre...@mirantis.com> > >> wrote: > >> > On 08.02.2016 17:05, Igor Kalnitsky wrote: > >> >> Hey Fuelers, > >> >> > >> >> When we are going to enable it? I think since HCF is passed for > >> >> stable/8.0, it's time to enable task-based deployment for master > >> >> branch. > >> >> > >> >> Opinion? > >> > > >> > This must be done for the 9.0, IMHO. > >> > > >> >> > >> >> - Igor > >> >> > >> >> On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya > >> >> wrote: > >> >>> On 02.02.2016 17:35, Alexey Shtokolov wrote: > >> Hi Fuelers! > >> > >> As you may be aware, since [0] Fuel has implemented a new > >> orchestration > >> engine [1] > >> We switched the deployment paradigm from role-based (aka granular) > to > >> task-based and now Fuel can deploy all nodes simultaneously using > >> cross-node dependencies between deployment tasks. > >> >>> > >> >>> That is great news! Please do not forget about docs updates as well. > >> >>> Those docs are always forgotten like poor orphans... I submitted a > >> >>> patch > >> >>> [0] to MOS docs, please review and add more details, if possible, > for > >> >>> plugins impact as well. > >> >>> > >> >>> [0] https://review.fuel-infra.org/#/c/16509/ > >> >>> > >> > >> This feature is experimental in Fuel 8.0 and will be enabled by > >> default > >> for Fuel 9.0 > >> > >> Allow me to show you the results. We made some benchmarks on our > bare > >> metal lab [2] > >> > >> Case #1. 3 controllers + 7 computes w/ ceph. > >> Task-based deployment takes *~38* minutes vs *~1h15m* for granular > >> (*~2* > >> times faster) > >> Here and below the deployment time is average time for 10 runs > >> > >> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. > >> Task-based deployment takes *~41* minutes vs *~1h32m* for granular > >> (*~2.24* times faster) > >> > >> > >> > >> Also we took measurements for Fuel CI test cases. Standard BVT > >> (Master > >> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on > one > >> host) > >> > >> Fuel CI slaves with *4 *cores *~1.1* times faster > >> In case of 4 cores for 7 VMs they are fighting for CPU resources > and > >> it > >> marginalizes the gain of task-based deployment > >> > >> Fuel CI slaves with *6* cores *~1.6* times faster > >> > >> Fuel CI slaves with *12* cores *~1.7* times faster > >> >>> > >> >>> These are really outstanding results! > >> >>> (tl;dr) > >> >>> I believe the next step may be to leverage the "external install & > svc > >> >>> management" feature (example [1]) of the Liberty release (7.0.0) of > >> >>> Puppet-Openstack (PO) modules. So we could use separate concurrent > >> >>> cross-depends based tasks *within a single node* as well, like: > >> >>> - task: install_all_packages - a singleton task for a node, > >> >>> - task: [configure_x, for each x] - concurrent for a node, > >> >>> - task: [manage_service_x, for each x] - some may be concurrent for > a > >> >>> node, while another shall be serialized. > >> >>> > >> >>> So, one might use the "--tags" separator for concurrent puppet runs > to > >> >>> make things go even faster, for example: > >> >>> > >> >>> # cat test.pp > >> >>> notify > >> >>> {"A": tag => "a" } > >> >>> notify > >> >>> {"B": tag => "b" } > >> >>> > >> >>> # puppet apply test.pp > >> >>> Notice: A
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
Hey Fuelers, When we are going to enable it? I think since HCF is passed for stable/8.0, it's time to enable task-based deployment for master branch. Opinion? - Igor On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelyawrote: > On 02.02.2016 17:35, Alexey Shtokolov wrote: >> Hi Fuelers! >> >> As you may be aware, since [0] Fuel has implemented a new orchestration >> engine [1] >> We switched the deployment paradigm from role-based (aka granular) to >> task-based and now Fuel can deploy all nodes simultaneously using >> cross-node dependencies between deployment tasks. > > That is great news! Please do not forget about docs updates as well. > Those docs are always forgotten like poor orphans... I submitted a patch > [0] to MOS docs, please review and add more details, if possible, for > plugins impact as well. > > [0] https://review.fuel-infra.org/#/c/16509/ > >> >> This feature is experimental in Fuel 8.0 and will be enabled by default >> for Fuel 9.0 >> >> Allow me to show you the results. We made some benchmarks on our bare >> metal lab [2] >> >> Case #1. 3 controllers + 7 computes w/ ceph. >> Task-based deployment takes *~38* minutes vs *~1h15m* for granular (*~2* >> times faster) >> Here and below the deployment time is average time for 10 runs >> >> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >> Task-based deployment takes *~41* minutes vs *~1h32m* for granular >> (*~2.24* times faster) >> >> >> >> Also we took measurements for Fuel CI test cases. Standard BVT (Master >> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one host) >> >> Fuel CI slaves with *4 *cores *~1.1* times faster >> In case of 4 cores for 7 VMs they are fighting for CPU resources and it >> marginalizes the gain of task-based deployment >> >> Fuel CI slaves with *6* cores *~1.6* times faster >> >> Fuel CI slaves with *12* cores *~1.7* times faster > > These are really outstanding results! > (tl;dr) > I believe the next step may be to leverage the "external install & svc > management" feature (example [1]) of the Liberty release (7.0.0) of > Puppet-Openstack (PO) modules. So we could use separate concurrent > cross-depends based tasks *within a single node* as well, like: > - task: install_all_packages - a singleton task for a node, > - task: [configure_x, for each x] - concurrent for a node, > - task: [manage_service_x, for each x] - some may be concurrent for a > node, while another shall be serialized. > > So, one might use the "--tags" separator for concurrent puppet runs to > make things go even faster, for example: > > # cat test.pp > notify > {"A": tag => "a" } > notify > {"B": tag => "b" } > > # puppet apply test.pp > Notice: A > Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' > Notice: B > Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' > > # puppet apply test.pp --tags a > Notice: A > Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' > > # puppet apply test.pp --tags a & puppet apply test.pp --tags b > Notice: B > Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' > Notice: A > Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' > > Which is supposed to be faster, although not for this example. > > [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp > >> >> You can see additional information and charts in the presentation [3]. >> >> [0] >> - >> http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html >> [1] >> - >> https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html >> [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x HP >> ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) >> [3] - >> https://docs.google.com/presentation/d/1jZCFZlXHs_VhjtVYS2VuWgdxge5Q6sOMLz4bRLuw7YE >> >> --- >> WBR, Alexey Shtokolov >> >> >> __ >> OpenStack Development Mailing List (not for usage questions) >> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe >> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev >> > > > -- > Best regards, > Bogdan Dobrelya, > Irc #bogdando > > __ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
+1. Regards, Bulat Gaifullin Mirantis Inc. > On 08 Feb 2016, at 19:05, Igor Kalnitskywrote: > > Hey Fuelers, > > When we are going to enable it? I think since HCF is passed for > stable/8.0, it's time to enable task-based deployment for master > branch. > > Opinion? > > - Igor > > On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya > wrote: >> On 02.02.2016 17:35, Alexey Shtokolov wrote: >>> Hi Fuelers! >>> >>> As you may be aware, since [0] Fuel has implemented a new orchestration >>> engine [1] >>> We switched the deployment paradigm from role-based (aka granular) to >>> task-based and now Fuel can deploy all nodes simultaneously using >>> cross-node dependencies between deployment tasks. >> >> That is great news! Please do not forget about docs updates as well. >> Those docs are always forgotten like poor orphans... I submitted a patch >> [0] to MOS docs, please review and add more details, if possible, for >> plugins impact as well. >> >> [0] https://review.fuel-infra.org/#/c/16509/ >> >>> >>> This feature is experimental in Fuel 8.0 and will be enabled by default >>> for Fuel 9.0 >>> >>> Allow me to show you the results. We made some benchmarks on our bare >>> metal lab [2] >>> >>> Case #1. 3 controllers + 7 computes w/ ceph. >>> Task-based deployment takes *~38* minutes vs *~1h15m* for granular (*~2* >>> times faster) >>> Here and below the deployment time is average time for 10 runs >>> >>> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. >>> Task-based deployment takes *~41* minutes vs *~1h32m* for granular >>> (*~2.24* times faster) >>> >>> >>> >>> Also we took measurements for Fuel CI test cases. Standard BVT (Master >>> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one host) >>> >>> Fuel CI slaves with *4 *cores *~1.1* times faster >>> In case of 4 cores for 7 VMs they are fighting for CPU resources and it >>> marginalizes the gain of task-based deployment >>> >>> Fuel CI slaves with *6* cores *~1.6* times faster >>> >>> Fuel CI slaves with *12* cores *~1.7* times faster >> >> These are really outstanding results! >> (tl;dr) >> I believe the next step may be to leverage the "external install & svc >> management" feature (example [1]) of the Liberty release (7.0.0) of >> Puppet-Openstack (PO) modules. So we could use separate concurrent >> cross-depends based tasks *within a single node* as well, like: >> - task: install_all_packages - a singleton task for a node, >> - task: [configure_x, for each x] - concurrent for a node, >> - task: [manage_service_x, for each x] - some may be concurrent for a >> node, while another shall be serialized. >> >> So, one might use the "--tags" separator for concurrent puppet runs to >> make things go even faster, for example: >> >> # cat test.pp >> notify >> {"A": tag => "a" } >> notify >> {"B": tag => "b" } >> >> # puppet apply test.pp >> Notice: A >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> Notice: B >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> >> # puppet apply test.pp --tags a >> Notice: A >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >> # puppet apply test.pp --tags a & puppet apply test.pp --tags b >> Notice: B >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' >> Notice: A >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' >> >> Which is supposed to be faster, although not for this example. >> >> [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp >> >>> >>> You can see additional information and charts in the presentation [3]. >>> >>> [0] >>> - >>> http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html >>> [1] >>> - >>> https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html >>> [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x HP >>> ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) >>> [3] - >>> https://docs.google.com/presentation/d/1jZCFZlXHs_VhjtVYS2VuWgdxge5Q6sOMLz4bRLuw7YE >>> >>> --- >>> WBR, Alexey Shtokolov >>> >>> >>> __ >>> OpenStack Development Mailing List (not for usage questions) >>> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe >>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev >>> >> >> >> -- >> Best regards, >> Bogdan Dobrelya, >> Irc #bogdando >> >> __ >> OpenStack Development Mailing List (not for usage questions) >> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe >> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > > __ > OpenStack Development Mailing List (not for usage questions) >
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
+1 to enable it ASAP. It will also affect our deployment tests (~1 hour vs. ~2.5 hours). Vladimir Kozhukalov On Mon, Feb 8, 2016 at 7:35 PM, Bulat Gaifullinwrote: > +1. > > Regards, > Bulat Gaifullin > Mirantis Inc. > > > > > On 08 Feb 2016, at 19:05, Igor Kalnitsky > wrote: > > > > Hey Fuelers, > > > > When we are going to enable it? I think since HCF is passed for > > stable/8.0, it's time to enable task-based deployment for master > > branch. > > > > Opinion? > > > > - Igor > > > > On Wed, Feb 3, 2016 at 12:31 PM, Bogdan Dobrelya > wrote: > >> On 02.02.2016 17:35, Alexey Shtokolov wrote: > >>> Hi Fuelers! > >>> > >>> As you may be aware, since [0] Fuel has implemented a new orchestration > >>> engine [1] > >>> We switched the deployment paradigm from role-based (aka granular) to > >>> task-based and now Fuel can deploy all nodes simultaneously using > >>> cross-node dependencies between deployment tasks. > >> > >> That is great news! Please do not forget about docs updates as well. > >> Those docs are always forgotten like poor orphans... I submitted a patch > >> [0] to MOS docs, please review and add more details, if possible, for > >> plugins impact as well. > >> > >> [0] https://review.fuel-infra.org/#/c/16509/ > >> > >>> > >>> This feature is experimental in Fuel 8.0 and will be enabled by default > >>> for Fuel 9.0 > >>> > >>> Allow me to show you the results. We made some benchmarks on our bare > >>> metal lab [2] > >>> > >>> Case #1. 3 controllers + 7 computes w/ ceph. > >>> Task-based deployment takes *~38* minutes vs *~1h15m* for granular > (*~2* > >>> times faster) > >>> Here and below the deployment time is average time for 10 runs > >>> > >>> Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. > >>> Task-based deployment takes *~41* minutes vs *~1h32m* for granular > >>> (*~2.24* times faster) > >>> > >>> > >>> > >>> Also we took measurements for Fuel CI test cases. Standard BVT (Master > >>> node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one > host) > >>> > >>> Fuel CI slaves with *4 *cores *~1.1* times faster > >>> In case of 4 cores for 7 VMs they are fighting for CPU resources and it > >>> marginalizes the gain of task-based deployment > >>> > >>> Fuel CI slaves with *6* cores *~1.6* times faster > >>> > >>> Fuel CI slaves with *12* cores *~1.7* times faster > >> > >> These are really outstanding results! > >> (tl;dr) > >> I believe the next step may be to leverage the "external install & svc > >> management" feature (example [1]) of the Liberty release (7.0.0) of > >> Puppet-Openstack (PO) modules. So we could use separate concurrent > >> cross-depends based tasks *within a single node* as well, like: > >> - task: install_all_packages - a singleton task for a node, > >> - task: [configure_x, for each x] - concurrent for a node, > >> - task: [manage_service_x, for each x] - some may be concurrent for a > >> node, while another shall be serialized. > >> > >> So, one might use the "--tags" separator for concurrent puppet runs to > >> make things go even faster, for example: > >> > >> # cat test.pp > >> notify > >> {"A": tag => "a" } > >> notify > >> {"B": tag => "b" } > >> > >> # puppet apply test.pp > >> Notice: A > >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' > >> Notice: B > >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' > >> > >> # puppet apply test.pp --tags a > >> Notice: A > >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' > >> > >> # puppet apply test.pp --tags a & puppet apply test.pp --tags b > >> Notice: B > >> Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' > >> Notice: A > >> Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' > >> > >> Which is supposed to be faster, although not for this example. > >> > >> [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp > >> > >>> > >>> You can see additional information and charts in the presentation [3]. > >>> > >>> [0] > >>> - > http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html > >>> [1] > >>> - > https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html > >>> [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x HP > >>> ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) > >>> [3] - > >>> > https://docs.google.com/presentation/d/1jZCFZlXHs_VhjtVYS2VuWgdxge5Q6sOMLz4bRLuw7YE > >>> > >>> --- > >>> WBR, Alexey Shtokolov > >>> > >>> > >>> > __ > >>> OpenStack Development Mailing List (not for usage questions) > >>> Unsubscribe: > openstack-dev-requ...@lists.openstack.org?subject:unsubscribe > >>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > >>> > >> > >> > >> -- > >> Best regards, > >> Bogdan Dobrelya, > >> Irc #bogdando > >> > >> >
Re: [openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
On 02.02.2016 17:35, Alexey Shtokolov wrote: > Hi Fuelers! > > As you may be aware, since [0] Fuel has implemented a new orchestration > engine [1] > We switched the deployment paradigm from role-based (aka granular) to > task-based and now Fuel can deploy all nodes simultaneously using > cross-node dependencies between deployment tasks. That is great news! Please do not forget about docs updates as well. Those docs are always forgotten like poor orphans... I submitted a patch [0] to MOS docs, please review and add more details, if possible, for plugins impact as well. [0] https://review.fuel-infra.org/#/c/16509/ > > This feature is experimental in Fuel 8.0 and will be enabled by default > for Fuel 9.0 > > Allow me to show you the results. We made some benchmarks on our bare > metal lab [2] > > Case #1. 3 controllers + 7 computes w/ ceph. > Task-based deployment takes *~38* minutes vs *~1h15m* for granular (*~2* > times faster) > Here and below the deployment time is average time for 10 runs > > Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. > Task-based deployment takes *~41* minutes vs *~1h32m* for granular > (*~2.24* times faster) > > > > Also we took measurements for Fuel CI test cases. Standard BVT (Master > node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one host) > > Fuel CI slaves with *4 *cores *~1.1* times faster > In case of 4 cores for 7 VMs they are fighting for CPU resources and it > marginalizes the gain of task-based deployment > > Fuel CI slaves with *6* cores *~1.6* times faster > > Fuel CI slaves with *12* cores *~1.7* times faster These are really outstanding results! (tl;dr) I believe the next step may be to leverage the "external install & svc management" feature (example [1]) of the Liberty release (7.0.0) of Puppet-Openstack (PO) modules. So we could use separate concurrent cross-depends based tasks *within a single node* as well, like: - task: install_all_packages - a singleton task for a node, - task: [configure_x, for each x] - concurrent for a node, - task: [manage_service_x, for each x] - some may be concurrent for a node, while another shall be serialized. So, one might use the "--tags" separator for concurrent puppet runs to make things go even faster, for example: # cat test.pp notify {"A": tag => "a" } notify {"B": tag => "b" } # puppet apply test.pp Notice: A Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' Notice: B Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' # puppet apply test.pp --tags a Notice: A Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' # puppet apply test.pp --tags a & puppet apply test.pp --tags b Notice: B Notice: /Stage[main]/Main/Notify[B]/message: defined 'message' as 'B' Notice: A Notice: /Stage[main]/Main/Notify[A]/message: defined 'message' as 'A' Which is supposed to be faster, although not for this example. [1] https://review.openstack.org/#/c/216926/3/manifests/init.pp > > You can see additional information and charts in the presentation [3]. > > [0] > - http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html > [1] > - > https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html > [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x HP > ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) > [3] - > https://docs.google.com/presentation/d/1jZCFZlXHs_VhjtVYS2VuWgdxge5Q6sOMLz4bRLuw7YE > > --- > WBR, Alexey Shtokolov > > > __ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > -- Best regards, Bogdan Dobrelya, Irc #bogdando __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [Fuel] Task Based Deployment Is at Least Twice Faster
Hi Fuelers! As you may be aware, since [0] Fuel has implemented a new orchestration engine [1] We switched the deployment paradigm from role-based (aka granular) to task-based and now Fuel can deploy all nodes simultaneously using cross-node dependencies between deployment tasks. This feature is experimental in Fuel 8.0 and will be enabled by default for Fuel 9.0 Allow me to show you the results. We made some benchmarks on our bare metal lab [2] Case #1. 3 controllers + 7 computes w/ ceph. Task-based deployment takes *~38* minutes vs *~1h15m* for granular (*~2* times faster) Here and below the deployment time is average time for 10 runs Case #2. 3 controllers + 3 mongodb + 4 computes w/ ceph. Task-based deployment takes *~41* minutes vs *~1h32m* for granular (*~2.24* times faster) Also we took measurements for Fuel CI test cases. Standard BVT (Master node + 3 controllers + 3 computes w/ ceph. All are in qemu VMs on one host) Fuel CI slaves with *4 *cores *~1.1* times faster In case of 4 cores for 7 VMs they are fighting for CPU resources and it marginalizes the gain of task-based deployment Fuel CI slaves with *6* cores *~1.6* times faster Fuel CI slaves with *12* cores *~1.7* times faster You can see additional information and charts in the presentation [3]. [0] - http://lists.openstack.org/pipermail/openstack-dev/2015-December/082093.html [1] - https://specs.openstack.org/openstack/fuel-specs/specs/8.0/task-based-deployment-mvp.html [2] - 3 x HP ProLiant DL360p Gen8 (XeonE5 6 cores/64GB/SSD) + 7 x HP ProLiant DL320p Gen8 (XeonE3 4 cores/8-16GB/HDD) [3] - https://docs.google.com/presentation/d/1jZCFZlXHs_VhjtVYS2VuWgdxge5Q6sOMLz4bRLuw7YE --- WBR, Alexey Shtokolov __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev