Re: [openstack-dev] [Nova] Frustrations with review wait times
Sure, I can see both ways. It's not easy to find a perfect solution, especially in opensource with such a diverse community. How do other projects handle this? I would think the kernel would have a similar issue, or hadoop or other diverse and large opensource projects. Sent from my really tiny device... On Aug 27, 2013, at 9:46 PM, Mike Spreitzer mspre...@us.ibm.commailto:mspre...@us.ibm.com wrote: Joshua, I do not think such a strict and coarse scheduling is a practical way to manage developers, who have highly individualized talents, backgrounds, and interests. Regards, Mike ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.orgmailto:OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] New Nova hypervisor: Docker
On Wed, Aug 28, 2013 at 4:18 AM, Sam Alba sam.a...@gmail.com wrote: Hi all, We've been working hard during the last couple of weeks with some people. Brian Waldon helped a lot designing the Glance integration and driver testing. Dean Troyer helped a lot on bringing Docker support in Devstack[1]. On top of that, we got several feedback on the Nova code review which definitely helped to improve the code. The blueprint[2] explains what Docker brings to Nova and how to use it. I have to say that this blueprint is a fantastic example of how we should be writing design documents. It addressed almost all of my questions about the integration. However, it would be nice for this to be on Launchpad, that being where we track blueprints. Would it be possible for you to move it over there? Or just link to the design doc from there? Thanks, Michael -- Rackspace Australia ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [TRIPLEO] Derekh for tripleo core
So - calling for votes for Derek to become a TripleO core reviewer! +1 ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On Wed, Aug 28, 2013 at 12:45:26PM +1000, Michael Still wrote: [Concerns over review wait times in the nova project] I think that we're also seeing the fact that nova-core's are also developers. nova-core members have the same feature freeze deadline, and that means that to a certain extent we need to stop reviewing in order to get our own code ready by the deadline. The strength of nova-core is that its members are active developers, so I think a reviewer caste would be a mistake. I am also not saying that nova-core should get different deadlines (although more leniency with exceptions would be nice). Agreed, I think it is very important for the core reviewers to also be active developers, since working on the code is how you gain the knowledge required to do high quality reviews. So, I think lower review rates around deadlines are just a fact of life. This is a fairly common problem across all open source projects really. People consistently wait until just before review deadlines to submit their code. You have to actively encourage people to submit their code well before deadlines / discourage them from waiting till the last minute. Sometimes the best way to get people to learn this is the hard way, by postponing their feature if submitted too close to the dealine and too much other stuff is ahead of it in the queue. IOW we should prioritize review of work whose authors submitted earlier to encourage good practice with early submission. Regards, Daniel -- |: http://berrange.com -o-http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On Wed, Aug 28, 2013 at 03:43:21AM +, Joshua Harlow wrote: Why not a rotation though, I could see it beneficial to say have a group of active developers code for say a release then those developers rotate to a reviewer position only (and rotate again for every release). This allows for a flow of knowledge between reviewers and a different set of coders (instead of a looping flow since reviewers are also coders). For a big project like nova the workload could be spread out more like that. I don't think any kind of rotation system like that is really practical. Core team members need to have the flexibility to balance their various conflicting workloads in a way that maximises their own productivity. Regards, Daniel -- |: http://berrange.com -o-http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [Nova] Scheduler support for PCI passthrough
Hi, Whilst reviewing the code I think that I have stumbled on an issue (I hope that I am mistaken). The change set (https://review.openstack.org/#/c/35749/) expects pci stats to be returned from the host. There are a number of issues here that I have concern with an would like to know what the process is for addressing the fact that the compute node may not provide these statistics, for example, this could be due to the fact that the driver has not been updated to return the pci stats or this could be if the scheduler has been upgraded prior to the compute node (what is the process for the upgrade). Thanks Gary ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] New Nova hypervisor: Docker
On Wed, Aug 28, 2013 at 06:00:50PM +1000, Michael Still wrote: On Wed, Aug 28, 2013 at 4:18 AM, Sam Alba sam.a...@gmail.com wrote: Hi all, We've been working hard during the last couple of weeks with some people. Brian Waldon helped a lot designing the Glance integration and driver testing. Dean Troyer helped a lot on bringing Docker support in Devstack[1]. On top of that, we got several feedback on the Nova code review which definitely helped to improve the code. The blueprint[2] explains what Docker brings to Nova and how to use it. I have to say that this blueprint is a fantastic example of how we should be writing design documents. It addressed almost all of my questions about the integration. Yes, Sam ( any of the other Docker guys involved) have been great at responding to reviewers' requests to expand their design document. The latest update has really helped in understanding how this driver works in the context of openstack from an architectural and functional POV. However, it would be nice for this to be on Launchpad, that being where we track blueprints. Would it be possible for you to move it over there? Or just link to the design doc from there? Their blueprint on launchpad already links to the doc in fact, so no change is needed https://blueprints.launchpad.net/nova/+spec/new-hypervisor-docker links to https://github.com/dotcloud/openstack-docker/blob/master/docs/nova_blueprint.md Regards, Daniel -- |: http://berrange.com -o-http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [savanna] team meeting reminder August 29 18:00 UTC
Hi folks, We'll be have the Savanna team meeting as usual in #openstack-meeting-alt channel. Agenda: https://wiki.openstack.org/wiki/Meetings/SavannaAgenda#Agenda_for_August.2C_29 http://www.timeanddate.com/worldclock/fixedtime.html?msg=Savanna+Meetingiso=20130829T18 Sincerely yours, Sergey Lukjanov Savanna Technical Lead Mirantis Inc. ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [Ceilometer] Meeting agenda for Wed Aug 28 at 2100 UTC
The Ceilometer project team holds a meeting in #openstack-meeting, see https://wiki.openstack.org/wiki/Meetings/MeteringAgenda for more details. Next meeting is on Wed Aug 28 at 2100 UTC Please add your name with the agenda item, so we know who to call on during the meeting. * Review Havana-3 milestone * https://launchpad.net/ceilometer/+milestone/havana-3 * Ditch Alembic for Havana? (sandy) * any plans on expanding Ceilometer coverage (ie. capacity planning, optimization, dashboard, analytics) -- gordc * are they items for other projects/products to cover? * expanding metrics captured -- gordc * plans to support beyond KVM(libvirt)? ie. VMware, IBM Power, IBM z/VM, Hyper-V, Cirtix Xen * Release python-ceilometerclient? * Open discussion If you are not able to attend or have additional topic(s) you would like to add, please update the agenda on the wiki. Cheers, -- Julien Danjou ;; Free Software hacker ; independent consultant ;; http://julien.danjou.info signature.asc Description: PGP signature ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On 28 August 2013 21:13, Daniel P. Berrange berra...@redhat.com wrote: On Wed, Aug 28, 2013 at 03:43:21AM +, Joshua Harlow wrote: For a big project like nova the workload could be spread out more like that. I don't think any kind of rotation system like that is really practical. Core team members need to have the flexibility to balance their various conflicting workloads in a way that maximises their own productivity. So does everyone else, surely? Are you saying 'I don't think I can commit to regular reviewing', or are you saying 'all reviewers will be unable to commit to regular reviewing'? Or something else? There are what - 300? ATC's for nova, and 20 core *reviewers*. http://russellbryant.net/openstack-stats/nova-reviewers-90.txt (taking 90 days to avoid some of the peak bulge). Total reviews: 11327 (6311 by core) Total reviewers: 290 Thats 125 reviews a day, or 6 per core reviewer, if core reviewers did every single review. Or 3 reviews a day at the moment. Adjusting up by 2/7 to cover weekends thats 8 per day if core did every review, and 4 per day for the ones they actually did over that period. Say it takes 20m to do a good review; thats 2.5 hours, more or less. Thats daily - thats certainly a large enough time period that I can see a rotation being potentially useful, for folk that need to discuss their patch in realtime (to reduce roundtrips etc - I think everyone knows how useful that can be). Separately, look at the math - if we assume that core reviewers are twice as productive within Nova due to familiarity with more code, we can expect at most 40 peoples worth of contributions from nova-core, vs 280 odd from ~nova-core, all other things things being equal. If reviewing is (say) 1/10th the time of writing the code, then 260 contributors would create a review load that can fully saturate 52 reviewers (/10 * 2 for the two +2s). Are we there yet? Arguably yes - there are lots of active reviewers doing as many reviews as most of the core team. And it's taking a week to review things at the moment, which means plenty of time for things to change under the patch and actually cause more review work due to rework. OTOH those are very coarse numbers with lots of assumptions. My main point is that scaling Nova development is hard, the problems are real, and right now it's a significant time investment needed for anyone wanting to become a core reviewer. -Rob -- Robert Collins rbtcoll...@hp.com Distinguished Technologist HP Converged Cloud ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On Wed, Aug 28, 2013 at 10:29:23PM +1200, Robert Collins wrote: On 28 August 2013 21:13, Daniel P. Berrange berra...@redhat.com wrote: On Wed, Aug 28, 2013 at 03:43:21AM +, Joshua Harlow wrote: For a big project like nova the workload could be spread out more like that. I don't think any kind of rotation system like that is really practical. Core team members need to have the flexibility to balance their various conflicting workloads in a way that maximises their own productivity. So does everyone else, surely? Are you saying 'I don't think I can commit to regular reviewing', or are you saying 'all reviewers will be unable to commit to regular reviewing'? Or something else? No, IIUC, Joshua was suggesting that core team members spend one cycle doing reviews only, with no coding, and then reverse for the next cycle. That is just far too coarse/crude. Core team members need to be free to balance their time between reviews and coding work on an ongoing basis, just as any other member of the community can. Regards, Daniel -- |: http://berrange.com -o-http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On 28 August 2013 22:39, Daniel P. Berrange berra...@redhat.com wrote: No, IIUC, Joshua was suggesting that core team members spend one cycle doing reviews only, with no coding, and then reverse for the next cycle. That is just far too coarse/crude. Core team members need to be free to balance their time between reviews and coding work on an ongoing basis, just as any other member of the community can. Oh! Yes, thats way too heavy. I do wonder about tweaking the balance more [or scaling the review team :)], but only-reviews would drive anyone batty. -Rob -- Robert Collins rbtcoll...@hp.com Distinguished Technologist HP Converged Cloud ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [horizon] Add a library js for creating charts
The Rickshaw library is in the Master. Building of the reusable charts on top of it is in the progress. On 08/27/2013 02:51 PM, Chmouel Boudjnah wrote: Julien Danjou jul...@danjou.info writes: It sounds like a good plan to pick Rickshaw. Better building on top of it, contributing back to it, rather than starting cold or building a new wheel. +1 Chmouel. ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [Swift] Container-sync automation + enabling a placement engine
Hi, We see a need for enhancing Swift to support a federation of swift clusters such that a set of clusters can work as a unified namespace and allow control over placement between the clusters. This requires multiple extensions to swift. To promote community inputs and work in this area, I have uploaded today https://github.com/davidhadas/Autosync. The uploaded autosync.py middleware allows working with two clusters as a single entity. It uses the container-sync mechanism but removes the need for the admin to configure each and every container with appropriate sync information and keys. This code was designed with multiple clusters in mind and assumes that there is a placement engine deciding which container is placed where (not included). As an alternative, configuration directives can be used to define the default placement (shown as an example configuration) Yet currently is is aimed only for a first case in which there is a primary and a backup (secondary) clusters. Much more work is needed to support a federation of clusters, yet this code may be useful to some even as is (with some cleanups etc). Anyone seeking to join and help push this direction forward is most welcomed. DH Regards, David Hadas, Openstack Swift ATC, Architect, Master Inventor IBM Research Labs, Haifa Tel:Int+972-4-829-6104 Fax: Int+972-4-829-6112 ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] multiple-scheduler-drivers blueprint
It seems that the main concern was that the overridden scheduler properties are taken from the flavor, and not from the aggregate. In fact, there was a consensus that this is not optimal. I think that we can still make some progress in Havana towards per-aggregate overrides, generalizing on the recently merged changes that do just that -- for cpu and for memory with FilterScheduler (and leveraging a bit from the original multi-sched patch). As follows: 1. individual filters will call get_config('abc') instead of CONF.abc (already implemented in the current version of the multi-sched patch, e.g., https://review.openstack.org/#/c/37407/30/nova/scheduler/filters/io_ops_filter.py ) 2. get_config() will check whether abc is defined in the aggregate, and if so will return the value from the aggregate, and CONF.abc otherwise (already implemented in recently merged AggregateCoreFilter and AggregateRamFilter -- e.g., https://review.openstack.org/#/c/33949/2/nova/scheduler/filters/core_filter.py ). 3. add a global flag that would enable or disable aggregate-based overrides This seems to be a relatively simple rafactoring of existing code, still achieving important portion of the original goals of this blueprint. Of course, we should still discuss the longer-term plan around scheduling policies at the summit. Thoughts? Regards, Alex From: Russell Bryant rbry...@redhat.com To: OpenStack Development Mailing List openstack-dev@lists.openstack.org, Date: 27/08/2013 10:48 PM Subject:[openstack-dev] [Nova] multiple-scheduler-drivers blueprint Greetings, One of the important things to strive for in our community is consensus. When there's not consensus, we should take a step back and see if we need to change directions. There has been a lot of iterating on this feature, and I'm afraid we still don't have consensus around the design. Phil Day has been posting some really good feedback on the review. I asked Joe Gordon to take a look and provide another opinion. He agreed with Phil that we really need to have scheduler policies be a first class API citizen. So, that pushes this feature out to Icehouse, as it doesn't seem possible to get this done in the required timeframe for Havana. If you'd really like to push to get this into Havana, please make your case. :-) Thanks, -- Russell Bryant ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
tl;dr at the end... I can ramble on a bit. I agree with Daniel. I'm not a core reviewer, but I'm trying to think like one. Over the last few weeks I've divested myself of almost all coding tasks, instead trying to increase the size of the community that is actively contributing to my area of expertise. I have indeed gone batty! I've caught myself a few times, and the frustration of feeling like I couldn't contribute code even if I wanted to is getting to be a bit too much. The common refrain I've heard is, That's OpenSource as if this is a natural state of affairs for OpenSource projects. I've been either on or around OpenSource projects for nearly 20 years at this point and I really feel this doesn't have to be the case. Any project is doomed to have an internal structure that mirrors the organization that maintains it. That means software beyond a certain scale becomes part engineering and part state-craft. In OpenSource projects that I have worked on recently, the way scale was handled was to break up the project into pieces small enough for teams of 1 to 5 to handle. The core framework developers worked on exposing API to the plugin developers. Each plugin developer would then focus on how their plugin could expose both additional API and leverage framework API. Feedback went from the application developers to the plugin developers and up to the core developers. This whole divide-and-conquer strategy was aided by the fact that we could lean heavily on a custom dependency management and code/binary distribution system leveraged inside the framework itself. It meant that package structure and distribution could be controlled by the community directly to suit its needs. That makes a powerful combination for building a flexible system but requires a fair amount of infrastructure in code and hardware. It wasn't a perfect solution. This strategy meant that an application or deployment became the coordination of plugins mixed at run-time. While efforts were made to test common combinations, it was impossible to test all combinations. That often meant people in the field were using combinations that nobody on official teams had ever considered. Because the plugins weren't on the same release cycle as the core framework (and even in different code repositories and release infrastructures) a plugin could release weekly or once every few years depending on its needs and sub-community. There is a separate dysfunction you'll see if you go down this path. Core API must necessarily lead plugin implementation ... which means you sometimes get nonsense API with no backing. To solve this a few plugins are deemed core plugins and march in-step with the API release cycle. Then there's the added burden of longer backward compatibility cycles that necessarily stretch longer and longer leaving deprecated API lying around for years as plugin developers are coaxed into leaving them behind (and subsequent plugin users are coaxed to upgrade). Some things slow down while others speed up. The core API's evolution slows, the plugin/driver speeds up. Is that a fair trade off? It's a judgement call. No right answer. In the end you trade one kind of problem for another and one kind of coordination for another. There's no clean answer that just works for this kind of problem and its why we have so many different kinds of governments in the world. Because, ultimately, that's what human coordination becomes if you don't watch out and one size does not fit all. Based on my experiences this last cycle, I think nova is pretty well broken down already. Each driver is practically its own little group, the scheduler is surprisingly well fenced off, as are cells. As for our sub-team I think we could have moved much faster if we had been able to approve our own driver blueprints some of which have been in review since May and others which have 30+ revisions updated every few days hoping for attention. It's part of why I moved to watching over people's work instead of doing my own and I now spend most of my time giving feedback to reviews other people are working on and seeking out expert opinions on other people's efforts. It's not a pleasant place to be and every time I pick up something to work on I either get pulled away or someone else picks up the job and finishes before I can even get started. I imagine this is much like what it is to be a core developer and that this contest of interest is the same strain the core-reviewers feel. You end up picking your own work and neglecting others or falling on the sword so other people can do their work and doing none of your own. Frankly, I don't want to use this strategy next cycle because it is far too unsatisfying for me. BTW: Anyone interested in this on an academic level, most of these ideas I have are from vague recollections of college readings of the work of W. Edwards Deming, Coase theorem, and more humorously and
Re: [openstack-dev] [Nova] Frustrations with review wait times
Robert Collins wrote: So I'd like to throw two ideas into the mix. Firstly, consider having a rota - ideally 24x5 but that will need some more geographical coverage I suspect for many projects - of folk who spend a dedicated time period only reviewing. We have been doing that in the past for Nova, with little success. The reason why reviewday is called reviewday is because... well... there were review days. The wiki page was a bit eaten by the wiki conversion, but you can still read it at: https://wiki.openstack.org/wiki/Nova/ReviewDays In the end, a strict rotation didn't work out because people just didn't review on their review day, but rather when they have one hour free waiting for a patch to pass gate or whatever. In the end, the rotation gave us way worse results than random ad-hoc reviewing, because people would stop reviewing on days other than their review day, and would regularly skip their review day altogether. Furthermore, there is some specialization going on: I prefer the two Xen experts in nova-core to review one hour every two days rather than one day every two weeks... because then Xen patches get better review roundtrip times. So I'm not convinced *at all* that a reboot of this would yield better results. Launchpad [the project, not the site] did this with considerable success : every qualified reviewer committed to a time slot and didn't *try* to code - they focused on reviews. The key difference is that every qualified reviewer was employed by the same company, and the review day was enforced by their management. The amount of patches is also significantly lower, and there is less specialization effect. -- Thierry Carrez (ttx) ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On Aug 28, 2013 9:42 AM, Shawn Hartsock hartso...@vmware.com wrote: tl;dr at the end... I can ramble on a bit. I agree with Daniel. I'm not a core reviewer, but I'm trying to think like one. Over the last few weeks I've divested myself of almost all coding tasks, instead trying to increase the size of the community that is actively contributing to my area of expertise. I have indeed gone batty! I've caught myself a few times, and the frustration of feeling like I couldn't contribute code even if I wanted to is getting to be a bit too much. The common refrain I've heard is, That's OpenSource as if this is a natural state of affairs for OpenSource projects. I've been either on or around OpenSource projects for nearly 20 years at this point and I really feel this doesn't have to be the case. Any project is doomed to have an internal structure that mirrors the organization that maintains it. That means software beyond a certain scale becomes part engineering and part state-craft. In OpenSource projects that I have worked on recently, the way scale was handled was to break up the project into pieces small enough for teams of 1 to 5 to handle. The core framework developers worked on exposing API to the plugin developers. Each plugin developer would then focus on how their plugin could expose both additional API and leverage framework API. Feedback went from the application developers to the plugin developers and up to the core developers. This whole divide-and-conquer strategy was aided by the fact that we could lean heavily on a custom dependency management and code/binary distribution system leveraged inside the framework itself. It meant that package structure and distribution could be controlled by the community directly to suit its needs. That makes a powerful combination for building a flexible system but requires a fair amount of infrastructure in code and hardware. It wasn't a perfect solution. This strategy meant that an application or deployment became the coordination of plugins mixed at run-time. While efforts were made to test common combinations, it was impossible to test all combinations. That often meant people in the field were using combinations that nobody on official teams had ever considered. Because the plugins weren't on the same release cycle as the core framework (and even in different code repositories and release infrastructures) a plugin could release weekly or once every few years depending on its needs and sub-community. There is a separate dysfunction you'll see if you go down this path. Core API must necessarily lead plugin implementation ... which means you sometimes get nonsense API with no backing. To solve this a few plugins are deemed core plugins and march in-step with the API release cycle. Then there's the added burden of longer backward compatibility cycles that necessarily stretch longer and longer leaving deprecated API lying around for years as plugin developers are coaxed into leaving them behind (and subsequent plugin users are coaxed to upgrade). Some things slow down while others speed up. The core API's evolution slows, the plugin/driver speeds up. Is that a fair trade off? It's a judgement call. No right answer. In the end you trade one kind of problem for another and one kind of coordination for another. There's no clean answer that just works for this kind of problem and its why we have so many different kinds of governments in the world. Because, ultimately, that's what human coordination becomes if you don't watch out and one size does not fit all. Based on my experiences this last cycle, I think nova is pretty well broken down already. Each driver is practically its own little group, the scheduler is surprisingly well fenced off, as are cells. As for our sub-team I think we could have moved much faster if we had been able to approve our own driver blueprints some of which have been in review since May and others which have 30+ revisions updated every few days hoping for attention. It's part of why I moved to watching over people's work instead of doing my own and I now spend most of my time giving feedback to reviews other people are working on and seeking out expert opinions on other people's efforts. Updating patches for attention often does the opposite for me. When I see a patch set being revised every few days, that makes me think ( perhaps incorrectly) that the patch is still in active development and I am inclined to review something else. On a related note, I really like when the developer adds a gerrit comment saying why the revision, that makes my life as a reviewer easier. It's not a pleasant place to be and every time I pick up something to work on I either get pulled away or someone else picks up the job and finishes before I can even get started. I imagine this is much like what it is to be a core developer and that this contest of interest is the same strain the core-reviewers feel. You end up picking your own work and
Re: [openstack-dev] New Nova hypervisor: Docker
On 08/28/2013 05:18 AM, Daniel P. Berrange wrote: On Wed, Aug 28, 2013 at 06:00:50PM +1000, Michael Still wrote: On Wed, Aug 28, 2013 at 4:18 AM, Sam Alba sam.a...@gmail.com wrote: Hi all, We've been working hard during the last couple of weeks with some people. Brian Waldon helped a lot designing the Glance integration and driver testing. Dean Troyer helped a lot on bringing Docker support in Devstack[1]. On top of that, we got several feedback on the Nova code review which definitely helped to improve the code. The blueprint[2] explains what Docker brings to Nova and how to use it. I have to say that this blueprint is a fantastic example of how we should be writing design documents. It addressed almost all of my questions about the integration. Yes, Sam ( any of the other Docker guys involved) have been great at responding to reviewers' requests to expand their design document. The latest update has really helped in understanding how this driver works in the context of openstack from an architectural and functional POV. They've been great in responding to my requests, as well. The biggest thing was that I wanted to see devstack support so that it's easily testable, both by developers and by CI. They delivered. So, in general, I'm good with this going in. It's just a matter of getting the code review completed in the next week before feature freeze. I'm going to try to help with it this week. -- Russell Bryant ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Ceilometer] Need help with Alembic...
On Tue, Aug 27, 2013 at 12:30 PM, Jay Pipes jaypi...@gmail.com wrote: On 08/27/2013 04:32 AM, Boris Pavlovic wrote: Jay, I should probably share to you about our work around DB. Migrations should be run only in production and only for production backends (e.g. psql and mysql) In tests we should use Schemas created by Models (BASE.metadata.create_all()) Agree on both. We are not able to use in this approach in moment because we don't have any mechanism to check that MODELS and SCHEMAS are EQUAL. And actually MODELS and SCHEMAS are DIFFERENT. Sorry, I don't understand the connection... how does not having a codified way of determining the difference between model and schema (BTW, this does exist in sqlalchemy-migrate... look at the compare_model_to_db method) not allow you to use metadata.create_all() in tests or mean that you can't run migrations only in production? E.g. in Celiometer we have BP that syncs models and migration https://blueprints.launchpad.**net/ceilometer/+spec/** ceilometer-db-sync-models-**with-migrationshttps://blueprints.launchpad.net/ceilometer/+spec/ceilometer-db-sync-models-with-migrations (in other projects we are doing the same) And also we are working around (oslo) generic tests that checks that models and migrations are equal: https://review.openstack.org/#**/c/42307/https://review.openstack.org/#/c/42307/ OK, cool. So in our roadmap (in this case is): 1) Soft switch to alembic (with code that allows to have sqla-migrate and alembic migration in the same time) I don't see the point in this at all... I would rather see patches that just switch to Alembic and get rid of SQLAlchemy-migrate. Create an initial Alembic migration that has the last state of the database schema under SQLAlchemy-migrate... and then delete SA-Migrate. We had a rather long discussion about this on the mailing list a while back. We decided not to spend time changing the existing migrations because we didn't want to introduce differences for anyone doing continuous deployment. The work to add alembic was supposed to mark the soft-switch, and it looks like you're the (unlucky) first person to try to create an actual alembic migration script. We have an agenda item on the ceilometer meeting for today to discuss what to do. At this point I think we should stick with sqlalchemy-migrate to avoid causing delays before the H3 deadline. Dou 2) Sync Models and Migrations (fix DB schemas also) 3) Add from oslo generic test that checks all this stuff 4) Use BASE.create_all() for Schema creation instead of migrations. This is already done in some projects, IIRC... (Glance used to be this way, at least) But in OpenStack is not so simple to implement such huge changes, so it take some time=) Best regards, Boris Pavlovic --- Mirantis Inc. On Tue, Aug 27, 2013 at 12:02 AM, Jay Pipes jaypi...@gmail.com mailto:jaypi...@gmail.com wrote: On 08/26/2013 03:40 PM, Herndon, John Luke (HPCS - Ft. Collins) wrote: Jay - It looks there is an error in the migration script that causes it to abort: AttributeError: 'ForeignKeyConstraint' object has no attribute 'drop' My guess is the migration runs on the first test, creates event types table fine, but exits with the above error, so migration is not complete. Thus every subsequent test tries to migrate the db, and notices that event types already exists. I'd corrected that particular mistake and pushed an updated migration script. Best, -jay -john On 8/26/13 1:15 PM, Jay Pipes jaypi...@gmail.com mailto:jaypi...@gmail.com wrote: I just noticed that every single test case for SQL-driver storage is executing every single migration upgrade before every single test case run: https://github.com/openstack/_**_ceilometer/blob/master/__** ceilometer/tests/db.pyhttps://github.com/openstack/__ceilometer/blob/master/__ceilometer/tests/db.py https://github.com/openstack/**ceilometer/blob/master/** ceilometer/tests/db.pyhttps://github.com/openstack/ceilometer/blob/master/ceilometer/tests/db.py #L46 https://github.com/openstack/_**_ceilometer/blob/master/__** ceilometer/storage/imphttps://github.com/openstack/__ceilometer/blob/master/__ceilometer/storage/imp https://github.com/openstack/**ceilometer/blob/master/** ceilometer/storage/imphttps://github.com/openstack/ceilometer/blob/master/ceilometer/storage/imp l_sqlalchemy.py#L153 instead of simply creating a new database schema from the models in the current source code base using a call to sqlalchemy.MetaData.create___**all(). This results in re-running migrations over and over again,
Re: [openstack-dev] [Nova] Frustrations with review wait times
On 08/28/2013 10:58 AM, Thierry Carrez wrote: Robert Collins wrote: So I'd like to throw two ideas into the mix. Firstly, consider having a rota - ideally 24x5 but that will need some more geographical coverage I suspect for many projects - of folk who spend a dedicated time period only reviewing. We have been doing that in the past for Nova, with little success. The reason why reviewday is called reviewday is because... well... there were review days. The wiki page was a bit eaten by the wiki conversion, but you can still read it at: https://wiki.openstack.org/wiki/Nova/ReviewDays In the end, a strict rotation didn't work out because people just didn't review on their review day, but rather when they have one hour free waiting for a patch to pass gate or whatever. In the end, the rotation gave us way worse results than random ad-hoc reviewing, because people would stop reviewing on days other than their review day, and would regularly skip their review day altogether. Furthermore, there is some specialization going on: I prefer the two Xen experts in nova-core to review one hour every two days rather than one day every two weeks... because then Xen patches get better review roundtrip times. So I'm not convinced *at all* that a reboot of this would yield better results. +1 That said, I think the reason my reviews dropped off from Nova was not having a dedicated day for it. But that was my fault, not the fault of the process. With Ceilometer, I try to set aside one fixed day a week for reviews (with moderate success ;) Launchpad [the project, not the site] did this with considerable success : every qualified reviewer committed to a time slot and didn't *try* to code - they focused on reviews. The key difference is that every qualified reviewer was employed by the same company, and the review day was enforced by their management. The amount of patches is also significantly lower, and there is less specialization effect. ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
Hi, I am not sure that there is a good solution. I guess that we all need to 'vasbyt' (that is Afrikaans for bite the bullet) and wait for the code posted to be reviewed. In Neutron when we were heading towards the end of a cycle and there were a ton of BP's being added the PTL would ensure that there were at least two reviewer on each BP. This would address the problem in two ways: 1. Accountability for the review process in the critical time period 2. The coder was able to have a person that he/she could be in touch with. The above would enhance the cadence of the reviews. I personally am spending a few hours a day reviewing code. I hope that it is helping move things forwards. A review not only means just looking at the code (there are some cases that it is simple), but it means running and testing the code. In some cases it is not possible to test (for example a Mellanox vif driver). In cases when a reviewer does not have an option to test the code would a tempest run help the reviewer with his/her decision? Thanks and Alut a continua Gary -Original Message- From: Sandy Walsh [mailto:sandy.wa...@rackspace.com] Sent: Wednesday, August 28, 2013 5:15 PM To: OpenStack Development Mailing List Subject: Re: [openstack-dev] [Nova] Frustrations with review wait times On 08/28/2013 10:58 AM, Thierry Carrez wrote: Robert Collins wrote: So I'd like to throw two ideas into the mix. Firstly, consider having a rota - ideally 24x5 but that will need some more geographical coverage I suspect for many projects - of folk who spend a dedicated time period only reviewing. We have been doing that in the past for Nova, with little success. The reason why reviewday is called reviewday is because... well... there were review days. The wiki page was a bit eaten by the wiki conversion, but you can still read it at: https://wiki.openstack.org/wiki/Nova/ReviewDays In the end, a strict rotation didn't work out because people just didn't review on their review day, but rather when they have one hour free waiting for a patch to pass gate or whatever. In the end, the rotation gave us way worse results than random ad-hoc reviewing, because people would stop reviewing on days other than their review day, and would regularly skip their review day altogether. Furthermore, there is some specialization going on: I prefer the two Xen experts in nova-core to review one hour every two days rather than one day every two weeks... because then Xen patches get better review roundtrip times. So I'm not convinced *at all* that a reboot of this would yield better results. +1 That said, I think the reason my reviews dropped off from Nova was not having a dedicated day for it. But that was my fault, not the fault of the process. With Ceilometer, I try to set aside one fixed day a week for reviews (with moderate success ;) Launchpad [the project, not the site] did this with considerable success : every qualified reviewer committed to a time slot and didn't *try* to code - they focused on reviews. The key difference is that every qualified reviewer was employed by the same company, and the review day was enforced by their management. The amount of patches is also significantly lower, and there is less specialization effect. ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [Ceilometer] mid-Icehouse-cycle meet-up (was: [Nova] Interested in a mid-Icehouse-cycle Nova meet-up?)
On Tue, Aug 27 2013, Thierry Carrez wrote: Daniel P. Berrange wrote: Is openstack looking to have a strong presence at FOSDEM 2014 ? I didn't make it to FOSDEM this year, but IIUC, there were quite a few openstack contributors talks in 2013. Yes, we are aiming for a devroom again at FOSDEM this year. IOW, should we consider holding the meetup in Brussels just before/after FOSDEM, so that people who want/need to attend both can try to maximise utilization of their often limited travel budgets and/or minimise the number of days lost to travelling ? I would certainly like that, but I'm not sure the center of gravity for Nova contributors is in Europe :) That could work for Ceilometer folks, I think we could gather half the core team at FOSDEM ;-) If people are interested, let's keep it in our mind to try to come up with something by then. -- Julien Danjou /* Free Software hacker * independent consultant http://julien.danjou.info */ signature.asc Description: PGP signature ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [TRIPLEO] Derekh for tripleo core
+1 On Tue, Aug 27, 2013 at 5:25 PM, Robert Collins robe...@robertcollins.netwrote: http://russellbryant.net/openstack-stats/tripleo-reviewers-30.txt http://russellbryant.net/openstack-stats/tripleo-reviewers-90.txt - Derek is reviewing fairly regularly and has got a sense of the culture etc now, I think. So - calling for votes for Derek to become a TripleO core reviewer! I think we're nearly at the point where we can switch to the 'two +2's' model - what do you think? Also tsk! to those cores who aren't reviewing as regularly :) Cheers, Rob -- Robert Collins rbtcoll...@hp.com Distinguished Technologist HP Converged Cloud ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- -- James Slagle -- ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [nova][tempest][rdo] looking for opinions on change 43298
On 08/27/2013 03:52 PM, Matt Riedemann wrote: This change: _https://review.openstack.org/#/c/43298/_ Is attempting to fix a bug where a tempest test fails when nova-manage --version is different from nova-manage version when using a RHEL 6 installation rather than devstack. Pavel points out an RDO bug that was filed back in April to address the issue: _https://bugzilla.redhat.com/show_bug.cgi?id=952811_ That RDO bug hasn't gotten any attention though (I wasn't aware of it when I reported the launchpad bug). So my question is, is this worth changing in Tempest or should we expect that nova-manage --version will always equal nova-manage version? I'm not even really sure how they are getting their values, one appears to be coming from the python distribution and one from the rpm (looks like argparse must do something there). My opinion is that nova-manage version and nova-manage --version should return the same thing, because user interfaces should look like they were designed on purpose, and nobody would intentionally design those two almost identical commands to return different things. I think it's a bug that needs to be fixed in nova-manage rather than something Tempest should have to work around. -- David Ripton Red Hat drip...@redhat.com ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On 08/28/2013 05:10 AM, Daniel P. Berrange wrote: IOW we should prioritize review of work whose authors submitted earlier to encourage good practice with early submission. +1. Can we reconfigure Gerrit to show oldest first rather than newest first by default? (next-review does this. next-review is awesome. Everyone should try next-review. But we should try to make Gerrit do the right thing too, just in case some people prefer it to next-review.) -- David Ripton Red Hat drip...@redhat.com ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On 08/28/2013 10:31 AM, Gary Kotton wrote: Hi, I am not sure that there is a good solution. I guess that we all need to 'vasbyt' (that is Afrikaans for bite the bullet) and wait for the code posted to be reviewed. In Neutron when we were heading towards the end of a cycle and there were a ton of BP's being added the PTL would ensure that there were at least two reviewer on each BP. This would address the problem in two ways: 1. Accountability for the review process in the critical time period 2. The coder was able to have a person that he/she could be in touch with. The above would enhance the cadence of the reviews. I personally am spending a few hours a day reviewing code. I hope that it is helping move things forwards. A review not only means just looking at the code (there are some cases that it is simple), but it means running and testing the code. In some cases it is not possible to test (for example a Mellanox vif driver). In cases when a reviewer does not have an option to test the code would a tempest run help the reviewer with his/her decision? Thanks and Alut a continua Gary Well, in general almost all of the tempest tests are gating on all projects. So if jenkins says +1 then tempest has passed. The unfortunate exception is that the jenkins job that runs all tempest tests for a neutron configuration has never passed and was non-voting. This week it was further demoted to the experimental queue where it will only run if some one tells it to :-( . I also suspect that tempest coverage of neutron is not as good as for other projects. -David ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
Shrinking that rotation granularity would be reasonable to. Rotate once every 2 weeks or some other time period still seems useful to me. Sent from my really tiny device... On Aug 28, 2013, at 3:43 AM, Daniel P. Berrange berra...@redhat.com wrote: On Wed, Aug 28, 2013 at 10:29:23PM +1200, Robert Collins wrote: On 28 August 2013 21:13, Daniel P. Berrange berra...@redhat.com wrote: On Wed, Aug 28, 2013 at 03:43:21AM +, Joshua Harlow wrote: For a big project like nova the workload could be spread out more like that. I don't think any kind of rotation system like that is really practical. Core team members need to have the flexibility to balance their various conflicting workloads in a way that maximises their own productivity. So does everyone else, surely? Are you saying 'I don't think I can commit to regular reviewing', or are you saying 'all reviewers will be unable to commit to regular reviewing'? Or something else? No, IIUC, Joshua was suggesting that core team members spend one cycle doing reviews only, with no coding, and then reverse for the next cycle. That is just far too coarse/crude. Core team members need to be free to balance their time between reviews and coding work on an ongoing basis, just as any other member of the community can. Regards, Daniel -- |: http://berrange.com -o-http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] New Nova hypervisor: Docker
Thanks a lot everyone for the nice feedback. I am going to work hard to get all those new comments addressed to be able to re-submit a new patchset today or tomorrow (the later). On Wed, Aug 28, 2013 at 7:02 AM, Russell Bryant rbry...@redhat.com wrote: On 08/28/2013 05:18 AM, Daniel P. Berrange wrote: On Wed, Aug 28, 2013 at 06:00:50PM +1000, Michael Still wrote: On Wed, Aug 28, 2013 at 4:18 AM, Sam Alba sam.a...@gmail.com wrote: Hi all, We've been working hard during the last couple of weeks with some people. Brian Waldon helped a lot designing the Glance integration and driver testing. Dean Troyer helped a lot on bringing Docker support in Devstack[1]. On top of that, we got several feedback on the Nova code review which definitely helped to improve the code. The blueprint[2] explains what Docker brings to Nova and how to use it. I have to say that this blueprint is a fantastic example of how we should be writing design documents. It addressed almost all of my questions about the integration. Yes, Sam ( any of the other Docker guys involved) have been great at responding to reviewers' requests to expand their design document. The latest update has really helped in understanding how this driver works in the context of openstack from an architectural and functional POV. They've been great in responding to my requests, as well. The biggest thing was that I wanted to see devstack support so that it's easily testable, both by developers and by CI. They delivered. So, in general, I'm good with this going in. It's just a matter of getting the code review completed in the next week before feature freeze. I'm going to try to help with it this week. -- Russell Bryant ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- @sam_alba ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
Shawn, If each little group had at least one active Nova core member, i think it would speed things up way faster IMHO. -- dims On Wed, Aug 28, 2013 at 9:40 AM, Shawn Hartsock hartso...@vmware.comwrote: tl;dr at the end... I can ramble on a bit. I agree with Daniel. I'm not a core reviewer, but I'm trying to think like one. Over the last few weeks I've divested myself of almost all coding tasks, instead trying to increase the size of the community that is actively contributing to my area of expertise. I have indeed gone batty! I've caught myself a few times, and the frustration of feeling like I couldn't contribute code even if I wanted to is getting to be a bit too much. The common refrain I've heard is, That's OpenSource as if this is a natural state of affairs for OpenSource projects. I've been either on or around OpenSource projects for nearly 20 years at this point and I really feel this doesn't have to be the case. Any project is doomed to have an internal structure that mirrors the organization that maintains it. That means software beyond a certain scale becomes part engineering and part state-craft. In OpenSource projects that I have worked on recently, the way scale was handled was to break up the project into pieces small enough for teams of 1 to 5 to handle. The core framework developers worked on exposing API to the plugin developers. Each plugin developer would then focus on how their plugin could expose both additional API and leverage framework API. Feedback went from the application developers to the plugin developers and up to the core developers. This whole divide-and-conquer strategy was aided by the fact that we could lean heavily on a custom dependency management and code/binary distribution system leveraged inside the framework itself. It meant that package structure and distribution could be controlled by the community directly to suit its needs. That makes a powerful combination for building a flexible system but requires a fair amount of infrastructure in code and hardware. It wasn't a perfect solution. This strategy meant that an application or deployment became the coordination of plugins mixed at run-time. While efforts were made to test common combinations, it was impossible to test all combinations. That often meant people in the field were using combinations that nobody on official teams had ever considered. Because the plugins weren't on the same release cycle as the core framework (and even in different code repositories and release infrastructures) a plugin could release weekly or once every few years depending on its needs and sub-community. There is a separate dysfunction you'll see if you go down this path. Core API must necessarily lead plugin implementation ... which means you sometimes get nonsense API with no backing. To solve this a few plugins are deemed core plugins and march in-step with the API release cycle. Then there's the added burden of longer backward compatibility cycles that necessarily stretch longer and longer leaving deprecated API lying around for years as plugin developers are coaxed into leaving them behind (and subsequent plugin users are coaxed to upgrade). Some things slow down while others speed up. The core API's evolution slows, the plugin/driver speeds up. Is that a fair trade off? It's a judgement call. No right answer. In the end you trade one kind of problem for another and one kind of coordination for another. There's no clean answer that just works for this kind of problem and its why we have so many different kinds of governments in the world. Because, ultimately, that's what human coordination becomes if you don't watch out and one size does not fit all. Based on my experiences this last cycle, I think nova is pretty well broken down already. Each driver is practically its own little group, the scheduler is surprisingly well fenced off, as are cells. As for our sub-team I think we could have moved much faster if we had been able to approve our own driver blueprints some of which have been in review since May and others which have 30+ revisions updated every few days hoping for attention. It's part of why I moved to watching over people's work instead of doing my own and I now spend most of my time giving feedback to reviews other people are working on and seeking out expert opinions on other people's efforts. It's not a pleasant place to be and every time I pick up something to work on I either get pulled away or someone else picks up the job and finishes before I can even get started. I imagine this is much like what it is to be a core developer and that this contest of interest is the same strain the core-reviewers feel. You end up picking your own work and neglecting others or falling on the sword so other people can do their work and doing none of your own. Frankly, I don't want to use this strategy next cycle because it is far too unsatisfying
Re: [openstack-dev] [Nova] Frustrations with review wait times
On 08/28/2013 10:31 AM, Gary Kotton wrote: Hi, I am not sure that there is a good solution. I guess that we all need to 'vasbyt' (that is Afrikaans for bite the bullet) and wait for the code posted to be reviewed. In Neutron when we were heading towards the end of a cycle and there were a ton of BP's being added the PTL would ensure that there were at least two reviewer on each BP. This would address the problem in two ways: 1. Accountability for the review process in the critical time period 2. The coder was able to have a person that he/she could be in touch with. The above would enhance the cadence of the reviews. I personally am spending a few hours a day reviewing code. I hope that it is helping move things forwards. A review not only means just looking at the code (there are some cases that it is simple), but it means running and testing the code. In some cases it is not possible to test (for example a Mellanox vif driver). In cases when a reviewer does not have an option to test the code would a tempest run help the reviewer with his/her decision? Thanks and Alut a continua Gary Just to clarify my last message, there is still a gating job called gate-tempest-devstack-vm-neutron http://logs.openstack.org/58/43658/4/check/gate-tempest-devstack-vm-neutron/10f1a5a but it only runs the smoke tests which is a small subset of tempest. -David ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On 08/28/2013 12:25 PM, Davanum Srinivas wrote: If each little group had at least one active Nova core member, i think it would speed things up way faster IMHO. Agreed, in theory. However, we should not add someone just for the sake of having someone on the team from a certain area. They need to be held to the same standards as the rest of the team. -- Russell Bryant ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
+1000 Russell -- dims On Wed, Aug 28, 2013 at 1:06 PM, Russell Bryant rbry...@redhat.com wrote: On 08/28/2013 12:25 PM, Davanum Srinivas wrote: If each little group had at least one active Nova core member, i think it would speed things up way faster IMHO. Agreed, in theory. However, we should not add someone just for the sake of having someone on the team from a certain area. They need to be held to the same standards as the rest of the team. -- Russell Bryant ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Davanum Srinivas :: http://davanum.wordpress.com ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] endpoint registration
The downside of doing version discovery in the client is that it adds a third round trip... though the client can cache the support versions I guess. On 18 August 2013 00:14, Joshua Harlow harlo...@yahoo-inc.com wrote: +3 Sent from my really tiny device... On Aug 17, 2013, at 8:33 AM, Dolph Mathews dolph.math...@gmail.com wrote: On Sat, Aug 17, 2013 at 4:18 AM, Julien Danjou jul...@danjou.info wrote: On Fri, Aug 16 2013, Doug Hellmann wrote: If you're saying that you want to register URLs without version info embedded in them, and let the client work that part out by talking to the service in question (or getting a version number from the caller), then yes, please. Yes yes yes. I already started for Swift about that a while back which got 0 reply. There's no point in registering URL with version suffix and others stuff. We should stop doing that in all places, including documentation. +2! -- Julien Danjou /* Free Software hacker * freelance consultant http://julien.danjou.info */ ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- -Dolph ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Duncan Thomas ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On Wed, Aug 28, 2013, Russell Bryant rbry...@redhat.com wrote: On 08/28/2013 12:25 PM, Davanum Srinivas wrote: If each little group had at least one active Nova core member, i think it would speed things up way faster IMHO. Agreed, in theory. However, we should not add someone just for the sake of having someone on the team from a certain area. They need to be held to the same standards as the rest of the team. Do you mean the nova-core standards? I had a soft understanding that nova-core members were trusted to give +2 and -2 reviews and that they actually needed to do reviews. I did a quick search and didn't find anything more than that, but maybe I missed a web page somewhere. JE ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Scheduler support for PCI passthrough
Gary Firstly, thanks for your review very much. The pci_stats is calculated in the resource tracker in the compute node and is also saved in compute_node, I think currently the scheduler depends on the information provided by the compute_node table so this method should fit into current framework. Please notice that the scheduler only decide the host that can meet the requirement, and it's the resource tracker in compute node that will do the real device allocation. So if scheduler does not get the latest information, it may either can't find the host, or, it finds a host with wrong information and the re-try mechanism should work. Anyway, this is same to other compute nodes information like free_ram or free_vcpus, right? But you does remind me one thing, that if a hot plug happens , after the resource tracker select the device and before the instance is really created. Possibly the virt driver need check the requirement before create the domain. But this race condition chain will not end. After all, there are window between the virt driver check and the instance creation, and no idea how can we guarantee this. Thanks --jyh From: Gary Kotton [mailto:gkot...@vmware.com] Sent: Wednesday, August 28, 2013 2:19 AM To: OpenStack Development Mailing List (openstack-dev@lists.openstack.org) Subject: [openstack-dev] [Nova] Scheduler support for PCI passthrough Hi, Whilst reviewing the code I think that I have stumbled on an issue (I hope that I am mistaken). The change set (https://review.openstack.org/#/c/35749/) expects pci stats to be returned from the host. There are a number of issues here that I have concern with an would like to know what the process is for addressing the fact that the compute node may not provide these statistics, for example, this could be due to the fact that the driver has not been updated to return the pci stats or this could be if the scheduler has been upgraded prior to the compute node (what is the process for the upgrade). Thanks Gary ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] About multihost patch review
On Aug 26, 2013, at 6:14 PM, Maru Newby ma...@redhat.com wrote: On Aug 26, 2013, at 4:06 PM, Edgar Magana emag...@plumgrid.com wrote: Hi Developers, Let me explain my point of view on this topic and please share your thoughts in order to merge this new feature ASAP. My understanding is that multi-host is nova-network HA and we are implementing this bp https://blueprints.launchpad.net/neutron/+spec/quantum-multihost for the same reason. So, If in neutron configuration admin enables multi-host: etc/dhcp_agent.ini # Support multi host networks # enable_multihost = False Why do tenants needs to be aware of this? They should just create networks in the way they normally do and not by adding the multihost extension. I was pretty confused until I looked at the nova-network HA doc [1]. The proposed design would seem to emulate nova-network's multi-host HA option, where it was necessary to both run nova-network on every compute node and create a network explicitly as multi-host. I'm not sure why nova-network was implemented in this way, since it would appear that multi-host is basically all-or-nothing. Once nova-network services are running on every compute node, what does it mean to create a network that is not multi-host? Just to add a little background to the nova-network multi-host: The fact that the multi_host flag is stored per-network as opposed to a configuration was an implementation detail. While in theory this would support configurations where some networks are multi_host and other ones are not, I am not aware of any deployments where both are used together. That said, If there is potential value in offering both, it seems like it should be under the control of the deployer not the user. In other words the deployer should be able to set the default network type and enforce whether setting the type is exposed to the user at all. Also, one final point. In my mind, multi-host is strictly better than single host, if I were to redesign nova-network today, I would get rid of the single host mode completely. Vish So, to Edgar's question - is there a reason other than 'be like nova-network' for requiring neutron multi-host to be configured per-network? m. 1: http://docs.openstack.org/trunk/openstack-compute/admin/content/existing-ha-networking-options.html I could be totally wrong and crazy, so please provide some feedback. Thanks, Edgar From: Yongsheng Gong gong...@unitedstack.com Date: Monday, August 26, 2013 2:58 PM To: Kyle Mestery (kmestery) kmest...@cisco.com, Aaron Rosen aro...@nicira.com, Armando Migliaccio amigliac...@vmware.com, Akihiro MOTOKI amot...@gmail.com, Edgar Magana emag...@plumgrid.com, Maru Newby ma...@redhat.com, Nachi Ueno na...@nttmcl.com, Salvatore Orlando sorla...@nicira.com, Sumit Naiksatam sumit.naiksa...@bigswitch.com, Mark McClain mark.mccl...@dreamhost.com, Gary Kotton gkot...@vmware.com, Robert Kukura rkuk...@redhat.com Cc: OpenStack List openstack-dev@lists.openstack.org Subject: Re: About multihost patch review Hi, Edgar Magana has commented to say: 'This is the part that for me is confusing and I will need some clarification from the community. Do we expect to have the multi-host feature as an extension or something that will natural work as long as the deployment include more than one Network Node. In my opinion, Neutron deployments with more than one Network Node by default should call DHCP agents in all those nodes without the need to use an extension. If the community has decided to do this by extensions, then I am fine' at https://review.openstack.org/#/c/37919/11/neutron/extensions/multihostnetwork.py I have commented back, what is your opinion about it? Regards, Yong Sheng Gong On Fri, Aug 16, 2013 at 9:28 PM, Kyle Mestery (kmestery) kmest...@cisco.com wrote: Hi Yong: I'll review this and try it out today. Thanks, Kyle On Aug 15, 2013, at 10:01 PM, Yongsheng Gong gong...@unitedstack.com wrote: The multihost patch is there for a long long time, can someone help to review? https://review.openstack.org/#/c/37919/ ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] MultiStrOpt opts can be problematic...
So I recently ran into a fun config issue in trying to configure Nova to work w/ Ceilometer using Puppet: https://bugs.launchpad.net/puppet-ceilometer/+bug/1217867 Today, what you need to do to make Nova work with ceilometer is add this to your nova.conf file: notification_driver=nova.openstack.common.notifier.rpc_notifier notification_driver=ceilometer.compute.nova_notifier As it turns out Multi-line config entries aren't very fun to deal with in the config management world. The puppet nova_config provider doesn't (yet) have a good way to support them. The core of the issue is they pose all sorts of problems in knowing whether a given tool should modify the existing config values. In the short term we can look into doing one of these Puppet land: -Using a conf.d directory for config (would require a change to the nova-compute init script to use --config-dir) -String together various resources in puppet to make it work, etc. (file_line, augeus, etc) Long term though I'm thinking what if MultiStrOpt's were to go away? They seem to be more trouble than they are worth... Dan ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] Incubation Request: Marconi
On Thu, Aug 22, 2013 at 12:29 PM, Kurt Griffiths kurt.griffi...@rackspace.com wrote: What was wrong with qpid, rabbitmq, activemq, zeromq, ${your favorite queue here} that required marconi? That's a good question. The features supported by AMQP brokers, ZMQ, and Marconi certainly do overlap in some areas. At the same time, however, each of these options offer distinct features that may or may not align with what a web developer is trying to accomplish. Here are a few of Marconi's unique features, relative to the other options you mentioned: * Multi-tenant * Keystone integration * 100% Python * First-class, stateless, firewall-friendly HTTP(S) transport driver * Simple protocol, easy for clients to implement * Scales to an unlimited number of queues and clients * Per-queue stats, useful for monitoring and autoscale * Tag-based message filtering (planned) Relative to SQS, Marconi: * Is open-source and community-driven * Supports private and hybrid deployments * Offers hybrid pub-sub and producer-consumer semantics * Provides a clean, modern HTTP API * Can route messages to multiple queues (planned) * Can perform custom message transformations (planned) Anyway, that's my $0.02 - others may chime in with their own thoughts. I assume the rabbitmq vs sqs debate ( http://notes.variogr.am/post/67710296/replacing-amazon-sqs-with-something-faster-and-cheaper) is the same for rabbitmq vs marconi? @kgriffs ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] multiple-scheduler-drivers blueprint
On Wed, Aug 28, 2013 at 9:12 AM, Alex Glikson glik...@il.ibm.com wrote: It seems that the main concern was that the overridden scheduler properties are taken from the flavor, and not from the aggregate. In fact, there was a consensus that this is not optimal. I think that we can still make some progress in Havana towards per-aggregate overrides, generalizing on the recently merged changes that do just that -- for cpu and for memory with FilterScheduler (and leveraging a bit from the original multi-sched patch). As follows: 1. individual filters will call get_config('abc') instead of CONF.abc (already implemented in the current version of the multi-sched patch, e.g., https://review.openstack.org/#/c/37407/30/nova/scheduler/filters/core_filter.py * https://review.openstack.org/#/c/37407/30/nova/scheduler/filters/io_ops_filter.py *https://review.openstack.org/#/c/37407/30/nova/scheduler/filters/io_ops_filter.py ) 2. get_config() will check whether abc is defined in the aggregate, and if so will return the value from the aggregate, and CONF.abc otherwise (already implemented in recently merged AggregateCoreFilter and AggregateRamFilter -- e.g., * https://review.openstack.org/#/c/33949/2/nova/scheduler/filters/core_filter.py *https://review.openstack.org/#/c/33949/2/nova/scheduler/filters/core_filter.py ). 3. add a global flag that would enable or disable aggregate-based overrides Why can't something like this be done with just different filters, see such as for AggregateRamFilter? This seems to be a relatively simple rafactoring of existing code, still achieving important portion of the original goals of this blueprint. Of course, we should still discuss the longer-term plan around scheduling policies at the summit. Thoughts? Regards, Alex From:Russell Bryant rbry...@redhat.com To:OpenStack Development Mailing List openstack-dev@lists.openstack.org, Date:27/08/2013 10:48 PM Subject:[openstack-dev] [Nova] multiple-scheduler-drivers blueprint -- Greetings, One of the important things to strive for in our community is consensus. When there's not consensus, we should take a step back and see if we need to change directions. There has been a lot of iterating on this feature, and I'm afraid we still don't have consensus around the design. Phil Day has been posting some really good feedback on the review. I asked Joe Gordon to take a look and provide another opinion. He agreed with Phil that we really need to have scheduler policies be a first class API citizen. So, that pushes this feature out to Icehouse, as it doesn't seem possible to get this done in the required timeframe for Havana. If you'd really like to push to get this into Havana, please make your case. :-) Thanks, -- Russell Bryant ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [TRIPLEO] Derekh for tripleo core
As an outsider to OOO, but a Keystone core, let me endorse Derek's work on provioning in general. He is an outstanding developer. On 08/28/2013 10:31 AM, James Slagle wrote: +1 On Tue, Aug 27, 2013 at 5:25 PM, Robert Collins robe...@robertcollins.net mailto:robe...@robertcollins.net wrote: http://russellbryant.net/openstack-stats/tripleo-reviewers-30.txt http://russellbryant.net/openstack-stats/tripleo-reviewers-90.txt - Derek is reviewing fairly regularly and has got a sense of the culture etc now, I think. So - calling for votes for Derek to become a TripleO core reviewer! I think we're nearly at the point where we can switch to the 'two +2's' model - what do you think? Also tsk! to those cores who aren't reviewing as regularly :) Cheers, Rob -- Robert Collins rbtcoll...@hp.com mailto:rbtcoll...@hp.com Distinguished Technologist HP Converged Cloud ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org mailto:OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- -- James Slagle -- ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] endpoint registration
On Wed, Aug 28, 2013 at 12:18 PM, Duncan Thomas duncan.tho...@gmail.comwrote: The downside of doing version discovery in the client is that it adds a third round trip... though the client can cache the support versions I guess. That's only for the Identity version discovery. Add more round trips for additional APIs to be used. Also, only Keystone allows unauthenticated access to /vXX endpoints. To query any other service you need to either know the root endpoint beforehand or auth to get one from the service catalog, and then you still don't know if the version is or is not included in that endpoint (no parsing-guessing here!) due to backward-compatibility for older deployments. dt -- Dean Troyer dtro...@gmail.com ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On 08/28/2013 01:22 PM, Johannes Erdfelt wrote: On Wed, Aug 28, 2013, Russell Bryant rbry...@redhat.com wrote: On 08/28/2013 12:25 PM, Davanum Srinivas wrote: If each little group had at least one active Nova core member, i think it would speed things up way faster IMHO. Agreed, in theory. However, we should not add someone just for the sake of having someone on the team from a certain area. They need to be held to the same standards as the rest of the team. Do you mean the nova-core standards? I had a soft understanding that nova-core members were trusted to give +2 and -2 reviews and that they actually needed to do reviews. I did a quick search and didn't find anything more than that, but maybe I missed a web page somewhere. You're right that much of this has been unwritten rules. I would really like to improve that. I started this page recently, but I'm sure there's more that could be added: https://wiki.openstack.org/wiki/Nova/CoreTeam -- Russell Bryant ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [keystone][horizon]Backend filtering in Keystone
While these are useful steps in isolation, I'm hesitant to just say go for it! because I see this as a problem that OpenStack as a whole needs to solve. Your implementation here is a good proof-of-concept that's likely worth vetting and then emulating elsewhere. However looking at it a different way it adds further fragmentation to the filtering, sorting, paging, etc. methods that Horizon has to attempt to support across all the projects. It's my intention to run a session on this at the summit and probably walk out of that summit with a Horizon will support X, so if you want people to have a good experience with your project in Horizon you should support X too kind of agreement that we can work towards across projects in Icehouse. That X will likely reflect what you've done here and the great discussions that happened recently about the possibility of doing away with pagination entirely. If the patch is ready to go then by all means merge it and we can start playing with it to see where it shines and where it needs polish. I'm all for it in principle. All the best, - Gabriel From: Henry Nash [mailto:hen...@linux.vnet.ibm.com] Sent: Wednesday, August 28, 2013 1:58 AM To: Gabriel Hurley Cc: OpenStack Development Mailing List; Dolph Mathews; Adam Young Subject: [keystone][horizon]Backend filtering in Keystone Hi Gabriel, Following up on our discussions on filtering and pagination, here's where we stand: 1) We have a patch ready to go into H that implements a framework that lets the keystone backend drivers implement filters (e.g. would be included in the SQL SELECT rather than being a post-prcessed filter on the full list, which is what happens today). See: https://review.openstack.org/#/c/43257/ . It includes the SQL drivers fixed up so they work with this, although it's unlikely we can get the LDAP one complete for H given the freeze (which just means queries to an LDAP backed entity will just work as they do today). 2) The above patch also lets a cloud provider set a limit on the number of rows return by a list query, to avoid excessively long responses and data in the case where the caller doesn't do a good job of filtering. We have two other changes ready, but are deferring to IceHouse: 3) The inexact filtering (e.g. GET /v3/users?name__startswitch=Hen) is coded and included in 1). However, since this is an API change we have it turned off, and will enable early in IceHouse. An API review for this is already posted (with you as one of the reviewers): https://review.openstack.org/#/c/43900/ 4) A separate patch is also ready for Pagination (https://review.openstack.org/#/c/43581/), using the simple page and per_page semantics. Given the contention over this, we'll discuss this at the HK summit I wanted to gauge how advantageous 1) and 2) are to you and the Horizon team? Some concerns have been raised (given how close we are to the freeze) as to whether we should push them in. Personally I'm OK with it, but wanted to balance that with real need (or not if you see these as only minor). Henry ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Horizon] Modal form without redirect
If you look at the code in the post()[1] method of the base workflow view you'll note that a response to a successful workflow POST is always a redirect[2] (caveat for when it's specifically adding data back to a field, which isn't relevant here). The reason for this is that in general when you POST via a standard browser request you want to send back a redirect so that reloading the page, etc. behave correctly and don't potentially result in double-POSTs. If you're submitting the workflow via a regular HTTP from submit POST then I'd say redirecting is correct; you simply want to redirect to the current page. If you're doing this via AJAX then you'll want to add some new code to otherwise signal a successful response (both to the code and to the user) and to take action accordingly. Hope that helps, - Gabriel [1] https://github.com/openstack/horizon/blob/master/horizon/workflows/views.py#L130 [2] https://github.com/openstack/horizon/blob/master/horizon/workflows/views.py#L156 -Original Message- From: Toshiyuki Hayashi [mailto:haya...@ntti3.com] Sent: Tuesday, August 27, 2013 2:26 PM To: OpenStack-dev@lists.openstack.org Subject: [openstack-dev] [Horizon] Modal form without redirect Hi all, I'm working on custmoizing modal form for topology view, I would like to prevent redirecting after submitting. https://github.com/openstack/horizon/blob/master/horizon/static/horizon/ js/horizon.modals.js#L110 According to this code, if there is a no redirect_header, the modal form won't redirect. But I couldn't figure out how to remove redirect information from http header. For example, if I want to remove redirect from LaunchInstance https://github.com/openstack/horizon/blob/master/openstack_dashboard/ dashboards/project/instances/workflows/create_instance.py#L508 How should I do that? I tried success_url = None, but it doesn't work. If you have any idea, that would be great. Regards, Toshiyuki ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [Neutron] Calling empty_chain in L3 agent
Salvatore, Two problems have been found that were caused by calling empty_chain and then failing to restore some rule in the chain that was just emptied. The first problem found was mine in my fix to this bug. https://bugs.launchpad.net/neutron/+bug/1209011 I filed a bug on the second problem today. We discovered it in our development environment yesterday. https://bugs.launchpad.net/neutron/+bug/1218040 Mine broke the gate and got reverted about two days after landing. Now, no one seems to want to touch it with a ten foot pole. It is tainted. :) The second one apparently has gone unnoticed for a while now. I would like to propose a strategy for addressing these problems for now and for the future. First, I propose that we accept the proposed fixes to these two bugs in time for H-3. My patch for the first bug has been up for a while. It is a good fix and I have fixed the problem that caused the gate to break and it has seen more runtime in our test environments. I'd really like to see it land. Second, I would like to discuss a more permanent solution with you since you and I are the authors of the code implicated in the two problems. I think some refactoring and better testing are in order here. The real problem is that in order to empty a chain, there has to be some way to know that we are reconstructing the chain with everything that needs to be in there. Maybe we could get this in for Havana final, maybe post-Havana. What are your thoughts? Regards, Carl Baldwin PS Below are the two reviews that I have to address these two scenarios. https://review.openstack.org/#/c/42412/ https://review.openstack.org/#/c/44133/ ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] multiple-scheduler-drivers blueprint
Why can't something like this be done with just different filters, see such as for AggregateRamFilter? Well, first, at the moment each of these filters today duplicate the code that handles aggregate-based overrides. So, it would make sense to have it in one place anyway. Second, why duplicating all the filters if this can be done with a single flag? Regards, Alex From: Joe Gordon joe.gord...@gmail.com To: OpenStack Development Mailing List openstack-dev@lists.openstack.org, Date: 28/08/2013 09:32 PM Subject:Re: [openstack-dev] [Nova] multiple-scheduler-drivers blueprint On Wed, Aug 28, 2013 at 9:12 AM, Alex Glikson glik...@il.ibm.com wrote: It seems that the main concern was that the overridden scheduler properties are taken from the flavor, and not from the aggregate. In fact, there was a consensus that this is not optimal. I think that we can still make some progress in Havana towards per-aggregate overrides, generalizing on the recently merged changes that do just that -- for cpu and for memory with FilterScheduler (and leveraging a bit from the original multi-sched patch). As follows: 1. individual filters will call get_config('abc') instead of CONF.abc (already implemented in the current version of the multi-sched patch, e.g., https://review.openstack.org/#/c/37407/30/nova/scheduler/filters/io_ops_filter.py ) 2. get_config() will check whether abc is defined in the aggregate, and if so will return the value from the aggregate, and CONF.abc otherwise (already implemented in recently merged AggregateCoreFilter and AggregateRamFilter -- e.g., https://review.openstack.org/#/c/33949/2/nova/scheduler/filters/core_filter.py ). 3. add a global flag that would enable or disable aggregate-based overrides Why can't something like this be done with just different filters, see such as for AggregateRamFilter? This seems to be a relatively simple rafactoring of existing code, still achieving important portion of the original goals of this blueprint. Of course, we should still discuss the longer-term plan around scheduling policies at the summit. Thoughts? Regards, Alex From:Russell Bryant rbry...@redhat.com To:OpenStack Development Mailing List openstack-dev@lists.openstack.org, Date:27/08/2013 10:48 PM Subject:[openstack-dev] [Nova] multiple-scheduler-drivers blueprint Greetings, One of the important things to strive for in our community is consensus. When there's not consensus, we should take a step back and see if we need to change directions. There has been a lot of iterating on this feature, and I'm afraid we still don't have consensus around the design. Phil Day has been posting some really good feedback on the review. I asked Joe Gordon to take a look and provide another opinion. He agreed with Phil that we really need to have scheduler policies be a first class API citizen. So, that pushes this feature out to Icehouse, as it doesn't seem possible to get this done in the required timeframe for Havana. If you'd really like to push to get this into Havana, please make your case. :-) Thanks, -- Russell Bryant ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] multiple-scheduler-drivers blueprint
On Wed, Aug 28, 2013 at 3:55 PM, Alex Glikson glik...@il.ibm.com wrote: Why can't something like this be done with just different filters, see such as for AggregateRamFilter*?* Well, first, at the moment each of these filters today duplicate the code that handles aggregate-based overrides. So, it would make sense to have it in one place anyway. Second, why duplicating all the filters if this can be done with a single flag? * We already have too many flags, and i don't want to introduce one that we plan on removing / deprecating in the near future if we can help it. * https://github.com/openstack/nova/blob/master/nova/scheduler/filters/ram_filter.py doesn't duplicate all the code, it uses a base class. The check the aggregate for the value logic is duplicated, but that is easy to fix. Regards, Alex From:Joe Gordon joe.gord...@gmail.com To:OpenStack Development Mailing List openstack-dev@lists.openstack.org, Date:28/08/2013 09:32 PM Subject:Re: [openstack-dev] [Nova] multiple-scheduler-drivers blueprint -- On Wed, Aug 28, 2013 at 9:12 AM, Alex Glikson *glik...@il.ibm.com*glik...@il.ibm.com wrote: It seems that the main concern was that the overridden scheduler properties are taken from the flavor, and not from the aggregate. In fact, there was a consensus that this is not optimal. I think that we can still make some progress in Havana towards per-aggregate overrides, generalizing on the recently merged changes that do just that -- for cpu and for memory with FilterScheduler (and leveraging a bit from the original multi-sched patch). As follows: 1. individual filters will call get_config('abc') instead of CONF.abc (already implemented in the current version of the multi-sched patch, e.g., https://review.openstack.org/#/c/37407/30/nova/scheduler/filters/core_filter.py * https://review.openstack.org/#/c/37407/30/nova/scheduler/filters/io_ops_filter.py *https://review.openstack.org/#/c/37407/30/nova/scheduler/filters/io_ops_filter.py ) 2. get_config() will check whether abc is defined in the aggregate, and if so will return the value from the aggregate, and CONF.abc otherwise (already implemented in recently merged AggregateCoreFilter and AggregateRamFilter -- e.g., * https://review.openstack.org/#/c/33949/2/nova/scheduler/filters/core_filter.py *https://review.openstack.org/#/c/33949/2/nova/scheduler/filters/core_filter.py ). 3. add a global flag that would enable or disable aggregate-based overrides Why can't something like this be done with just different filters, see such as for AggregateRamFilter*?* This seems to be a relatively simple rafactoring of existing code, still achieving important portion of the original goals of this blueprint. Of course, we should still discuss the longer-term plan around scheduling policies at the summit. Thoughts? Regards, Alex From:Russell Bryant *rbry...@redhat.com* rbry...@redhat.com To:OpenStack Development Mailing List * openstack-dev@lists.openstack.org* openstack-dev@lists.openstack.org, Date:27/08/2013 10:48 PM Subject:[openstack-dev] [Nova] multiple-scheduler-drivers blueprint -- Greetings, One of the important things to strive for in our community is consensus. When there's not consensus, we should take a step back and see if we need to change directions. There has been a lot of iterating on this feature, and I'm afraid we still don't have consensus around the design. Phil Day has been posting some really good feedback on the review. I asked Joe Gordon to take a look and provide another opinion. He agreed with Phil that we really need to have scheduler policies be a first class API citizen. So, that pushes this feature out to Icehouse, as it doesn't seem possible to get this done in the required timeframe for Havana. If you'd really like to push to get this into Havana, please make your case. :-) Thanks, -- Russell Bryant ___ OpenStack-dev mailing list* **OpenStack-dev@lists.openstack.org* OpenStack-dev@lists.openstack.org* **http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev*http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list* **OpenStack-dev@lists.openstack.org* OpenStack-dev@lists.openstack.org* **http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev*http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org
[openstack-dev] [Neutron] The three API server multi-worker process patches.
All, We've known for a while now that some duplication of work happened with respect to adding multiple worker processes to the neutron-server. There were a few mistakes made which led to three patches being done independently of each other. Can we settle on one and accept it? I have changed my patch at the suggestion of one of the other 2 authors, Peter Feiner, in attempt to find common ground. It now uses openstack common code and therefore it is more concise than any of the original three and should be pretty easy to review. I'll admit to some bias toward my own implementation but most importantly, I would like for one of these implementations to land and start seeing broad usage in the community earlier than later. Carl Baldwin PS Here are the two remaining patches. The third has been abandoned. https://review.openstack.org/#/c/37131/ https://review.openstack.org/#/c/36487/ ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] multiple-scheduler-drivers blueprint
Joe Gordon joe.gord...@gmail.com wrote on 28/08/2013 11:04:45 PM: Well, first, at the moment each of these filters today duplicate the code that handles aggregate-based overrides. So, it would make sense to have it in one place anyway. Second, why duplicating all the filters if this can be done with a single flag? We already have too many flags, and i don't want to introduce one that we plan on removing / deprecating in the near future if we can help it. Wouldn't it make sense to have a flag that enables/disables aggregate-based policy overrides anyway? https://github.com/openstack/nova/blob/master/nova/scheduler/ filters/ram_filter.py doesn't duplicate all the code, it uses a base class. The check the aggregate for the value logic is duplicated, but that is easy to fix. Yep, that's exactly what I'm saying -- the first step would be to put that logic in one place (e.g., scheduler/utils.py, like the get_config method we have been thinking to introduce originally), and then we can easily reuse it in all the other filters (regardless of the decision whether to do it within the existing filters, or to add an AggregateXYZ filter for each existing filter XYZ. Same potentially for weight functions, etc). Alex ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] multiple-scheduler-drivers blueprint
On Wed, Aug 28, 2013 at 4:34 PM, Alex Glikson glik...@il.ibm.com wrote: Joe Gordon joe.gord...@gmail.com wrote on 28/08/2013 11:04:45 PM: Well, first, at the moment each of these filters today duplicate the code that handles aggregate-based overrides. So, it would make sense to have it in one place anyway. Second, why duplicating all the filters if this can be done with a single flag? We already have too many flags, and i don't want to introduce one that we plan on removing / deprecating in the near future if we can help it. Wouldn't it make sense to have a flag that enables/disables aggregate-based policy overrides anyway? Why? FWIW that is my default answer to don't we need a flag to do x. https://github.com/openstack/nova/blob/master/nova/scheduler/ filters/ram_filter.py doesn't duplicate all the code, it uses a base class. The check the aggregate for the value logic is duplicated, but that is easy to fix. Yep, that's exactly what I'm saying -- the first step would be to put that logic in one place (e.g., scheduler/utils.py, like the get_config method we have been thinking to introduce originally), and then we can easily reuse it in all the other filters (regardless of the decision whether to do it within the existing filters, or to add an AggregateXYZ filter for each existing filter XYZ. Same potentially for weight functions, etc). Sounds like we are in agreement here Alex ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] About multihost patch review
That said, If there is potential value in offering both, it seems like it should be under the control of the deployer not the user. In other words the deployer should be able to set the default network type and enforce whether setting the type is exposed to the user at all. +1 From my perspective, multi-host is an option that should be in the hands of the deployer/operator, not exposed to the end user. My users should not have to know or care how DHCP, routing and floating IPs are implemented under the covers. Also, last I looked, I think this patch required admin role to create a multi-host network. That may be OK for a single flat network or small-scale multi-network environment, but it is not viable in a large-scale multi-tenant environment. - Jack -Original Message- From: Vishvananda Ishaya [mailto:vishvana...@gmail.com] Sent: Wednesday, August 28, 2013 1:29 PM To: OpenStack Development Mailing List Cc: Robert Kukura; Armando Migliaccio; Nachi Ueno; Sumit Naiksatam Subject: Re: [openstack-dev] About multihost patch review On Aug 26, 2013, at 6:14 PM, Maru Newby ma...@redhat.com wrote: On Aug 26, 2013, at 4:06 PM, Edgar Magana emag...@plumgrid.com wrote: Hi Developers, Let me explain my point of view on this topic and please share your thoughts in order to merge this new feature ASAP. My understanding is that multi-host is nova-network HA and we are implementing this bp https://blueprints.launchpad.net/neutron/+spec/quantum- multihost for the same reason. So, If in neutron configuration admin enables multi-host: etc/dhcp_agent.ini # Support multi host networks # enable_multihost = False Why do tenants needs to be aware of this? They should just create networks in the way they normally do and not by adding the multihost extension. I was pretty confused until I looked at the nova-network HA doc [1]. The proposed design would seem to emulate nova-network's multi-host HA option, where it was necessary to both run nova-network on every compute node and create a network explicitly as multi-host. I'm not sure why nova-network was implemented in this way, since it would appear that multi-host is basically all-or-nothing. Once nova-network services are running on every compute node, what does it mean to create a network that is not multi-host? Just to add a little background to the nova-network multi-host: The fact that the multi_host flag is stored per-network as opposed to a configuration was an implementation detail. While in theory this would support configurations where some networks are multi_host and other ones are not, I am not aware of any deployments where both are used together. That said, If there is potential value in offering both, it seems like it should be under the control of the deployer not the user. In other words the deployer should be able to set the default network type and enforce whether setting the type is exposed to the user at all. Also, one final point. In my mind, multi-host is strictly better than single host, if I were to redesign nova-network today, I would get rid of the single host mode completely. Vish So, to Edgar's question - is there a reason other than 'be like nova-network' for requiring neutron multi-host to be configured per-network? m. 1: http://docs.openstack.org/trunk/openstack-compute/admin/content/existing-ha- networking-options.html I could be totally wrong and crazy, so please provide some feedback. Thanks, Edgar From: Yongsheng Gong gong...@unitedstack.com Date: Monday, August 26, 2013 2:58 PM To: Kyle Mestery (kmestery) kmest...@cisco.com, Aaron Rosen aro...@nicira.com, Armando Migliaccio amigliac...@vmware.com, Akihiro MOTOKI amot...@gmail.com, Edgar Magana emag...@plumgrid.com, Maru Newby ma...@redhat.com, Nachi Ueno na...@nttmcl.com, Salvatore Orlando sorla...@nicira.com, Sumit Naiksatam sumit.naiksa...@bigswitch.com, Mark McClain mark.mccl...@dreamhost.com, Gary Kotton gkot...@vmware.com, Robert Kukura rkuk...@redhat.com Cc: OpenStack List openstack-dev@lists.openstack.org Subject: Re: About multihost patch review Hi, Edgar Magana has commented to say: 'This is the part that for me is confusing and I will need some clarification from the community. Do we expect to have the multi-host feature as an extension or something that will natural work as long as the deployment include more than one Network Node. In my opinion, Neutron deployments with more than one Network Node by default should call DHCP agents in all those nodes without the need to use an extension. If the community has decided to do this by extensions, then I am fine' at https://review.openstack.org/#/c/37919/11/neutron/extensions/multihostnetwork.py I have commented back, what is your opinion about it? Regards, Yong Sheng Gong On Fri, Aug 16, 2013 at 9:28 PM, Kyle Mestery
Re: [openstack-dev] [savanna] migration to pbr completed
On 08/27/2013 04:46 PM, Sergey Lukjanov wrote: Hi folks, migration of all Savanna sub projects to pbr has been completed. Please, inform us and/or create bugs for all packaging-related issues. Thanks. Sincerely yours, Sergey Lukjanov Savanna Technical Lead Mirantis Inc. Thanks for pushing this forward Sergey. I can confirm that pbr 0.5.19 works fine for building savanna, python-savannaclient and savanna-dashboard, with one minor hiccup in savanna that data_files glob'ing doesn't work. It's easily worked around in my spec though (sed -i 's,etc/savanna/\*,etc/savanna/savanna.conf.sample etc/savanna/savanna.conf.sample-full,' setup.cfg), and isn't an issue with 0.5.21. Best, matt ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] Proposal for Raksha, a Data Protection As a Service project
Hello Stackers, We would like to introduce a new project Raksha, a Data Protection As a Service (DPaaS) for OpenStack Cloud. Raksha's primary goal is to provide a comprehensive Data Protection for OpenStack by leveraging Nova, Swift, Glance and Cinder. Raksha has following key features: 1. Provide an enterprise grade data protection for OpenStack based clouds 2. Tenant administered backups and restores 3. Application consistent backups 4. Point In Time(PiT) full and incremental backups and restores 5. Dedupe at source for efficient backups 6. A job scheduler for periodic backups 7. Noninvasive backup solution that does not require service interruption during backup window You will find the rationale behind the need for Raksha in OpenStack in its Wiki. The wiki also has the preliminary design and the API description. Some of the Raksha functionality may overlap with Nova and Cinder projects and as a community lets work together to coordinate the features among these projects. We would like to seek out early feedback so we can address as many issues as we can in the first code drop. We are hoping to enlist the OpenStack community help in making Raksha a part of OpenStack. Raksha's project resources: Wiki: https://wiki.openstack.org/wiki/Raksha Launchpad: https://launchpad.net/raksha Github: https://github.com/DPaaS-Raksha/Raksha (We will upload a prototype code in few days) If you want to talk to us, send an email to openstack-...@lists.launchpad.net with [raksha] in the subject or use #openstack-raksha irc channel. Best Regards, Murali Balcha ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [State-Management] Agenda for tomorrows meeting at 2000 UTC
Hi all, The [state-management] project team holds a weekly meeting in #openstack-meeting on thursdays, 2000 UTC. The next meeting is tomorrow, 2013-08-29!!! As usual, everyone is welcome :-) Link: https://wiki.openstack.org/wiki/Meetings/StateManagement Taskflow: https://wiki.openstack.org/TaskFlow ## Agenda (30-60 mins): - Discuss any action items from last meeting. - Discuss ongoing status of the overall effort and any needed coordination. - Talk about open reviews (please look over https://review.openstack.org/#/q/topic:enp,n,z) - Discuss any integration problems or suggestions. - Discuss about any other potential new use-cases for said library. - Discuss about any other ideas, problems, issues, solutions, questions (and more). Any other topics are welcome :-) See you all soon! -- Joshua Harlow It's openstack, relax... | harlo...@yahoo-inc.com ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] Proposal for Raksha, a Data Protection As a Service project
+1 On Thu, Aug 29, 2013 at 6:12 AM, Murali Balcha murali.bal...@triliodata.com wrote: Hello Stackers, We would like to introduce a new project Raksha, a Data Protection As a Service (DPaaS) for OpenStack Cloud. Raksha’s primary goal is to provide a comprehensive Data Protection for OpenStack by leveraging Nova, Swift, Glance and Cinder. Raksha has following key features: 1. Provide an enterprise grade data protection for OpenStack based clouds 2. Tenant administered backups and restores 3. Application consistent backups 4. Point In Time(PiT) full and incremental backups and restores 5. Dedupe at source for efficient backups 6. A job scheduler for periodic backups 7. Noninvasive backup solution that does not require service interruption during backup window You will find the rationale behind the need for Raksha in OpenStack in its Wiki. The wiki also has the preliminary design and the API description. Some of the Raksha functionality may overlap with Nova and Cinder projects and as a community lets work together to coordinate the features among these projects. We would like to seek out early feedback so we can address as many issues as we can in the first code drop. We are hoping to enlist the OpenStack community help in making Raksha a part of OpenStack. Raksha’s project resources: Wiki: https://wiki.openstack.org/wiki/Raksha Launchpad: https://launchpad.net/raksha Github: https://github.com/DPaaS-Raksha/Raksha (We will upload a prototype code in few days) If you want to talk to us, send an email to openstack-...@lists.launchpad.net with [raksha] in the subject or use #openstack-raksha irc channel. Best Regards, Murali Balcha ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- blog: zqfan.github.com git: github.com/zqfan ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] Proposal for Raksha, a Data Protection As a Service project
Hi Murali, welcome to the OpenStack community. Some comments inline... On 08/28/2013 06:12 PM, Murali Balcha wrote: Hello Stackers, We would like to introduce a new project Raksha, a Data Protection As a Service (DPaaS) for OpenStack Cloud. Raksha’s primary goal is to provide a comprehensive Data Protection for OpenStack by leveraging Nova, Swift, Glance and Cinder. Raksha has following key features: 1.Provide an enterprise grade data protection for OpenStack based clouds What is enterprise grade? Any time I hear that term, I think of Deloitte and Touche salespeople trying to convince some sucker CIO that expensive == good. I'd prefer to just leave the whole enterprise thing for the marketing folks and stick to the specific engineering features ;) 2.Tenant administered backups and restores 3.Application consistent backups Can you expand on this a bit? Data is backed up, not applications... that's what source control is for :) 4.Point In Time(PiT) full and incremental backups and restores Cool, very useful. 5.Dedupe at source for efficient backups Hmmm... this would depend heavily on what is being backed up and the level of access that Raksha would have to the tenant's application domains. Unless you are going to limit yourself to just backing up and restoring instances or volumes? Is that the plan? 6.A job scheduler for periodic backups Cron? 7.Noninvasive backup solution that does not require service interruption during backup window By service, are you referring to the tenant's applications running on the instance? Or are you referring to something else? Also, one thing that is really good to expose/debate/discuss early on in the project's incubation is the RESTful API that the project would expose. I'd be really interested to see this. Finally, would be good to include in the wiki page some discussion about any interaction with Trove (DBaaS), especially since Trove's API already implements a backups/ resource [1]. Best, -jay [1] https://github.com/openstack/database-api/blob/master/openstack-database-api/src/markdown/database-api-v1.md#backups You will find the rationale behind the need for Raksha in OpenStack in its Wiki. The wiki also has the preliminary design and the API description.Some of the Raksha functionality may overlap with Nova and Cinder projects and as a community lets work together to coordinate the features among these projects. We would like to seek out early feedback so we can address as many issues as we can in the first code drop. We are hoping to enlist the OpenStack community help in making Raksha a part of OpenStack. Raksha’s project resources: Wiki: https://wiki.openstack.org/wiki/Raksha Launchpad: https://launchpad.net/raksha Github: https://github.com/DPaaS-Raksha/Raksha (We will upload a prototype code in few days) If you want to talk to us, send an email to openstack-...@lists.launchpad.net with [raksha] in the subject or use #openstack-raksha irc channel. Best Regards, Murali Balcha ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [neutron] Why does nova.network.neutronv2.get_client(context, admin=True) drop auth_token?
Hi stackers! Sorry for the stupid question, but why does nova.network.neutronv2.get_client() [1] drop auth_token for admin? Is it really necessary to make another check for username/password when trying to get a list of ports or floating IPs?.. When keystone is configured with LDAP backed this leads to a bunch of LDAP requests which tend to be quite slow. Plus those LDAP requests could have been simply skipped when keystone is configured with token cache enabled. Thanks, Roman [1] https://github.com/openstack/nova/blob/master/nova/network/neutronv2/__init__.py#L68 ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] Proposal for Raksha, a Data Protection As a Service project
Hi Murali, welcome to the OpenStack community. Some comments inline... Hi Jay, Thanks for your comments. I have been involved with OpenStack since Diablo time frame, but mostly monitoring the traffic. Always great to be part of the community. My comments are inline. On 08/28/2013 06:12 PM, Murali Balcha wrote: Hello Stackers, We would like to introduce a new project Raksha, a Data Protection As a Service (DPaaS) for OpenStack Cloud. Raksha’s primary goal is to provide a comprehensive Data Protection for OpenStack by leveraging Nova, Swift, Glance and Cinder. Raksha has following key features: 1.Provide an enterprise grade data protection for OpenStack based clouds What is enterprise grade? Any time I hear that term, I think of Deloitte and Touche salespeople trying to convince some sucker CIO that expensive == good. I'd prefer to just leave the whole enterprise thing for the marketing folks and stick to the specific engineering features ;) I hear you. Will take out the market jargon and stick to technical specifications. 2.Tenant administered backups and restores 3.Application consistent backups Can you expand on this a bit? Data is backed up, not applications... that's what source control is for :) I meant a VM and all its associated resources such as data volumes are backed up consistently. And if the application is a multi tier application that is spread across multiple vms, then all the vms should be backed up together consistently too. 4.Point In Time(PiT) full and incremental backups and restores Cool, very useful. 5.Dedupe at source for efficient backups Hmmm... this would depend heavily on what is being backed up and the level of access that Raksha would have to the tenant's application domains. Unless you are going to limit yourself to just backing up and restoring instances or volumes? Is that the plan? Yes, that is the plan. We don't have plans to look into vms and backup any individual resources with in a vm. 6.A job scheduler for periodic backups Cron? Yes, similar. 7.Noninvasive backup solution that does not require service interruption during backup window By service, are you referring to the tenant's applications running on the instance? Or are you referring to something else? A service similar to either nova or cinder. We don't want to run any agents inside tenant vms. Also, one thing that is really good to expose/debate/discuss early on in the project's incubation is the RESTful API that the project would expose. I'd be really interested to see this. Absolutely. We took a first stab of the list restful API that raksha need to support in our wiki @ http://wiki.openstack.org/wiki/raksha. Finally, would be good to include in the wiki page some discussion about any interaction with Trove (DBaaS), especially since Trove's API already implements a backups/ resource [1]. That is a good point. we will include that in the wiki. Best, -jay [1] https://github.com/openstack/database-api/blob/master/openstack-database-api/src/markdown/database-api-v1.md#backups You will find the rationale behind the need for Raksha in OpenStack in its Wiki. The wiki also has the preliminary design and the API description.Some of the Raksha functionality may overlap with Nova and Cinder projects and as a community lets work together to coordinate the features among these projects. We would like to seek out early feedback so we can address as many issues as we can in the first code drop. We are hoping to enlist the OpenStack community help in making Raksha a part of OpenStack. Raksha’s project resources: Wiki: https://wiki.openstack.org/wiki/Raksha Launchpad: https://launchpad.net/raksha Github: https://github.com/DPaaS-Raksha/Raksha (We will upload a prototype code in few days) If you want to talk to us, send an email to openstack-...@lists.launchpad.net with [raksha] in the subject or use #openstack-raksha irc channel. Best Regards, Murali Balcha ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [neutron] Why does nova.network.neutronv2.get_client(context, admin=True) drop auth_token?
For admin, we must use admin token. In general, the token from API context is not of role admin. I think the BP can help https://blueprints.launchpad.net/keystone/+spec/reuse-token On Thu, Aug 29, 2013 at 8:12 AM, Roman Verchikov rverchi...@mirantis.comwrote: Hi stackers! Sorry for the stupid question, but why does nova.network.neutronv2.get_client() [1] drop auth_token for admin? Is it really necessary to make another check for username/password when trying to get a list of ports or floating IPs?.. When keystone is configured with LDAP backed this leads to a bunch of LDAP requests which tend to be quite slow. Plus those LDAP requests could have been simply skipped when keystone is configured with token cache enabled. Thanks, Roman [1] https://github.com/openstack/nova/blob/master/nova/network/neutronv2/__init__.py#L68 ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] MultiStrOpt opts can be problematic...
Ha! My team totally ran into the same issue. I was hoping that Padraig's crudini would make my dreams come true but I don't think it's handling multi-string options. https://github.com/pixelb/crudini So +1 to getting rid of those. Thanks, MATT RIEDEMANN Advisory Software Engineer Cloud Solutions and OpenStack Development Phone: 1-507-253-7622 | Mobile: 1-507-990-1889 E-mail: mrie...@us.ibm.com 3605 Hwy 52 N Rochester, MN 55901-1407 United States From: Dan Prince dpri...@redhat.com To: OpenStack Development Mailing List openstack-dev@lists.openstack.org, Date: 08/28/2013 12:54 PM Subject:[openstack-dev] MultiStrOpt opts can be problematic... So I recently ran into a fun config issue in trying to configure Nova to work w/ Ceilometer using Puppet: https://bugs.launchpad.net/puppet-ceilometer/+bug/1217867 Today, what you need to do to make Nova work with ceilometer is add this to your nova.conf file: notification_driver=nova.openstack.common.notifier.rpc_notifier notification_driver=ceilometer.compute.nova_notifier As it turns out Multi-line config entries aren't very fun to deal with in the config management world. The puppet nova_config provider doesn't (yet) have a good way to support them. The core of the issue is they pose all sorts of problems in knowing whether a given tool should modify the existing config values. In the short term we can look into doing one of these Puppet land: -Using a conf.d directory for config (would require a change to the nova-compute init script to use --config-dir) -String together various resources in puppet to make it work, etc. (file_line, augeus, etc) Long term though I'm thinking what if MultiStrOpt's were to go away? They seem to be more trouble than they are worth... Dan ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev image/gif___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [openstack]ASK: Summit talk on OpenStack on OpenStack/ OpenStack as a Service
Dear Stackers, I came across a proposed talk related to OpenStack on OpenStack/ OpenStack as a Service. Couldn't recollect if there were two different talks or the same. I am interested in learning more - could the speakers ping me offline please? -- Thanks, -Sriram ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] Glance revert to fix image deletes on F19
Dan, I saw you abandoned the patch because: Using Sqlalchemy 0.8.2 or downgrading to 0.7.10 seems to fix this issue. Given the glance requirement on SQLAlchemy: https://github.com/openstack/glance/blob/master/requirements.txt#L9 Why were you at those levels to begin with? Thanks, MATT RIEDEMANN Advisory Software Engineer Cloud Solutions and OpenStack Development Phone: 1-507-253-7622 | Mobile: 1-507-990-1889 E-mail: mrie...@us.ibm.com 3605 Hwy 52 N Rochester, MN 55901-1407 United States From: Dan Prince dpri...@redhat.com To: OpenStack Development Mailing List openstack-dev@lists.openstack.org, Date: 08/26/2013 07:21 AM Subject:[openstack-dev] Glance revert to fix image deletes on F19 Hi all, I'd like to highlight this Glance revert for an issue I started seeing on Fedora 19 last week. https://review.openstack.org/#/c/43542/ A revert like this should be a very safe thing to do so long as we do it quickly. Especially given this is a performance improvement... Dan ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev image/gif___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] MultiStrOpt opts can be problematic...
[/me grabs popcorn and pulls up a chair to watch...] On Wed, Aug 28, 2013 at 7:40 PM, Matt Riedemann mrie...@us.ibm.com wrote: Ha! My team totally ran into the same issue. I was hoping that Padraig's crudini would make my dreams come true but I don't think it's handling multi-string options. *https://github.com/pixelb/crudini* https://github.com/pixelb/crudini So +1 to getting rid of those. Ditto When we added support for these to DevStack it was for exactly those options Dan mentioned, and that is still the only place they are used... dt -- Dean Troyer dtro...@gmail.com ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [neutron] Why does nova.network.neutronv2.get_client(context, admin=True) drop auth_token?
On Wed, Aug 28, 2013 at 7:22 PM, Yongsheng Gong gong...@unitedstack.comwrote: For admin, we must use admin token. In general, the token from API context is not of role admin. So... because the authenticated user making the API request *may not* have admin access, you're dropping that authorization in favor of using CONF.neutron_admin_username, etc, to escalate the available privileges? Yikes. I think the BP can help https://blueprints.launchpad.net/keystone/+spec/reuse-token I don't see how? On Thu, Aug 29, 2013 at 8:12 AM, Roman Verchikov rverchi...@mirantis.comwrote: Hi stackers! Sorry for the stupid question, but why does nova.network.neutronv2.get_client() [1] drop auth_token for admin? Is it really necessary to make another check for username/password when trying to get a list of ports or floating IPs?.. When keystone is configured with LDAP backed this leads to a bunch of LDAP requests which tend to be quite slow. Plus those LDAP requests could have been simply skipped when keystone is configured with token cache enabled. Thanks, Roman [1] https://github.com/openstack/nova/blob/master/nova/network/neutronv2/__init__.py#L68 ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- -Dolph ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [savanna] cluster scaling on the 0.2 branch
Hi, I am trying to back port the HDP scaling implementation to the 0.2 branch and have run into a number of differences. At this point I am trying to figure out whether what I am observing is intended or symptoms of a bug. For a case in which I am adding one instance to an existing node group as well as an additional node group with one instance I am seeing the following arguments being passed to the scale_cluster method of the plugin: - A cluster object that contains the following set of node groups: [savanna.db.models.NodeGroup[object at 10d8bdd90] {created=datetime.datetime(2013, 8, 28, 21, 50, 5, 208003), updated=datetime.datetime(2013, 8, 28, 21, 50, 5, 208007), id=u'd6fadb7a-367b-41ed-989c-af40af2d3e3d', name=u'master', flavor_id=u'3', image_id=None, node_processes=[u'NAMENODE', u'SECONDARY_NAMENODE', u'GANGLIA_SERVER', u'GANGLIA_MONITOR', u'AMBARI_SERVER', u'AMBARI_AGENT', u'JOBTRACKER', u'NAGIOS_SERVER'], node_configs={}, volumes_per_node=0, volumes_size=10, volume_mount_prefix=u'/volumes/disk', count=1, cluster_id=u'e086d444-2a0f-4105-8ef2-51c56cdb70d2', node_group_template_id=u'15344a5c-5e83-496a-9648-d7b58f40ad1f'}, savanna.db.models.NodeGroup[object at 10d8bd950] {created=datetime.datetime(2013, 8, 28, 21, 50, 5, 210962), updated=datetime.datetime(2013, 8, 28, 22, 5, 1, 728402), id=u'672e5597-2a8d-4470-8f5d-8cc43c7bb28e', name=u'slave', flavor_id=u'3', image_id=None, node_processes=[u'DATANODE', u'HDFS_CLIENT', u'GANGLIA_MONITOR', u'AMBARI_AGENT', u'TASKTRACKER', u'MAPREDUCE_CLIENT'], node_configs={}, volumes_per_node=0, volumes_size=10, volume_mount_prefix=u'/volumes/disk', count=2, cluster_id=u'e086d444-2a0f-4105-8ef2-51c56cdb70d2', node_group_template_id=u'5dd6aa5a-496c-4dda-b94c-3b3752eb0efb'}, savanna.db.models.NodeGroup[object at 10d897f90] {created=datetime.datetime(2013, 8, 28, 22, 4, 59, 871379), updated=datetime.datetime(2013, 8, 28, 22, 4, 59, 871388), id=u'880e1b17-f4e4-456d-8421-31bf8ef1fb65', name=u'slave2', flavor_id=u'1', image_id=None, node_processes=[u'DATANODE', u'HDFS_CLIENT', u'GANGLIA_MONITOR', u'AMBARI_AGENT', u'TASKTRACKER', u'MAPREDUCE_CLIENT'], node_configs={}, volumes_per_node=0, volumes_size=10, volume_mount_prefix=u'/volumes/disk', count=1, cluster_id=u'e086d444-2a0f-4105-8ef2-51c56cdb70d2', node_group_template_id=u'd67da924-792b-4558-a5cb-cb97bba4107f'}] So it appears that the cluster is already configured with the three node groups (two original, one new) and the associated counts. - The list of instances. However, whereas the master branch was passing me two instances (one instance representing the addition to the existing group, one representing the new instance associated with the added node group), in the 0.2 branch I am only seeing one instance being passed (the one instance being added to the existing node group): (Pdb) p instances [savanna.db.models.Instance[object at 10d8bf050] {created=datetime.datetime(2013, 8, 28, 22, 5, 1, 725343), updated=datetime.datetime(2013, 8, 28, 22, 5, 47, 286665), extra=None, node_group_id=u'672e5597-2a8d-4470-8f5d-8cc43c7bb28e', instance_id=u'377694a2-a589-479b-860f-f1541d249624', instance_name=u'scale-slave-002', internal_ip=u'192.168.32.4', management_ip=u'172.18.3.5', volumes=[]}] (Pdb) p len(instances) 1 I am not certain why I am not getting a listing of instances representing the instances being added to the cluster as I do in the master branch. Is this intended? How do I obtain the instance reference for the instance being added to the new node group? -- Jon -- CONFIDENTIALITY NOTICE NOTICE: This message is intended for the use of the individual or entity to which it is addressed and may contain information that is confidential, privileged and exempt from disclosure under applicable law. If the reader of this message is not the intended recipient, you are hereby notified that any printing, copying, dissemination, distribution, disclosure or forwarding of this communication is strictly prohibited. If you have received this communication in error, please contact the sender immediately and delete it from your system. Thank You. ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [neutron] Why does nova.network.neutronv2.get_client(context, admin=True) drop auth_token?
On Wed, Aug 28, 2013 at 5:22 PM, Yongsheng Gong gong...@unitedstack.comwrote: For admin, we must use admin token. In general, the token from API context is not of role admin. If this functionality is supposed to be allowed to non-admin users, wouldn't it be easier to provide access to it to non-admin users, instead of escalating permissions (maybe RBAC)? I'll admit not knowing why this needs escalation, but it stands out as an odd approach in my mind. I think the BP can help https://blueprints.launchpad.net/keystone/+spec/reuse-token This isn't likely what you are looking for. It would still require lookups to the backend for a number of reasons (not listed, as I don't think it is relevant for this conversation). -- Morgan Fainberg IRC: morganfainberg ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [Ceilometer] what's in scope of Ceilometer
so we're in the process of selling Ceilometer to product teams so that they'll adopt it and we'll get more funding :). one item that comes up from product teams is 'what will Ceilometer be able to do and where does the product takeover and add value?' the first question is, Ceilometer currently does metering/alarming/maybe a few other things... will it go beyond that? specifically: capacity planning, optimization, dashboard(i assume this falls under horizon/ceilometer plugin work), analytics. they're pretty broad items so i would think they would probably end up being separate projects? another question is what metrics will we capture. some of the product teams we have collect metrics on datacenter memory/cpu utilization, cluster cpu/memory/vm, and a bunch of other clustered stuff. i'm a nova-idiot, but is this info possible to retrieve? is the consensus that Ceilometer will collect anything and everything the other projects allow for? cheers, gordon chung openstack, ibm software standards email: chungg [at] ca.ibm.com phone: 905.413.5072___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On Wed, 28 Aug 2013 09:58:48 -0400 Joe Gordon joe.gord...@gmail.com wrote: On a related note, I really like when the developer adds a gerrit comment saying why the revision, that makes my life as a reviewer easier. +1 - I try to remember to do this and from a reviewer point of view this is especially useful when there has been a rebase involved. Chris ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Neutron] Resource URL support for more than two levels
Hi, When compared to Nova URL implementations, It is observed that the Neutron URL support cannot be used for more than TWO levels. Applications which want to add as PLUG-IN may be restricted with this. We want to add support for changes required for supporting more than TWO Levels of URL by adding the support changes required in Core Neutron Files. Any comments/interest in this.? Regards, Balaji.P On Tue, Aug 27, 2013 at 5:04 PM, B Veera-B37207 b37...@freescale.comwrote: Hi, ** ** The current infrastructure provided in Quantum [Grizzly], while building Quantum API resource URL using the base function ‘base.create_resource()’ and RESOURCE_ATTRIBUTE_MAP/SUB_RESOURCE_ATTRIBUTE_MAP, supports only two level URI. Example: GET /lb/pools/pool_id/members/member_id ** ** Some applications may need more than two levels of URL support. Example: GET /lb/pools/pool_id/members/member_id/xyz/xyz_id ** ** If anybody is interested in this, we want to contribute for this as BP and make it upstream. ** ** Regards, Veera. ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] Frustrations with review wait times
On Wed, 28 Aug 2013 15:56:33 + Joshua Harlow harlo...@yahoo-inc.com wrote: Shrinking that rotation granularity would be reasonable to. Rotate once every 2 weeks or some other time period still seems useful to me. I wonder if the quality of reviewing would drop if someone was doing it all day long though. IIRC the link that Robert pointed to in another thread seemed to indicate that the ability for someone to pick up bugs reduces significantly if they are doing code reviews continuously. Chris ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] New Nova hypervisor: Docker
On Wed, Aug 28, 2013 at 9:12 AM, Sam Alba sam.a...@gmail.com wrote: Thanks a lot everyone for the nice feedback. I am going to work hard to get all those new comments addressed to be able to re-submit a new patchset today or tomorrow (the later). On Wed, Aug 28, 2013 at 7:02 AM, Russell Bryant rbry...@redhat.com wrote: On 08/28/2013 05:18 AM, Daniel P. Berrange wrote: On Wed, Aug 28, 2013 at 06:00:50PM +1000, Michael Still wrote: On Wed, Aug 28, 2013 at 4:18 AM, Sam Alba sam.a...@gmail.com wrote: Hi all, We've been working hard during the last couple of weeks with some people. Brian Waldon helped a lot designing the Glance integration and driver testing. Dean Troyer helped a lot on bringing Docker support in Devstack[1]. On top of that, we got several feedback on the Nova code review which definitely helped to improve the code. The blueprint[2] explains what Docker brings to Nova and how to use it. I have to say that this blueprint is a fantastic example of how we should be writing design documents. It addressed almost all of my questions about the integration. Yes, Sam ( any of the other Docker guys involved) have been great at responding to reviewers' requests to expand their design document. The latest update has really helped in understanding how this driver works in the context of openstack from an architectural and functional POV. They've been great in responding to my requests, as well. The biggest thing was that I wanted to see devstack support so that it's easily testable, both by developers and by CI. They delivered. So, in general, I'm good with this going in. It's just a matter of getting the code review completed in the next week before feature freeze. I'm going to try to help with it this week. If someone wants to take another look at https://review.openstack.org/#/c/32960/, we answered/fixed all previous comments. -- @sam_alba ___ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev