Re: [openstack-dev] [Fuel] Abandon changesets which hang for a while without updates
Folks, I'll push DevOps to run the script. However, what we need is to just go ahead and clean up, abandon manually what is not relevant anymore, provide comment. Please start with your patches. On Thu, Jul 16, 2015 at 11:08 PM Oleg Gelbukh ogelb...@mirantis.com wrote: Nicely put, Doug, you gave me laughs :) I can't see how a CR could hang for a month without anyone paying attention if it worths merging. If this really happens (which I'm not aware of), auto-abandon definitely won't make things any worse. -- Best regards, Oleg Gelbukh On Fri, Jul 17, 2015 at 6:10 AM, Doug Wiegley doug...@parksidesoftware.com wrote: Just adding an experience from another project, Neutron. We had similar debates, and prepping for the long apocalyptic winter of changeset death, Kyle decimated the world and ran the abandon script. The debates were far more intense than the reality, and my large stockpile of Rad-X and Nuka Cola went to waste. Every few weeks, I get a few emails of things being abandoned. And if I care about something, mine or not, I click through and tap ‘Restore’. If one person in the entire community can’t be bothered to click one button, I’m not sure how it’d ever be kept up-to-date, much less merge. Thanks, doug On Jul 16, 2015, at 8:36 PM, Dmitry Borodaenko dborodae...@mirantis.com wrote: I'm with Stanislaw on this one: abandoning reviews just to make numbers *look* better will accomplish nothing. The only benefit I can see is cleaning up reviews that we *know* don't need to be considered, so that it's easier for reviewers to find the reviews that still need attention. I don't see this as that much of a problem, finding stuff to review in Fuel Review Inbox [0] is not hard at all. [0] https://wiki.openstack.org/wiki/Fuel#Development_related_links And the state of our review backlog is such that it's not safe to auto-abandon reviews without looking at them, and if a contributor has spent time looking at a review, abandoning it manually is one click away. If we do go with setting up an auto-abandon rule, it should be extremely conservative, for example: CR has a negative vote from a core reviewer AND there were no comments or positive votes from anyone after that AND it has not been touched in any way for 2 months. On Wed, Jul 15, 2015 at 5:10 PM Mike Scherbakov mscherba...@mirantis.com wrote: Folks, let's execute here. Numbers are still large. Did we have a chance to look over the whole queue? Can we go ahead and abandon changes having -1 or -2 from reviewers for over than a months or so? I'm all for just following standard OpenStack process [1], and then change it only if there is good reason for it. [1] https://wiki.openstack.org/wiki/Puppet#Patch_abandonment_policy On Thu, Jul 9, 2015 at 6:27 PM Stanislaw Bogatkin sbogat...@mirantis.com wrote: 2 weeks seems too small for me. We easy can be in situation when fix for medium bug is done, but SCF starts. And gap between SCF and release easily can be more than a month. So, 2 months seems okay for me if speaking about forcibly applying auto-abandon by major vote. And I'm personally against such innovation at all. On Thu, Jul 9, 2015 at 5:37 PM, Davanum Srinivas dava...@gmail.com wrote: That's a very good plan (Initial feedback/triage) Mike. thanks, dims On Thu, Jul 9, 2015 at 3:23 PM, Mike Scherbakov mscherba...@mirantis.com wrote: +1 for just reusing existing script, and adjust it on the way. No need to immediately switch from infinite time to a couple of weeks, we can always adjust it later. But 1-2 month should be a good start already. Our current stats [1] look just terrible. Before we enable an auto-abandon, we need to go every single patch first, and review it / provide comment to authors. The idea is not to abandon some good patches, and not to offend contributors... Let's think how we can approach it. Should we have core reviewers to check their corresponding components? [1] http://stackalytics.com/report/reviews/fuel-group/open On Wed, Jul 8, 2015 at 1:13 PM Sean M. Collins s...@coreitpro.com wrote: Let's keep it at 4 weeks without comment, and Jenkins failed - similar to the script that Kyle Mestery uses for Neutron. In fact, we could actually just use his script ;) https://github.com/openstack/neutron/blob/master/tools/abandon_old_reviews.sh -- Sean M. Collins __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://openstack-dev-requ...@lists.openstack.org/?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe:
Re: [openstack-dev] [Fuel] Abandon changesets which hang for a while without updates
Nicely put, Doug, you gave me laughs :) I can't see how a CR could hang for a month without anyone paying attention if it worths merging. If this really happens (which I'm not aware of), auto-abandon definitely won't make things any worse. -- Best regards, Oleg Gelbukh On Fri, Jul 17, 2015 at 6:10 AM, Doug Wiegley doug...@parksidesoftware.com wrote: Just adding an experience from another project, Neutron. We had similar debates, and prepping for the long apocalyptic winter of changeset death, Kyle decimated the world and ran the abandon script. The debates were far more intense than the reality, and my large stockpile of Rad-X and Nuka Cola went to waste. Every few weeks, I get a few emails of things being abandoned. And if I care about something, mine or not, I click through and tap ‘Restore’. If one person in the entire community can’t be bothered to click one button, I’m not sure how it’d ever be kept up-to-date, much less merge. Thanks, doug On Jul 16, 2015, at 8:36 PM, Dmitry Borodaenko dborodae...@mirantis.com wrote: I'm with Stanislaw on this one: abandoning reviews just to make numbers *look* better will accomplish nothing. The only benefit I can see is cleaning up reviews that we *know* don't need to be considered, so that it's easier for reviewers to find the reviews that still need attention. I don't see this as that much of a problem, finding stuff to review in Fuel Review Inbox [0] is not hard at all. [0] https://wiki.openstack.org/wiki/Fuel#Development_related_links And the state of our review backlog is such that it's not safe to auto-abandon reviews without looking at them, and if a contributor has spent time looking at a review, abandoning it manually is one click away. If we do go with setting up an auto-abandon rule, it should be extremely conservative, for example: CR has a negative vote from a core reviewer AND there were no comments or positive votes from anyone after that AND it has not been touched in any way for 2 months. On Wed, Jul 15, 2015 at 5:10 PM Mike Scherbakov mscherba...@mirantis.com wrote: Folks, let's execute here. Numbers are still large. Did we have a chance to look over the whole queue? Can we go ahead and abandon changes having -1 or -2 from reviewers for over than a months or so? I'm all for just following standard OpenStack process [1], and then change it only if there is good reason for it. [1] https://wiki.openstack.org/wiki/Puppet#Patch_abandonment_policy On Thu, Jul 9, 2015 at 6:27 PM Stanislaw Bogatkin sbogat...@mirantis.com wrote: 2 weeks seems too small for me. We easy can be in situation when fix for medium bug is done, but SCF starts. And gap between SCF and release easily can be more than a month. So, 2 months seems okay for me if speaking about forcibly applying auto-abandon by major vote. And I'm personally against such innovation at all. On Thu, Jul 9, 2015 at 5:37 PM, Davanum Srinivas dava...@gmail.com wrote: That's a very good plan (Initial feedback/triage) Mike. thanks, dims On Thu, Jul 9, 2015 at 3:23 PM, Mike Scherbakov mscherba...@mirantis.com wrote: +1 for just reusing existing script, and adjust it on the way. No need to immediately switch from infinite time to a couple of weeks, we can always adjust it later. But 1-2 month should be a good start already. Our current stats [1] look just terrible. Before we enable an auto-abandon, we need to go every single patch first, and review it / provide comment to authors. The idea is not to abandon some good patches, and not to offend contributors... Let's think how we can approach it. Should we have core reviewers to check their corresponding components? [1] http://stackalytics.com/report/reviews/fuel-group/open On Wed, Jul 8, 2015 at 1:13 PM Sean M. Collins s...@coreitpro.com wrote: Let's keep it at 4 weeks without comment, and Jenkins failed - similar to the script that Kyle Mestery uses for Neutron. In fact, we could actually just use his script ;) https://github.com/openstack/neutron/blob/master/tools/abandon_old_reviews.sh -- Sean M. Collins __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://openstack-dev-requ...@lists.openstack.org/?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://openstack-dev-requ...@lists.openstack.org/?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Davanum Srinivas :: https://twitter.com/dims
Re: [openstack-dev] [Fuel] Abandon changesets which hang for a while without updates
I'm with Stanislaw on this one: abandoning reviews just to make numbers *look* better will accomplish nothing. The only benefit I can see is cleaning up reviews that we *know* don't need to be considered, so that it's easier for reviewers to find the reviews that still need attention. I don't see this as that much of a problem, finding stuff to review in Fuel Review Inbox [0] is not hard at all. [0] https://wiki.openstack.org/wiki/Fuel#Development_related_links And the state of our review backlog is such that it's not safe to auto-abandon reviews without looking at them, and if a contributor has spent time looking at a review, abandoning it manually is one click away. If we do go with setting up an auto-abandon rule, it should be extremely conservative, for example: CR has a negative vote from a core reviewer AND there were no comments or positive votes from anyone after that AND it has not been touched in any way for 2 months. On Wed, Jul 15, 2015 at 5:10 PM Mike Scherbakov mscherba...@mirantis.com wrote: Folks, let's execute here. Numbers are still large. Did we have a chance to look over the whole queue? Can we go ahead and abandon changes having -1 or -2 from reviewers for over than a months or so? I'm all for just following standard OpenStack process [1], and then change it only if there is good reason for it. [1] https://wiki.openstack.org/wiki/Puppet#Patch_abandonment_policy On Thu, Jul 9, 2015 at 6:27 PM Stanislaw Bogatkin sbogat...@mirantis.com wrote: 2 weeks seems too small for me. We easy can be in situation when fix for medium bug is done, but SCF starts. And gap between SCF and release easily can be more than a month. So, 2 months seems okay for me if speaking about forcibly applying auto-abandon by major vote. And I'm personally against such innovation at all. On Thu, Jul 9, 2015 at 5:37 PM, Davanum Srinivas dava...@gmail.com wrote: That's a very good plan (Initial feedback/triage) Mike. thanks, dims On Thu, Jul 9, 2015 at 3:23 PM, Mike Scherbakov mscherba...@mirantis.com wrote: +1 for just reusing existing script, and adjust it on the way. No need to immediately switch from infinite time to a couple of weeks, we can always adjust it later. But 1-2 month should be a good start already. Our current stats [1] look just terrible. Before we enable an auto-abandon, we need to go every single patch first, and review it / provide comment to authors. The idea is not to abandon some good patches, and not to offend contributors... Let's think how we can approach it. Should we have core reviewers to check their corresponding components? [1] http://stackalytics.com/report/reviews/fuel-group/open On Wed, Jul 8, 2015 at 1:13 PM Sean M. Collins s...@coreitpro.com wrote: Let's keep it at 4 weeks without comment, and Jenkins failed - similar to the script that Kyle Mestery uses for Neutron. In fact, we could actually just use his script ;) https://github.com/openstack/neutron/blob/master/tools/abandon_old_reviews.sh -- Sean M. Collins __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Davanum Srinivas :: https://twitter.com/dims __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Fuel] Abandon changesets which hang for a while without updates
Folks, let's execute here. Numbers are still large. Did we have a chance to look over the whole queue? Can we go ahead and abandon changes having -1 or -2 from reviewers for over than a months or so? I'm all for just following standard OpenStack process [1], and then change it only if there is good reason for it. [1] https://wiki.openstack.org/wiki/Puppet#Patch_abandonment_policy On Thu, Jul 9, 2015 at 6:27 PM Stanislaw Bogatkin sbogat...@mirantis.com wrote: 2 weeks seems too small for me. We easy can be in situation when fix for medium bug is done, but SCF starts. And gap between SCF and release easily can be more than a month. So, 2 months seems okay for me if speaking about forcibly applying auto-abandon by major vote. And I'm personally against such innovation at all. On Thu, Jul 9, 2015 at 5:37 PM, Davanum Srinivas dava...@gmail.com wrote: That's a very good plan (Initial feedback/triage) Mike. thanks, dims On Thu, Jul 9, 2015 at 3:23 PM, Mike Scherbakov mscherba...@mirantis.com wrote: +1 for just reusing existing script, and adjust it on the way. No need to immediately switch from infinite time to a couple of weeks, we can always adjust it later. But 1-2 month should be a good start already. Our current stats [1] look just terrible. Before we enable an auto-abandon, we need to go every single patch first, and review it / provide comment to authors. The idea is not to abandon some good patches, and not to offend contributors... Let's think how we can approach it. Should we have core reviewers to check their corresponding components? [1] http://stackalytics.com/report/reviews/fuel-group/open On Wed, Jul 8, 2015 at 1:13 PM Sean M. Collins s...@coreitpro.com wrote: Let's keep it at 4 weeks without comment, and Jenkins failed - similar to the script that Kyle Mestery uses for Neutron. In fact, we could actually just use his script ;) https://github.com/openstack/neutron/blob/master/tools/abandon_old_reviews.sh -- Sean M. Collins __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Davanum Srinivas :: https://twitter.com/dims __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Fuel] Abandon changesets which hang for a while without updates
2 weeks seems too small for me. We easy can be in situation when fix for medium bug is done, but SCF starts. And gap between SCF and release easily can be more than a month. So, 2 months seems okay for me if speaking about forcibly applying auto-abandon by major vote. And I'm personally against such innovation at all. On Thu, Jul 9, 2015 at 5:37 PM, Davanum Srinivas dava...@gmail.com wrote: That's a very good plan (Initial feedback/triage) Mike. thanks, dims On Thu, Jul 9, 2015 at 3:23 PM, Mike Scherbakov mscherba...@mirantis.com wrote: +1 for just reusing existing script, and adjust it on the way. No need to immediately switch from infinite time to a couple of weeks, we can always adjust it later. But 1-2 month should be a good start already. Our current stats [1] look just terrible. Before we enable an auto-abandon, we need to go every single patch first, and review it / provide comment to authors. The idea is not to abandon some good patches, and not to offend contributors... Let's think how we can approach it. Should we have core reviewers to check their corresponding components? [1] http://stackalytics.com/report/reviews/fuel-group/open On Wed, Jul 8, 2015 at 1:13 PM Sean M. Collins s...@coreitpro.com wrote: Let's keep it at 4 weeks without comment, and Jenkins failed - similar to the script that Kyle Mestery uses for Neutron. In fact, we could actually just use his script ;) https://github.com/openstack/neutron/blob/master/tools/abandon_old_reviews.sh -- Sean M. Collins __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Davanum Srinivas :: https://twitter.com/dims __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Fuel] Abandon changesets which hang for a while without updates
That's a very good plan (Initial feedback/triage) Mike. thanks, dims On Thu, Jul 9, 2015 at 3:23 PM, Mike Scherbakov mscherba...@mirantis.com wrote: +1 for just reusing existing script, and adjust it on the way. No need to immediately switch from infinite time to a couple of weeks, we can always adjust it later. But 1-2 month should be a good start already. Our current stats [1] look just terrible. Before we enable an auto-abandon, we need to go every single patch first, and review it / provide comment to authors. The idea is not to abandon some good patches, and not to offend contributors... Let's think how we can approach it. Should we have core reviewers to check their corresponding components? [1] http://stackalytics.com/report/reviews/fuel-group/open On Wed, Jul 8, 2015 at 1:13 PM Sean M. Collins s...@coreitpro.com wrote: Let's keep it at 4 weeks without comment, and Jenkins failed - similar to the script that Kyle Mestery uses for Neutron. In fact, we could actually just use his script ;) https://github.com/openstack/neutron/blob/master/tools/abandon_old_reviews.sh -- Sean M. Collins __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Davanum Srinivas :: https://twitter.com/dims __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Fuel] Abandon changesets which hang for a while without updates
+1 for just reusing existing script, and adjust it on the way. No need to immediately switch from infinite time to a couple of weeks, we can always adjust it later. But 1-2 month should be a good start already. Our current stats [1] look just terrible. Before we enable an auto-abandon, we need to go every single patch first, and review it / provide comment to authors. The idea is not to abandon some good patches, and not to offend contributors... Let's think how we can approach it. Should we have core reviewers to check their corresponding components? [1] http://stackalytics.com/report/reviews/fuel-group/open On Wed, Jul 8, 2015 at 1:13 PM Sean M. Collins s...@coreitpro.com wrote: Let's keep it at 4 weeks without comment, and Jenkins failed - similar to the script that Kyle Mestery uses for Neutron. In fact, we could actually just use his script ;) https://github.com/openstack/neutron/blob/master/tools/abandon_old_reviews.sh -- Sean M. Collins __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Fuel] Abandon changesets which hang for a while without updates
+1 for enabling auto-abandon Sergii, I think the period should be smaller, 2 weeks - 1 month, if patch is important, the author will unabandon it. Thanks, On Wed, Jul 8, 2015 at 3:50 AM, Sergii Golovatiuk sgolovat...@mirantis.com wrote: 3 Months should be a good period. Sometimes, people have one month vacation thus it might be a problem in such cases. -- Best regards, Sergii Golovatiuk, Skype #golserge IRC #holser On Tue, Jul 7, 2015 at 3:34 PM, Mike Scherbakov mscherba...@mirantis.com wrote: Hi, we have too many patches which hang for over a month without attention. Almost all are those which should be abandoned due to changes happened in master, and need complete rework. Or, those were just proposals to make something, which didn't fly. Unfortunately there are those which were good ideas, but didn't attract enough attention from cores and didn't land into master as a result. Anyway, I think we need to keep our review queues in better health. Can we enable auto-abandon back again, if patch is hanging out for a month or more? -- Mike Scherbakov #mihgen __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev