| *** sudipto has joined #openstack-dib | 03:00 | |
| *** sudipto_ has joined #openstack-dib | 03:00 | |
| *** brault has quit IRC | 03:04 | |
| openstackgerrit | Merged openstack/diskimage-builder: elements: Drop executable bits from environment files https://review.openstack.org/387441 | 03:18 |
|---|---|---|
| openstackgerrit | Merged openstack/diskimage-builder: Delete deprecated Hacking in tox.ini https://review.openstack.org/406555 | 03:19 |
| *** sudipto_ has quit IRC | 04:25 | |
| *** sudipto has quit IRC | 04:25 | |
| *** mordred has quit IRC | 04:26 | |
| *** greghaynes has quit IRC | 04:26 | |
| *** greghaynes has joined #openstack-dib | 05:03 | |
| *** mordred has joined #openstack-dib | 05:04 | |
| *** mordred has quit IRC | 05:18 | |
| *** sudipto has joined #openstack-dib | 05:18 | |
| *** sudipto_ has joined #openstack-dib | 05:18 | |
| *** mat128 has quit IRC | 05:19 | |
| *** mat128 has joined #openstack-dib | 05:23 | |
| *** mordred has joined #openstack-dib | 05:27 | |
| *** chhavi has joined #openstack-dib | 05:27 | |
| *** chhavi has quit IRC | 05:41 | |
| *** chhavi has joined #openstack-dib | 05:48 | |
| *** sudipto_ has quit IRC | 06:28 | |
| *** sudipto has quit IRC | 06:28 | |
| openstackgerrit | Noam Angel proposed openstack/diskimage-builder: redhat-common add option to select networking service NetworkManager/network https://review.openstack.org/392170 | 06:44 |
| *** sudipto has joined #openstack-dib | 06:51 | |
| *** sudipto_ has joined #openstack-dib | 06:52 | |
| *** mat128 has quit IRC | 07:22 | |
| *** mat128 has joined #openstack-dib | 07:23 | |
| *** mordred has quit IRC | 07:25 | |
| *** mordred has joined #openstack-dib | 07:34 | |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder: Preserve backwards compat for expand_dependencies https://review.openstack.org/407386 | 08:05 |
| *** brault has joined #openstack-dib | 08:13 | |
| *** chhavi has quit IRC | 08:45 | |
| *** strigazi_AFK is now known as strigazi | 09:28 | |
| *** sudipto_ has quit IRC | 09:47 | |
| *** sudipto has quit IRC | 09:47 | |
| *** chhavi has joined #openstack-dib | 10:32 | |
| *** jaianshu has joined #openstack-dib | 10:38 | |
| *** chhavi has quit IRC | 11:08 | |
| *** owalsh has joined #openstack-dib | 11:13 | |
| *** chhavi has joined #openstack-dib | 11:22 | |
| *** jaianshu_ has joined #openstack-dib | 11:33 | |
| *** jaianshu has quit IRC | 11:36 | |
| *** jaianshu_ has quit IRC | 12:49 | |
| *** pmannidi has quit IRC | 13:27 | |
| *** chhavi has quit IRC | 13:37 | |
| *** sudipto has joined #openstack-dib | 14:59 | |
| *** sudipto_ has joined #openstack-dib | 14:59 | |
| openstackgerrit | Antoine Musso proposed openstack/diskimage-builder: debootstrap: avoid duplicate network configuration https://review.openstack.org/407551 | 15:00 |
| openstackgerrit | Antoine Musso proposed openstack/diskimage-builder: debootstrap: avoid duplicate network configuration https://review.openstack.org/407551 | 15:05 |
| *** hashar has joined #openstack-dib | 15:07 | |
| johnsom | Hi DIB folks. Can we get some reviews on https://review.openstack.org/#/c/406413/ and https://review.openstack.org/#/c/406420/ ? | 15:16 |
| johnsom | These combined fix an issue impacting the Octavia project scenario tests / image builds. Thanks! | 15:16 |
| *** owalsh has quit IRC | 16:14 | |
| *** dmarlin has joined #openstack-dib | 16:28 | |
| *** chhavi has joined #openstack-dib | 17:14 | |
| *** sudipto_ has quit IRC | 17:32 | |
| *** sudipto has quit IRC | 17:32 | |
| *** hashar is now known as hasharEat | 18:02 | |
| *** chhavi has quit IRC | 18:32 | |
| *** hasharEat is now known as hashar | 19:05 | |
| *** pleia2 has quit IRC | 19:34 | |
| *** pleia2 has joined #openstack-dib | 19:34 | |
| openstackgerrit | greghaynes proposed openstack/diskimage-builder: Fix pip-and-virtualenv to work with python3 https://review.openstack.org/407723 | 21:51 |
| openstackgerrit | Andrey Shestakov proposed openstack/diskimage-builder: Fix dhcp-all-interfaces for ubuntu-minimal xenial https://review.openstack.org/407725 | 22:03 |
| greghaynes | ianw: I replied to you on https://review.openstack.org/#/c/406413/2 - its messy but thats the best I could come up with | 22:04 |
| openstackgerrit | Andrey Shestakov proposed openstack/diskimage-builder: Fix dhcp-all-interfaces for ubuntu-minimal xenial https://review.openstack.org/407725 | 22:05 |
| ianw | johnsom greghaynes : so it does get left in the image (sysctl-write-value) at /usr/local/bin as a tool for later potentional use, right? | 22:12 |
| greghaynes | yes | 22:13 |
| greghaynes | the issue is - the logic bug means its broken | 22:13 |
| ianw | oh, i agree it's not safe to use during the build. it's just super confusing to me if a host built with tripleo elements has the same script but it does different things | 22:14 |
| johnsom | I agree, it is confusing. | 22:15 |
| greghaynes | actually, are we sure it gets left in the image? | 22:15 |
| greghaynes | I thought it did, but now I cant remember if bin/ automagically does that | 22:15 |
| johnsom | I guess I missed that everything in bin just gets picked up. The tripleo element has an install.d so I figure that is where it was getting installed | 22:15 |
| ianw | if it's in bin/ then it gets copied to /usr/local/bin by install-bin | 22:15 |
| greghaynes | ok | 22:16 |
| ianw | the other idea, if we can be more radical, is to support elements having their own sysctl.conf snippets | 22:16 |
| ianw | and making this element merge them into the final sysctl.conf | 22:16 |
| greghaynes | oh definitely, this is doing it the old, wrong way | 22:17 |
| greghaynes | I just want to try and back us out of the backwards compat mess somehow :/ | 22:17 |
| ianw | that way, nothing gets left on the host, and we don't try to apply settings to the build kernel | 22:17 |
| johnsom | Yeah, that makes sense. How fast can we pull that off though? Right now our gates ground to a halt after that os- change last week | 22:18 |
| greghaynes | johnsom: for your gates - you all just need to switch from sysctl-set-value to sysctl-write-value, right? | 22:19 |
| ianw | johnsom: what's calling "-apply" during the build? | 22:19 |
| ianw | sorry, "set-value" (which does the apply) | 22:19 |
| ianw | yeah, i think same point as greghaynes | 22:19 |
| johnsom | I guess I could, but would lose the check for duplicate settings. | 22:20 |
| johnsom | duplicate/conflicting | 22:21 |
| greghaynes | ok. So maybe the fix is we deprecate this whole element, make a new element (dib-sysctl) and add a dib-sysctl-set-value there which DTRT. Eventually when we do it 'the right way' it can live in that element and we can deprecte the dib-sysctl-set-value | 22:22 |
| greghaynes | and the reason for deprecating the whole element is its the only way we can really give a proper deprecation warning (we dont know if someone will use the tripleo or dib version of the element) | 22:22 |
| ianw | why don't we add the duplicate setting check to the write-value command? | 22:24 |
| ianw | i think that was the prior changeset? | 22:24 |
| greghaynes | I think latest does it, it just also changes -set. The only issue with that is the potential deal where youre using the tripleo version | 22:24 |
| greghaynes | if were fine with that potential bug though I think thats ok | 22:24 |
| greghaynes | actually, yea - the check shouldnt ever hurt, it just adds a safeguard | 22:25 |
| ianw | yeah, if we leave the "set" command alone, it's compatible? | 22:26 |
| greghaynes | ok, now I want to write the element to do this correctly so we can get rid of this | 22:26 |
| greghaynes | or if we just remove it | 22:26 |
| greghaynes | oh I see what youre saying, if we make one that errors its wierd | 22:26 |
| greghaynes | ok yea, just leave it | 22:27 |
| greghaynes | johnsom: sorry for the roundabout. I think leave the -set for now (maybe with a comment about how this needs to die) and then keep you -write fix | 22:27 |
| johnsom | Ok | 22:27 |
| ianw | yep, and johnsom's build needs to use "-write" ... not "-set" ... that's a bug that should be mentioned in the README.rst | 22:28 |
| ianw | v2 would be nice here :) makes it much easier to walk the element list and cat the files together | 22:29 |
| greghaynes | definitely | 22:29 |
| johnsom | I think we should keep the seperate sysctl files and not try to mess with sysctl.conf | 22:30 |
| ianw | oh yeah, that's even easier just dumping the files into sysctl.d | 22:30 |
| greghaynes | johnsom: agreed - I think the plan is more to let elements write out sysctl.conf snippets then have an element which copies them all in place | 22:30 |
| johnsom | +1 | 22:31 |
| openstackgerrit | Ian Wienand proposed openstack/diskimage-builder: Copy sysctl.d snippets https://review.openstack.org/407739 | 22:40 |
| ianw | greghaynes johnsom : ^ version 0.1 :) | 22:40 |
| *** hashar has quit IRC | 22:46 | |
| johnsom | greghaynes Looks like something is missing the python-dev package as a dependency: https://review.openstack.org/#/c/407723/ | 22:51 |
| johnsom | ianw Mind if I take that and add the conflict check? | 22:52 |
| *** yolanda has quit IRC | 22:52 | |
| *** yolanda has joined #openstack-dib | 22:54 | |
| johnsom | I would just do a simple one for now, just aborting on duplicate filenames, but I guess that isn't so safe given people can call the files whatever they want this way | 22:54 |
| ianw | johnsom: no, of course, edit away! :) | 22:57 |
| openstackgerrit | Michael Johnson proposed openstack/diskimage-builder: Update sysctl-write-value to do conflict checking https://review.openstack.org/406413 | 23:00 |
| johnsom | That one just updates the write to do conflict checking | 23:00 |
| *** yolanda has quit IRC | 23:04 | |
| greghaynes | johnsom: oh thats fun | 23:11 |
| *** yolanda has joined #openstack-dib | 23:13 | |
| *** pmannidi has joined #openstack-dib | 23:13 | |
| *** yolanda has quit IRC | 23:22 | |
| openstackgerrit | greghaynes proposed openstack/diskimage-builder: Fix pip-and-virtualenv to work with python3 https://review.openstack.org/407723 | 23:26 |
| openstackgerrit | greghaynes proposed openstack/diskimage-builder: Fix pip-and-virtualenv to work with python3 https://review.openstack.org/407723 | 23:30 |
| *** yolanda has joined #openstack-dib | 23:35 | |
| openstackgerrit | greghaynes proposed openstack/diskimage-builder: Fix pip-and-virtualenv to work with python3 https://review.openstack.org/407723 | 23:38 |
| *** dmarlin has quit IRC | 23:39 | |
| *** yolanda has quit IRC | 23:43 | |
| openstackgerrit | Michael Johnson proposed openstack/diskimage-builder: Copy sysctl.d snippets https://review.openstack.org/407739 | 23:44 |
| johnsom | ianw What do you think about that? ^^^ Do I need an awk dependency? | 23:45 |
| greghaynes | wow thats gnarly | 23:45 |
| greghaynes | fwiw its trivial to use python in place of bash if that ends up being easier | 23:46 |
| johnsom | Too gnarly? | 23:46 |
| greghaynes | uhh, i dont think so | 23:46 |
| greghaynes | but yes you should add an awk dep | 23:46 |
| johnsom | Basically it's "skip comments, make sure the line has an =, then look for a duplicate and exit if you find one" | 23:46 |
| greghaynes | maybe a comment about that too | 23:47 |
| ianw | johnsom: not bad, you could split it over lines to make it a little clearer. i'd also like to see it exit with some sort of sensible message describing the issue | 23:48 |
| johnsom | I guess it has a bug in that if the duplicate has an identical value it will exit | 23:49 |
| ianw | i think that's probably ok ... if two things are setting the same value (even to the same thing) it seems to indicate they should get together and sort it out for real | 23:50 |
| ianw | https://git.openstack.org/cgit/openstack-dev/devstack/tree/inc/meta-config#n97 are some gnarly awk's that are at least a little readable | 23:50 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!