| *** rcernin_ has joined #openstack-ironic | 00:02 | |
| *** rcernin has quit IRC | 00:03 | |
| *** mjturek has quit IRC | 00:12 | |
| *** mbeierl has joined #openstack-ironic | 00:15 | |
| *** namnh has quit IRC | 00:17 | |
| *** rcernin has joined #openstack-ironic | 00:29 | |
| *** rcernin has quit IRC | 00:29 | |
| *** rcernin has joined #openstack-ironic | 00:30 | |
| *** rcernin_ has quit IRC | 00:32 | |
| *** rh-jelabarre has quit IRC | 00:57 | |
| *** marlinc has quit IRC | 00:57 | |
| *** ddurst has quit IRC | 00:57 | |
| *** gmann has quit IRC | 00:57 | |
| *** lennyb has quit IRC | 00:57 | |
| *** rh-jelabarre has joined #openstack-ironic | 00:59 | |
| *** namnh has joined #openstack-ironic | 00:59 | |
| *** marlinc has joined #openstack-ironic | 01:04 | |
| *** ddurst has joined #openstack-ironic | 01:04 | |
| *** gmann has joined #openstack-ironic | 01:04 | |
| *** namnh has quit IRC | 01:04 | |
| *** openstackgerrit has quit IRC | 01:06 | |
| *** openstackgerrit has joined #openstack-ironic | 01:20 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic master: Disable periodic tasks if interval set to 0 https://review.openstack.org/569166 | 01:20 |
|---|---|---|
| *** phuongnh has joined #openstack-ironic | 01:23 | |
| *** rloo has quit IRC | 01:31 | |
| *** namnh has joined #openstack-ironic | 01:44 | |
| *** jiapei has joined #openstack-ironic | 01:49 | |
| openstackgerrit | yuan liang proposed openstack/ironic-python-agent master: Parallel erase disk devices https://review.openstack.org/346723 | 01:54 |
| *** markguz_ has joined #openstack-ironic | 02:19 | |
| *** rh-jelabarre has quit IRC | 02:30 | |
| *** Bhujay has joined #openstack-ironic | 03:04 | |
| *** markguz_ has quit IRC | 03:05 | |
| openstackgerrit | yuan liang proposed openstack/ironic-python-agent master: Parallel erase disk devices https://review.openstack.org/346723 | 03:25 |
| *** Bhujay has quit IRC | 03:40 | |
| *** namnh has quit IRC | 03:40 | |
| *** jiapei has quit IRC | 03:59 | |
| *** Bhujay has joined #openstack-ironic | 04:21 | |
| *** gyee has quit IRC | 05:05 | |
| *** yolanda has quit IRC | 05:25 | |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic master: Prevents deletion of ports for active nodes https://review.openstack.org/470273 | 05:58 |
| *** yolanda has joined #openstack-ironic | 06:17 | |
| *** phuongnh has quit IRC | 06:29 | |
| *** pcaruana has joined #openstack-ironic | 06:44 | |
| openstackgerrit | yolanda.robla proposed openstack/ironic master: Add functionality for individual cleanup on nodes https://review.openstack.org/586277 | 06:50 |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic-inspector master: Multiple backends for introspection data: database https://review.openstack.org/583930 | 06:52 |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic-inspector master: Multiple backends for introspection data: processing https://review.openstack.org/514552 | 06:59 |
| *** rcernin has quit IRC | 07:02 | |
| *** e0ne has joined #openstack-ironic | 07:18 | |
| yolanda | hi, silly question, what's the alternative for "ironic node-set-provsion-state deleted", using openstack baremetal api? i try to just delete a node on deploy failed with the openstack baremetal cli, but i can't figure how to move from deploying to deleted state | 07:31 |
| yolanda | when i try the barmetal node delete there, i just get "cannot delete node while it is in state deploy failed" | 07:33 |
| *** marlinc has quit IRC | 07:39 | |
| *** olivierbourdon38 has joined #openstack-ironic | 07:39 | |
| etingof | yolanda, may be you can try turning it into 'manage' first? | 07:43 |
| etingof | yolanda, no, according to the state diagram you should be able to delete the node from the 'deploy failed' state -- https://docs.openstack.org/ironic/pike/_images/states.svg | 07:45 |
| etingof | yolanda, wait, if you are in 'deploying' state, you probably have to wait till it fails and transitions into 'deploy failed' state from where you should be able to delete the node | 07:46 |
| *** marlinc has joined #openstack-ironic | 07:48 | |
| *** w-miller has joined #openstack-ironic | 07:49 | |
| yolanda | it is in deploy failed, but i cannot delete when it's there | 07:54 |
| *** tssurya has joined #openstack-ironic | 07:54 | |
| yolanda | mm again it may be because of an old ironic version? i shoud try upgrading to queens | 07:55 |
| *** Bhujay has quit IRC | 08:10 | |
| *** ajya has joined #openstack-ironic | 08:11 | |
| *** openstackstatus has quit IRC | 08:12 | |
| *** dtantsur|afk is now known as dtantsur | 08:15 | |
| dtantsur | morning ironic | 08:15 |
| dtantsur | TheJulia, rpioso|afk, jroll, my take on this is that's it's a bit risky patch due to its size | 08:16 |
| dtantsur | and while Ocata won't be automatically closed this months, maintaining the CI for it will be increasingly harder | 08:17 |
| dtantsur | e.g. I'm not sure devstack will support Ocata after that point | 08:17 |
| dtantsur | so if we screw up now, we may not be able to fix it. | 08:17 |
| *** serlex has joined #openstack-ironic | 08:17 | |
| *** derekh has joined #openstack-ironic | 08:29 | |
| *** Bhujay has joined #openstack-ironic | 08:34 | |
| dtantsur | rpioso|afk: btw, we'll need a rocky release for dracclient. lemme know if you need help with cutting it. | 08:42 |
| etingof | morning dtantsur \o | 08:44 |
| *** MattMan_1 has quit IRC | 08:49 | |
| *** MattMan_1 has joined #openstack-ironic | 08:49 | |
| dtantsur | o/ | 08:53 |
| *** mgoddard has joined #openstack-ironic | 09:07 | |
| *** etingof is now known as etingof|brb | 09:11 | |
| *** gkadam has joined #openstack-ironic | 09:18 | |
| *** ajya has quit IRC | 09:29 | |
| *** ajya has joined #openstack-ironic | 09:29 | |
| *** openstackstatus has joined #openstack-ironic | 09:41 | |
| *** ChanServ sets mode: +v openstackstatus | 09:41 | |
| *** hoonetorg has quit IRC | 09:43 | |
| dtantsur | yolanda: the counterpart of "deleted" is "baremetal node undeploy". | 09:48 |
| dtantsur | ("deleted" is a very confusing name indeed) | 09:48 |
| yolanda | undeploy? | 09:48 |
| yolanda | first time i hear! | 09:48 |
| yolanda | dtantsur, i guess documentation needs a refresh | 09:48 |
| dtantsur | probably? where did you spot the old command? | 09:49 |
| yolanda | in queens doc | 09:49 |
| yolanda | dtantsur, https://docs.openstack.org/ironic/queens/contributor/states.html | 09:50 |
| yolanda | no undeploy there | 09:50 |
| yolanda | when you reach deploy failed, it mentions to use the deleted verb, but this didn't work for me | 09:51 |
| yolanda | it was just a fatal failure, so it was just on deploy failed forever | 09:52 |
| *** hoonetorg has joined #openstack-ironic | 09:57 | |
| dtantsur | yolanda: you're confusing states with CLI commands | 09:57 |
| dtantsur | yolanda: https://docs.openstack.org/python-ironicclient/latest/cli/osc/v1/index.html#baremetal-node-undeploy | 09:58 |
| yolanda | mm, it should be useful to mention this on the states doc. The equivalent between verb and the cli command | 09:59 |
| yolanda | some of them are mostly equivalente | 09:59 |
| yolanda | equivalent | 09:59 |
| yolanda | but for this one, for sure i didn't get it | 09:59 |
| yolanda | i mean, if you want to set the rebuild verb you have the rebuild cli, same for manage... but in this specific undeploy case it's not as obvious | 10:01 |
| *** S4ren has joined #openstack-ironic | 10:03 | |
| dtantsur | well, it's downside of having "deleted", which we cannot change without breaking the API | 10:08 |
| openstackgerrit | Aija Jaunteva proposed openstack/sushy master: Update to use mapped field for Message Severity field https://review.openstack.org/591612 | 10:09 |
| yolanda | dtantsur, i mean, a documentation change there could help to understand | 10:15 |
| dtantsur | agreed | 10:21 |
| yolanda | i can propose a patch. I was explaining last week about that diagram, and i missed that undeploy action as well, i'm pretty sure lots of people don't know about it. With the previous node-set provision state api, it was more clear | 10:34 |
| yolanda | not that i like, but the command was more obvious for that case | 10:34 |
| *** ajya has quit IRC | 10:40 | |
| *** ajya has joined #openstack-ironic | 10:40 | |
| dtantsur | yeah, but it was obvious if and only if you were familiar with the state machine action | 10:41 |
| dtantsur | otherwise a lot of people confused 'node-set-provision-state deleted' and 'node-delete' | 10:41 |
| *** etingof|brb is now known as etingof | 10:53 | |
| *** ajya has quit IRC | 10:59 | |
| *** ajya has joined #openstack-ironic | 11:00 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironicclient master: Minor fixes to README.rst https://review.openstack.org/591624 | 11:10 |
| jroll | morning | 11:17 |
| jroll | dtantsur: thanks, so the patch is the problem, not us maintaining ocata, right? | 11:17 |
| jroll | and from my POV if our CI hits issues that are difficult to fix, we remove the jobs and it's up to consumers to test | 11:18 |
| *** serlex has quit IRC | 11:24 | |
| *** ajya3 has joined #openstack-ironic | 11:31 | |
| *** ajya has quit IRC | 11:31 | |
| *** rh-jelabarre has joined #openstack-ironic | 11:34 | |
| openstackgerrit | Aija Jaunteva proposed openstack/sushy master: WIP: Add @Redfish.Settings update status and expose it for BIOS https://review.openstack.org/591635 | 11:49 |
| *** slagle has joined #openstack-ironic | 11:53 | |
| *** serlex has joined #openstack-ironic | 12:04 | |
| dtantsur | jroll: morning. the problem is, if we allow a regression now, it will get into our very last release of ocata. | 12:05 |
| dtantsur | any later fixes will not be released | 12:05 |
| jroll | dtantsur: right | 12:05 |
| jroll | I just wanted to be clear that you were not saying we (upstream ironic people) should not maintain ocata | 12:05 |
| dtantsur | which is why I'd personally prefer to stay very conservative until that point, then allow the blocked fixes to land. | 12:05 |
| jroll | ++ | 12:05 |
| dtantsur | I'm fine with maintaining ocata | 12:06 |
| jroll | I figured. just checking. :) | 12:06 |
| dtantsur | just noting that if/when devstack drops ocata support, we're only left with unit tests | 12:06 |
| jroll | rpioso|afk: ^ clarity for you | 12:06 |
| jroll | yep | 12:06 |
| *** bfournie has quit IRC | 12:15 | |
| *** ajya3 has quit IRC | 12:19 | |
| *** ajya has joined #openstack-ironic | 12:20 | |
| *** ajya has quit IRC | 12:22 | |
| *** ajya3 has joined #openstack-ironic | 12:22 | |
| *** S4ren has quit IRC | 12:24 | |
| *** S4ren has joined #openstack-ironic | 12:29 | |
| *** trown|outtypewww is now known as trown | 12:32 | |
| openstackgerrit | Olivier Bourdon proposed openstack/bifrost master: Fix deployment when improper version of python-openssl is installed https://review.openstack.org/591646 | 12:33 |
| olivierbourdon38 | Hello ironicers, could you please have a look at https://gist.github.com/obourdon/104ca78baab77fb5ac753ec3e0a9f105 which states the problem I am trying to solve in https://review.openstack.org/591646 and tell me your thoughts about this ? | 12:33 |
| patchbot | patch 591646 - bifrost - Fix deployment when improper version of python-ope... | 12:33 |
| olivierbourdon38 | I have been testing this patch for the last few days successfully | 12:34 |
| olivierbourdon38 | this should also remind you of an issue I reported earlier last month when the pymysql package version was updated in OpenStack requirements which lead to a somewhat but less nasty/imbricated issue which was solved by forcing the use pf using when deploying Bifrost in my case. | 12:36 |
| olivierbourdon38 | However this was not sufficient this time | 12:36 |
| olivierbourdon38 | many thanks in advance for your time, advices and comments | 12:37 |
| *** rnoriega has quit IRC | 12:39 | |
| TheJulia | olivierbourdon38: seems like the python-openssl must be pulling in something else and possibly squashing pip since your using a dpkg and the package managers are going to have no clue about additional packages installed via pip, so then you end up in a situation where it looks like pip can't even start up :( | 12:49 |
| olivierbourdon38 | @TheJulia from what I experienced installing only python-openssl package in addition to a working platform does not drag any other dependencies | 12:50 |
| olivierbourdon38 | however the fact that shade is installed in /usr/local and that the urllib3 is using pyopenssl 18.0.0 version (and not 15.0.0 like in xenial base system) is the issue | 12:52 |
| *** bfournie has joined #openstack-ironic | 12:52 | |
| olivierbourdon38 | the patchset I submitted (even if still incomplete if I refer to the current Zuul CI errors) is adressing these | 12:52 |
| olivierbourdon38 | currently forging patchset 2 | 12:52 |
| arne_wiebalck | :q | 12:53 |
| arne_wiebalck | (sorry, wrong window) | 12:53 |
| TheJulia | olivierbourdon38: so that is a conundrum since shade is needed where the ansible plugins execute :\ | 12:54 |
| TheJulia | well, now openstacksdk | 12:54 |
| olivierbourdon38 | conundrum ? | 12:55 |
| olivierbourdon38 | git show | 12:56 |
| TheJulia | well, not a conundrum in the sense that something happened 7 days ago | 12:56 |
| TheJulia | But more so the additional python code was required to be present for the module to be able to execute via a local connection because it is performed outside of any venv. :\ | 12:57 |
| TheJulia | or is it the remote install case where that was needed | 12:58 |
| TheJulia | gah, better inline comments | 12:58 |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironic-inspector-client master: CI: stop trying to pull the tempest plugin from ironic-inspector https://review.openstack.org/591659 | 13:03 |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironic-inspector-client master: CI: stop trying to pull the tempest plugin from ironic-inspector https://review.openstack.org/591659 | 13:04 |
| openstackgerrit | Olivier Bourdon proposed openstack/bifrost master: Fix deployment when improper version of python-openssl is installed https://review.openstack.org/591646 | 13:05 |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironic-inspector-client stable/queens: CI: stop trying to pull the tempest plugin from ironic-inspector https://review.openstack.org/591661 | 13:05 |
| dtantsur | le sigh ^^^ | 13:05 |
| TheJulia | :( | 13:06 |
| *** Bhujay has quit IRC | 13:07 | |
| *** Bhujay has joined #openstack-ironic | 13:07 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironic-inspector-client stable/rocky: Update .gitreview for stable/rocky https://review.openstack.org/585995 | 13:09 |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironic-inspector-client master: CI: stop trying to pull the tempest plugin from ironic-inspector https://review.openstack.org/591659 | 13:10 |
| dtantsur | omg, I think this is finally done for master - queens | 13:10 |
| *** rloo has joined #openstack-ironic | 13:15 | |
| *** jcoufal has joined #openstack-ironic | 13:23 | |
| *** Bhujay has quit IRC | 13:40 | |
| openstackgerrit | Merged openstack/bifrost master: Add doc/requirements.txt to releasenotes tox environment https://review.openstack.org/591416 | 13:44 |
| *** cdearborn has joined #openstack-ironic | 13:53 | |
| *** mjturek has joined #openstack-ironic | 13:54 | |
| openstackgerrit | Derek Higgins proposed openstack/ironic master: Switch ironic-dsvm-functional to zuulv3 https://review.openstack.org/591675 | 13:55 |
| openstackgerrit | Derek Higgins proposed openstack/ironic master: Switch ironic-tempest-dsvm-parallel to zuulv3 https://review.openstack.org/591676 | 13:55 |
| *** baha has joined #openstack-ironic | 14:03 | |
| *** brtknr has quit IRC | 14:05 | |
| *** baha_ has joined #openstack-ironic | 14:05 | |
| *** baha has quit IRC | 14:09 | |
| *** olivierbourdon38 has quit IRC | 14:09 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/bifrost master: Allow extra packages to be installed with ironic https://review.openstack.org/591417 | 14:14 |
| mordred | TheJulia: reading scrollback | 14:25 |
| mordred | oh, yeah. that | 14:26 |
| * mordred just slinks out the back | 14:26 | |
| openstackgerrit | Derek Higgins proposed openstack/ironic master: Switch ironic-tempest-dsvm-parallel to zuulv3 https://review.openstack.org/591676 | 14:27 |
| pabelanger | morning, could I get a +3 on refactoring of bifrost tests https://review.openstack.org/591471/ hoping to remove legacy-opensuse-423 nodeset from openstack-infra | 14:29 |
| patchbot | patch 591471 - bifrost - Refactor bifrost-integration-dhcp jobs - 0h 16m 0s spent in CI | 14:29 |
| dtantsur | in addition to ^^, can I get some reviews on https://review.openstack.org/#/c/591659/ please, it fixes the CI | 14:33 |
| patchbot | patch 591659 - python-ironic-inspector-client - CI: stop trying to pull the tempest plugin from ir... - 0h 52m 44s spent in CI | 14:33 |
| S4ren | Hi guys, I am trying to write a script using the ironic-pythonclient. After running 'session.node.list()[0]' I get a node record, but when I try to access the uuid value by using 'session.node.list()[0]['uuid'] I get an exception saying the Node object is not subscriptable. How do I access the values inside this object? | 14:34 |
| dtantsur | S4ren: via dot, i.e. .uuid | 14:34 |
| S4ren | That makes a lot of sense, thanks dtantsur | 14:35 |
| *** baha_ is now known as baha | 14:37 | |
| *** olivierbourdon38 has joined #openstack-ironic | 14:40 | |
| *** olivierbourdon38 has quit IRC | 14:43 | |
| *** Bhujay has joined #openstack-ironic | 14:52 | |
| *** ajya3 has quit IRC | 15:06 | |
| *** ajya3 has joined #openstack-ironic | 15:06 | |
| *** rpioso|afk is now known as rpioso | 15:07 | |
| rpioso | Good morning | 15:07 |
| *** ajya3 has quit IRC | 15:10 | |
| *** ajya has joined #openstack-ironic | 15:11 | |
| etingof | good morning rpioso \o | 15:14 |
| rpioso | etingof: o/ | 15:14 |
| TheJulia | good morning rpioso | 15:15 |
| *** e0ne has quit IRC | 15:16 | |
| rpioso | TheJulia: Morning :) | 15:19 |
| *** e0ne has joined #openstack-ironic | 15:20 | |
| * rpioso catches up with the backscroll | 15:22 | |
| S4ren | Pardon dtantsur, I am looking through the documentation but I cant seem to find the appropriate command to change a nodes name | 15:28 |
| dtantsur | morning rpioso | 15:28 |
| S4ren | There doesnt seem to be anything referring to that in node.py | 15:28 |
| dtantsur | S4ren: via Python or via CLI? | 15:28 |
| S4ren | Python | 15:28 |
| dtantsur | all fields are changed via node.update | 15:29 |
| dtantsur | the format is a bit fancy, it's JSON patch | 15:29 |
| dtantsur | so, something like ironic.node.update(node_id, [{'path': '/name', 'op': 'replace', 'value': <new name>}]) | 15:29 |
| S4ren | Ah I see, and if I needed to add ipmi credentials to driver_info, would it be ironic.node.update(node_id, [{'path': '/name', 'op': 'add', 'value': {u'deploy_kernel':'http;//1.1.1.1:8080', 'ipmi_address':'2.2.2.2'} etc...])? | 15:32 |
| TheJulia | well path of /driver_info, but I think you can add each value on at a time as /driver_info/deploy_kernel instead of pre-formatting a dictionary | 15:35 |
| S4ren | let me give that a go, thank you TheJulia and dtantsur | 15:36 |
| dtantsur | S4ren: feel free to adapt bits from https://github.com/openstack/metalsmith to your needs. it does not have node enrolling, but does show case deployment and the overall approach. | 15:39 |
| arne_wiebalck | Hi Ironic! When looking for a root device, the lsblk command run by the IPA explicitly excludes dependent devices (such as md) ... does anyone know the reason why? | 15:41 |
| S4ren | Thanks dtantsur, will have a look at that repo now | 15:41 |
| rpioso | dtantsur: :) | 15:43 |
| dtantsur | arne_wiebalck: not sure, but I suspect we're only interested in actual hardware | 15:44 |
| rpioso | dtantsur, jroll: Thank you for the clarification. | 15:46 |
| arne_wiebalck | dtantsur: Hmm. I’d think installing on top of a s/W RAID device is sth that you might be interested in ;) | 15:48 |
| arne_wiebalck | dtantsur: Esp. as the dosc on root devices lists “/dev/md0” as an example for the “name” hint :-D | 15:48 |
| arne_wiebalck | s/dosc/docs/ | 15:49 |
| TheJulia | arne_wiebalck: It might make sense to have an override then... | 15:49 |
| TheJulia | if a path at least, trust the path maybe... | 15:49 |
| TheJulia | well, we already have the override in a sense, we're just not honoring it because we're looking for matches | 15:49 |
| arne_wiebalck | TheJulia: With override you mean to explicitly pass the device name? | 15:50 |
| TheJulia | as the hint, yeah | 15:50 |
| dtantsur | arne_wiebalck: is it about firmware RAID? | 15:50 |
| dtantsur | we've just had people interested in it downstream, but I'm not sure how exactly it will work with ironic/inspector | 15:51 |
| arne_wiebalck | dtantsur: No, I’m trying to get s/w RAID working. | 15:51 |
| dtantsur | arne_wiebalck: what is configuring software RAID when IPA is running? custom hw manager? | 15:51 |
| arne_wiebalck | dtantsur: yes | 15:52 |
| *** e0ne has quit IRC | 15:52 | |
| dtantsur | ah, I see | 15:52 |
| arne_wiebalck | dtantsur: We have a custom manager that creates the RAID, and currently I’m overwriting the get_os_install_device function to define which device to use. | 15:53 |
| arne_wiebalck | dtantsur: Was just traying to understand why the root device hints did not work. | 15:53 |
| rpioso | dtantsur, jroll: Regarding the change in question, I feel the risk is very, very low. The change is to the idrac's management interface, which is used by its classic drivers and h/w type. It's been thoroughly tested with the h/w type. We're actively working on the Dell EMC 3d party CI's tests for stable/ocata so we can demonstrate that it works with the classic drivers. All that said, it's not a biggie either way. It | 15:53 |
| rpioso | would be nice to provide to operators who are on stable/ocata. | 15:53 |
| arne_wiebalck | dtantsur: Then I saw the lsblk command. | 15:54 |
| *** tssurya has quit IRC | 15:54 | |
| dtantsur | arne_wiebalck: I don't know for sure why we prevent that. maybe jroll knows, IPA came from his former team :) | 15:54 |
| arne_wiebalck | dtantsur: Is overwriting get_os_install_device in our manager a sensible approach to get around this? | 15:55 |
| dtantsur | rpioso: okay, I'm fine with that. but please still make the CI pass. | 15:55 |
| arne_wiebalck | dtantsur: Or will I hit other issues? | 15:55 |
| dtantsur | arne_wiebalck: I think you need to go one level deeper and override list_hardware (or how is it called) | 15:55 |
| dtantsur | so that all parts get the correct view of what you consider to be disks | 15:56 |
| rpioso | dtantsur: Ack. We're working on that :-) | 15:56 |
| dtantsur | e.g. if you ever use ironic-inspector | 15:56 |
| dtantsur | rpioso: great :) I'll unblock it when I see it green | 15:56 |
| arne_wiebalck | dtantsur: Right. list_block_devices() it I guess. | 15:56 |
| dtantsur | yeah | 15:56 |
| arne_wiebalck | dtantsur: Yeah … I guess the installation of grub may also get confused otherwise. | 15:57 |
| arne_wiebalck | dtantsur: TheJulia : Thx, I’ll have a look. | 15:58 |
| TheJulia | arne_wiebalck: i think many people would be interested in software raid if you get something working that is viable | 15:58 |
| dtantsur | ++ maybe we should stop blacklisting md devices | 15:59 |
| dtantsur | but then we should be very careful to not expose their underlying devices | 15:59 |
| TheJulia | ++ | 15:59 |
| dtantsur | e.g. if RAID is build from /dev/sda1 and /dev/sda2, we should hide them and show /dev/md0 | 15:59 |
| dtantsur | right? | 15:59 |
| TheJulia | Yeah | 16:00 |
| arne_wiebalck | dtantsur: Show where? | 16:00 |
| dtantsur | arne_wiebalck: s/show/return from list_block_devices/ | 16:00 |
| arne_wiebalck | dtantsur: why? | 16:00 |
| dtantsur | arne_wiebalck: imagine we return /dev/sda1, /dev/sda2 and /dev/md0 and ironic picks /dev/sda2 for deployment. what will happen to /dev/md0? | 16:01 |
| TheJulia | dtantsur: it wouldn't matter now that I think about it | 16:01 |
| TheJulia | the deploy will just fail | 16:01 |
| dtantsur | maybe I misunderstand how the things works | 16:01 |
| TheJulia | because /dev/md0 has locked /dev/sdax | 16:01 |
| dtantsur | well, failed deployment is not an ideal outcome :) | 16:01 |
| *** gyee has joined #openstack-ironic | 16:02 | |
| TheJulia | no, it is not, but it is feedback that something is misconfigured | 16:02 |
| *** pcaruana has quit IRC | 16:02 | |
| TheJulia | and that some action needs to be taken | 16:02 |
| dtantsur | we should only consider devices for installation that we can use in practice | 16:02 |
| TheJulia | agreed, but we likely don't need to block one on the other | 16:02 |
| dtantsur | it may not be entirely obviously from something like "Device /dev/sda2 is locked" that one needs root device hints.. | 16:02 |
| TheJulia | yeah... | 16:02 |
| arne_wiebalck | is list_block_devices supposed to return all block devices or only the ones eligible for installation? | 16:03 |
| jroll | iirc we excluded raid devices in upstream because wiping them fails, but I could be wrong | 16:03 |
| TheJulia | jroll: when you guys had firmware raid, how was it displaying in the OS/ as a /dev/md[x] device? | 16:03 |
| jroll | TheJulia: yes, it was displaying as /dev/md$x, I think | 16:04 |
| TheJulia | ahh | 16:04 |
| TheJulia | see that makes sense then | 16:04 |
| jroll | I have to imagine our hardware manager that did raid things returned it correctly though | 16:04 |
| jroll | or just didn't care what list_block_devices did | 16:05 |
| TheJulia | was that the one upstreamed? | 16:05 |
| TheJulia | or a separate one? | 16:05 |
| jroll | yeah, list_block_devices is overridable | 16:05 |
| jroll | I don't believe it was OSS'd | 16:05 |
| * jroll looking | 16:05 | |
| TheJulia | I remember an example one with a bunch of cleaning stuff... | 16:05 |
| TheJulia | but that is it | 16:05 |
| jroll | right, this was v1 https://github.com/rackerlabs/onmetal-ironic-hardware-manager | 16:06 |
| jroll | v2 had the raids | 16:06 |
| jroll | and also a ton of proprietary vendor-related junk | 16:06 |
| jroll | I don't think we ever split it | 16:06 |
| jroll | yeah, not seeing it :( | 16:06 |
| TheJulia | yeah, reasons start piling up with vendor stuff | 16:07 |
| arne_wiebalck | jroll: thx | 16:08 |
| jroll | no problem | 16:09 |
| *** rpittau has quit IRC | 16:09 | |
| *** jrist has quit IRC | 16:12 | |
| dtantsur | jroll, TheJulia, this is needed to repair ironic-inspector-client gate: https://review.openstack.org/#/c/591659/ https://review.openstack.org/#/c/591661/ https://review.openstack.org/#/c/585995/ | 16:12 |
| patchbot | patch 591659 - python-ironic-inspector-client - CI: stop trying to pull the tempest plugin from ir... - 0h 52m 44s spent in CI | 16:12 |
| patchbot | patch 591661 - python-ironic-inspector-client (stable/queens) - CI: stop trying to pull the tempest plugin from ir... - 0h 49m 16s spent in CI | 16:12 |
| patchbot | patch 585995 - python-ironic-inspector-client (stable/rocky) - Update .gitreview for stable/rocky - 3h 0m 19s spent in CI | 16:12 |
| * TheJulia looks | 16:12 | |
| JayF | jroll: TheJulia: Yeah, we never upstreamed anything from that repo except the basic concept unit tests i upstreamed | 16:13 |
| TheJulia | dtantsur: no rocky backport? | 16:14 |
| dtantsur | TheJulia: that's the .gitreview change | 16:14 |
| dtantsur | I had to squash them because without .gitreview I was not able to correctly submit the patch... | 16:14 |
| TheJulia | ahh, i see | 16:15 |
| TheJulia | k | 16:15 |
| JayF | jroll: TheJulia: arne_wiebalck: onmetal handled this by in one of the super early custom steps, we wiped raid metadata from the disks then rebooted -- then the raid device is gone | 16:16 |
| TheJulia | jroll: rloo: could one of you bless https://review.openstack.org/#/c/585995/ to proceed forth into the gate | 16:16 |
| patchbot | patch 585995 - python-ironic-inspector-client (stable/rocky) - Update .gitreview for stable/rocky - 3h 0m 19s spent in CI | 16:16 |
| JayF | jroll: TheJulia: arne_wiebalck: I'm fairly certain that's how we prevented erase from blowing up on raid devices ... ensuring those devices were gone by that step | 16:16 |
| rloo | TheJulia: done | 16:16 |
| TheJulia | rloo: thanks! | 16:16 |
| arne_wiebalck | JayF: thx! | 16:17 |
| JayF | arne_wiebalck: no problem, good luck! Can I ask what sorta fun things you're doing? | 16:18 |
| arne_wiebalck | JayF: Sure: we’ve introduced Ironic at CERN earlier this year. We almost exclusively use s/w RAID. For now, deploying nodes with Ironic means a double installation. | 16:19 |
| arne_wiebalck | JayF: That’s why we’re trying to integrate this. | 16:20 |
| JayF | Oh, awesome! CERN is always doing cool stuff :) Good luck! I'm not so active with Ironic anymore as my job has changed, but it's nice to see it all getting used :) | 16:20 |
| jroll | JayF: ah, you're right, thanks | 16:20 |
| *** jaganathan has quit IRC | 16:21 | |
| arne_wiebalck | JayF: We’re still learning :) We have 1’200 active nodes, but aim to go to around 10k in the next couple of months … let’s see :) | 16:23 |
| arne_wiebalck | JayF: Why did you need to reboot, i.e. why did the erase blow up after the RAID was stopped? | 16:25 |
| JayF | In this particular case: rebooting was the most straightforward way to stop the raid after we completely invalidated the block devices they were running on | 16:25 |
| JayF | while I'm sure we could've figured out how to make it work sans reboot.... rebooting functionality is built into the hardware manager, and it's the more clean solution... why not? | 16:26 |
| arne_wiebalck | JayF: Ah, wasn’t aware you could reboot inside the h/w manager! | 16:26 |
| JayF | https://github.com/rackerlabs/onmetal-ironic-hardware-manager -- linked earlier -- has some examples, although like jroll said it's not the best examples (those are under lock and key at $employer[-1]) | 16:28 |
| JayF | even just reading the steps can help understand what you can do https://github.com/rackerlabs/onmetal-ironic-hardware-manager/blob/master/onmetal_ironic_hardware_manager/__init__.py#L66 | 16:28 |
| JayF | and that hardware manager interface in the agent is exceedingly well documented | 16:28 |
| *** d0ugal has quit IRC | 16:28 | |
| JayF | I'd suggest reading up on it + some of the examples. You can do pretty much anything you can get working in a ramdisk during cleaning. | 16:29 |
| jroll | JayF: heh, we were looking for "reboot_required" recently and couldn't find it anywhere :/ | 16:29 |
| JayF | did that never get upstreamed!? oh noooo | 16:29 |
| jroll | not sure if/when/why it was removed, or that | 16:30 |
| dtantsur | jroll: we were looking for reboot_required, and this is reboot_requested | 16:30 |
| arne_wiebalck | JayF: Excellent, thanks a lot! | 16:30 |
| dtantsur | this may explain it :D | 16:30 |
| jroll | oh jeez | 16:30 |
| jroll | ahaha | 16:31 |
| jroll | yeah, that's it, thanks JayF :P | 16:31 |
| JayF | https://github.com/openstack/ironic/blob/d3236ef7dd45ee49edb2c2695c5bba4e77e6432e/ironic/drivers/modules/agent_base_vendor.py#L441 | 16:31 |
| dtantsur | I do see some code to back reboot_requested | 16:31 |
| jroll | yep | 16:31 |
| JayF | Man I leave for a year and you all start losing features in the couch cushions :P | 16:31 |
| dtantsur | lol | 16:31 |
| jroll | we didn't lose features, only our search index apparently | 16:31 |
| *** jrist has joined #openstack-ironic | 16:38 | |
| JayF | arne_wiebalck: one more useful thing -> https://github.com/openstack/ipa-example-hardware-managers was explicitly made to be examples / docs for folks implementing hardware managers. Has some explicit copy-paste ready examples. | 16:39 |
| *** mmethot has quit IRC | 16:41 | |
| NobodyCam | Good Morning Ironic'ers | 16:53 |
| *** serlex has quit IRC | 16:56 | |
| *** w-miller has quit IRC | 17:01 | |
| *** w-miller has joined #openstack-ironic | 17:02 | |
| *** derekh has quit IRC | 17:04 | |
| *** mgoddard has quit IRC | 17:05 | |
| rpioso | NobodyCam: Good morning-ish :) | 17:06 |
| *** w-miller has quit IRC | 17:06 | |
| NobodyCam | :) good morning rpioso :) | 17:07 |
| *** S4ren has quit IRC | 17:07 | |
| NobodyCam | Yea moving slow today :p | 17:07 |
| openstackgerrit | Merged openstack/python-ironic-inspector-client master: CI: stop trying to pull the tempest plugin from ironic-inspector https://review.openstack.org/591659 | 17:07 |
| openstackgerrit | Merged openstack/python-ironic-inspector-client master: Update reno for stable/rocky https://review.openstack.org/585997 | 17:07 |
| openstackgerrit | Merged openstack/python-ironic-inspector-client stable/rocky: Update .gitreview for stable/rocky https://review.openstack.org/585995 | 17:13 |
| dtantsur | morning NobodyCam | 17:14 |
| NobodyCam | hey hey dtantsur Good Morning :) | 17:14 |
| *** baha has quit IRC | 17:18 | |
| openstackgerrit | Paul Belanger proposed openstack/bifrost master: Refactor bifrost-integration-dibipa-debian jobs https://review.openstack.org/591795 | 17:21 |
| *** Bhujay has quit IRC | 17:22 | |
| dtantsur | bfournie: https://review.openstack.org/#/c/591661/ has not merged yet | 17:22 |
| patchbot | patch 591661 - python-ironic-inspector-client (stable/queens) - CI: stop trying to pull the tempest plugin from ir... - 0h 49m 16s spent in CI | 17:22 |
| dtantsur | rloo: if you have a second ^^ | 17:23 |
| dtantsur | bfournie: you can rebase your patch on top of the fix | 17:23 |
| rloo | dtantsur: in meeting, give me 30 min or so | 17:23 |
| TheJulia | I just marked it wf+1 | 17:23 |
| TheJulia | since the patch on master merged | 17:24 |
| bfournie | dtantsur: hmm, yeah I saw something else that had merged and thought it was that… | 17:24 |
| *** gkadam has quit IRC | 17:53 | |
| openstackgerrit | Aija Jaunteva proposed openstack/sushy master: WIP: Add Message parsing support with all registry loading https://review.openstack.org/590747 | 17:59 |
| jroll | should I expect everything in playbooks/roles/bifrost-ironic-install/defaults/main.yml to be present in a given ironic-install task? https://review.openstack.org/#/c/591417 is making me doubt myself | 17:59 |
| patchbot | patch 591417 - bifrost - Allow extra packages to be installed with ironic - 0h 21m 51s spent in CI | 17:59 |
| jroll | specifically I get 2018-08-14 14:23:33.569076 | primary | "msg": "Unexpected failure in finding the lookup named '{{ ironic_extra_packages|default([]) }}' in the available lookup plugins" | 18:01 |
| jroll | when ironic_extra_packages is defined in defaults/main.yml | 18:01 |
| openstackgerrit | Aija Jaunteva proposed openstack/sushy master: Update to use mapped field for Message Severity field https://review.openstack.org/591612 | 18:02 |
| openstackgerrit | Aija Jaunteva proposed openstack/sushy master: WIP: Add @Redfish.Settings update status and expose it for BIOS https://review.openstack.org/591635 | 18:06 |
| dtantsur | jroll: that's a weird message, "default" should cover an undefined variable | 18:06 |
| jroll | dtantsur: yep, and it should be found anyway :/ | 18:07 |
| jroll | I'm wonder if the include: in tasks/main.yml is what's killing it | 18:07 |
| dtantsur | jroll: I have a gut feeling that maybe "loop" does not behave the way you want | 18:09 |
| dtantsur | what if you try the older with_items? | 18:09 |
| *** trown is now known as trown|lunch | 18:11 | |
| jroll | dtantsur: yeah, will try, I thought that was deprecated | 18:13 |
| dtantsur | maybe? I'm not sure | 18:13 |
| dtantsur | I think with_<anything> is somehow converted into loop+lookup | 18:13 |
| jroll | ¯\_(ツ)_/¯ | 18:13 |
| jroll | we'll see, thanks for the tip | 18:14 |
| * jroll wonders why git-review is so slow on his machine | 18:14 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/bifrost master: Allow extra packages to be installed with ironic https://review.openstack.org/591417 | 18:14 |
| * dtantsur waves | 18:15 | |
| *** dtantsur is now known as dtantsur|afk | 18:15 | |
| openstackgerrit | Merged openstack/python-ironic-inspector-client stable/queens: CI: stop trying to pull the tempest plugin from ironic-inspector https://review.openstack.org/591661 | 18:18 |
| openstackgerrit | Aija Jaunteva proposed openstack/sushy master: WIP: Add @Redfish.Settings update status and expose it for BIOS https://review.openstack.org/591635 | 18:21 |
| TheJulia | jroll: default won't work with nested data structures unless you explicitly define them when the default yaml is parsed | 18:29 |
| TheJulia | jroll: if I'm grokking what your trying to do correctly after my brain context change | 18:30 |
| *** baha has joined #openstack-ironic | 18:36 | |
| jroll | TheJulia: I think you're grokking correctly but I'm not sure what you mean by 'unless you explicitly define them when the default yaml is parsed' | 18:41 |
| TheJulia | jroll: defaults/main.yaml gets loaded out of the gate, so he defaults for nested things need to be present or overriden then | 18:43 |
| * TheJulia hopes that makes sense | 18:43 | |
| jroll | TheJulia: that makes sense, but I have a default in defaults/main | 18:44 |
| jroll | or I'm misunderstanding you, and the fact I have the default doesn't matter :) | 18:45 |
| TheJulia | hmm :\ | 18:45 |
| TheJulia | hmmm | 18:46 |
| TheJulia | so if the yaml parses that into an empty list, I don't think the iterator will run and you don't need the default in the install.yaml file unless it is ever separated from the defaults | 18:48 |
| jroll | right, that's what I thought | 18:49 |
| jroll | but things blew up, so I added a default | 18:49 |
| jroll | and it still blows up in the same way | 18:49 |
| TheJulia | hmm | 18:49 |
| TheJulia | is value: [] valid for a list in yaml? | 18:49 |
| jroll | according to a yaml linter, yes | 18:49 |
| jroll | I should probably try it in python | 18:49 |
| TheJulia | hmm | 18:49 |
| jroll | but, I'm wondering if include: in tasks/main.yml does not use the vars for some reason | 18:50 |
| jroll | I can't imagine that's right, but it's all I've got | 18:50 |
| TheJulia | it could be default with_items is just not valid on the expansion that occurs for it | 18:52 |
| TheJulia | what was the error with an empty list? | 18:52 |
| TheJulia | ugh, ironic-ui job settings were never moved out of infra | 18:52 |
| *** trown|lunch is now known as trown | 18:53 | |
| jroll | TheJulia: same error, see PS2 | 18:54 |
| TheJulia | but have you tried with_items: "{{ ironic_extra_packages }}" | 18:55 |
| jroll | no, I can give that a shot | 18:56 |
| openstackgerrit | Julia Kreger proposed openstack/virtualbmc master: Change jobs for Python3 effort https://review.openstack.org/589295 | 18:57 |
| *** ajya has quit IRC | 18:57 | |
| *** ajya has joined #openstack-ironic | 18:57 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/bifrost master: Allow extra packages to be installed with ironic https://review.openstack.org/591417 | 19:00 |
| openstackgerrit | Julia Kreger proposed openstack/networking-generic-switch master: Change jobs for Python3 effort https://review.openstack.org/591825 | 19:01 |
| openstackgerrit | Julia Kreger proposed openstack/networking-generic-switch master: Switch n-g-s devstack job to python3 https://review.openstack.org/591826 | 19:01 |
| openstackgerrit | Julia Kreger proposed openstack/sushy master: Change jobs for Python3 effort https://review.openstack.org/591827 | 19:06 |
| openstackgerrit | Julia Kreger proposed openstack/sushy master: Change sushy devstack job to python3 https://review.openstack.org/591828 | 19:06 |
| openstackgerrit | Merged openstack/networking-generic-switch master: Detection of config errors for netmiko drivers https://review.openstack.org/589960 | 19:12 |
| *** ajya[m] has joined #openstack-ironic | 19:18 | |
| openstackgerrit | Julia Kreger proposed openstack/sushy-tools master: Change jobs for Python3 effort https://review.openstack.org/591830 | 19:19 |
| openstackgerrit | Julia Kreger proposed openstack/sushy-tools master: Switch sushy-tools devstack job to Python3 https://review.openstack.org/591831 | 19:19 |
| openstackgerrit | Julia Kreger proposed openstack/virtualbmc master: Change py35 to py3 so Python 3.6 can be picked up https://review.openstack.org/591833 | 19:22 |
| *** e0ne has joined #openstack-ironic | 19:23 | |
| *** bfournie has quit IRC | 19:26 | |
| *** e0ne has quit IRC | 19:26 | |
| *** e0ne has joined #openstack-ironic | 19:27 | |
| openstackgerrit | Julia Kreger proposed openstack/python-ironicclient master: Change jobs for Python3 effort https://review.openstack.org/591834 | 19:34 |
| openstackgerrit | Julia Kreger proposed openstack/python-ironicclient master: Switch ci jobs to Python3 https://review.openstack.org/591835 | 19:34 |
| openstackgerrit | Julia Kreger proposed openstack/python-ironic-inspector-client master: Change basepython to python3 https://review.openstack.org/591838 | 19:43 |
| openstackgerrit | Julia Kreger proposed openstack/python-ironic-inspector-client master: Change jobs for Python3 effort https://review.openstack.org/591839 | 19:43 |
| openstackgerrit | Julia Kreger proposed openstack/python-ironic-inspector-client master: Switch inspector client dsvm job to use python3 https://review.openstack.org/591840 | 19:43 |
| * TheJulia wonders if "TheJulia is evil" shirts will appear at the PTG | 19:45 | |
| *** ajya has quit IRC | 19:46 | |
| jroll | heh | 19:47 |
| jroll | TheJulia: didn't doug's email say those patches will be automated? | 19:47 |
| TheJulia | I didn't read it that way | 19:47 |
| TheJulia | well, that he wrote some tools to help too | 19:47 |
| TheJulia | most of it is really minor changes and mindless is sometimes what the brain needs | 19:48 |
| jroll | fair enough :) | 19:48 |
| * TheJulia misses the days of fixing datacenter wiring | 19:49 | |
| jroll | I'm referring to this, btw: http://lists.openstack.org/pipermail/openstack-dev/2018-August/133232.html | 19:49 |
| jroll | "The goal champions will start by preparing the patches for steps 1-5, and 8" | 19:49 |
| TheJulia | heh | 19:50 |
| TheJulia | sigh | 19:50 |
| TheJulia | I started based on the book that was the original goal documentation | 19:50 |
| TheJulia | maybe an elbow tatoo is less painful | 19:52 |
| jroll | lol | 20:02 |
| *** jcoufal has quit IRC | 20:19 | |
| *** baha has quit IRC | 20:22 | |
| *** bfournie has joined #openstack-ironic | 20:28 | |
| *** S4ren has joined #openstack-ironic | 20:56 | |
| *** trown is now known as trown|outtypewww | 21:06 | |
| *** S4ren has quit IRC | 21:18 | |
| *** slagle has quit IRC | 21:23 | |
| jroll | TheJulia: holy cow, your last suggestion works | 21:28 |
| TheJulia | \o/ | 21:29 |
| *** rh-jelabarre has quit IRC | 21:37 | |
| *** cdearborn has quit IRC | 22:25 | |
| *** mbeierl has quit IRC | 22:38 | |
| *** e0ne has quit IRC | 22:39 | |
| *** mbeierl has joined #openstack-ironic | 22:43 | |
| *** mjturek has quit IRC | 23:22 | |
| *** rpioso is now known as rpioso|afk | 23:30 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
