| * TheJulia has lots to review tomorrow morning | 00:10 | |
| *** gyee has quit IRC | 00:21 | |
| *** altlogbot_1 has quit IRC | 00:23 | |
| *** irclogbot_0 has quit IRC | 00:23 | |
| *** threestrands has quit IRC | 00:32 | |
| *** Goneri has quit IRC | 01:02 | |
| *** rloo has quit IRC | 01:05 | |
| *** igordc has quit IRC | 01:16 | |
| *** irclogbot_1 has joined #openstack-ironic | 01:34 | |
| *** irclogbot_1 has quit IRC | 01:39 | |
| *** irclogbot_3 has joined #openstack-ironic | 01:40 | |
| *** irclogbot_3 has quit IRC | 01:43 | |
| *** irclogbot_2 has joined #openstack-ironic | 02:54 | |
| *** irclogbot_2 has quit IRC | 02:59 | |
| *** rh-jelabarre has quit IRC | 03:04 | |
| *** ianychoi has quit IRC | 03:12 | |
| *** ianychoi has joined #openstack-ironic | 03:13 | |
| *** irclogbot_0 has joined #openstack-ironic | 03:34 | |
| *** irclogbot_0 has quit IRC | 03:39 | |
| *** mkrai has joined #openstack-ironic | 03:49 | |
| *** gkadam has joined #openstack-ironic | 03:59 | |
| *** gkadam_ has joined #openstack-ironic | 04:01 | |
| *** gkadam_ has quit IRC | 04:01 | |
| *** irclogbot_2 has joined #openstack-ironic | 04:02 | |
| *** gkadam has quit IRC | 04:03 | |
| *** irclogbot_2 has quit IRC | 04:07 | |
| *** ricolin_ has joined #openstack-ironic | 04:08 | |
| *** ricolin_ has quit IRC | 04:08 | |
| *** ricolin has joined #openstack-ironic | 04:10 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic master: iLO driver doc update https://review.opendev.org/679702 | 04:15 |
|---|---|---|
| *** irclogbot_0 has joined #openstack-ironic | 04:22 | |
| openstackgerrit | Merged openstack/ironic-lib master: Deprecate [disk_utils]iscsi_verify_attempts https://review.opendev.org/677558 | 04:31 |
| openstackgerrit | Merged openstack/ironic master: Add a release note for iscsi_verify_attempts deprecation https://review.opendev.org/679792 | 04:31 |
| *** irclogbot_0 has quit IRC | 04:33 | |
| *** rachit7 has joined #openstack-ironic | 05:17 | |
| *** gregoryo has joined #openstack-ironic | 05:20 | |
| *** irclogbot_3 has joined #openstack-ironic | 05:20 | |
| *** irclogbot_3 has quit IRC | 05:25 | |
| *** rcernin_ has joined #openstack-ironic | 05:31 | |
| *** rcernin has quit IRC | 05:33 | |
| openstackgerrit | Maksim Malchuk proposed openstack/networking-baremetal master: Add the option to choose the Ironic endpoint type https://review.opendev.org/670812 | 05:47 |
| *** michchap has quit IRC | 05:58 | |
| *** hjensas has quit IRC | 06:01 | |
| *** mbeierl has quit IRC | 06:03 | |
| *** jtomasek has joined #openstack-ironic | 06:06 | |
| *** ricolin_ has joined #openstack-ironic | 06:15 | |
| *** irclogbot_2 has joined #openstack-ironic | 06:16 | |
| *** ricolin has quit IRC | 06:17 | |
| *** irclogbot_2 has quit IRC | 06:21 | |
| *** irclogbot_2 has joined #openstack-ironic | 06:24 | |
| *** irclogbot_2 has quit IRC | 06:29 | |
| arne_wiebalck | Good morning, ironic! | 06:32 |
| *** diga has joined #openstack-ironic | 06:54 | |
| *** rcernin_ has quit IRC | 07:01 | |
| rachit7 | Good morning arne_wiebalck | 07:04 |
| *** stendulker has joined #openstack-ironic | 07:05 | |
| arne_wiebalck | Hey rachit7 o/ | 07:09 |
| openstackgerrit | Ankit Kumar proposed openstack/ironic master: Add iso less virtual media deploy https://review.opendev.org/677996 | 07:11 |
| *** toabctl has quit IRC | 07:11 | |
| *** hjensas has joined #openstack-ironic | 07:15 | |
| *** hjensas has quit IRC | 07:17 | |
| *** toabctl has joined #openstack-ironic | 07:19 | |
| *** irclogbot_2 has joined #openstack-ironic | 07:30 | |
| *** irclogbot_2 has quit IRC | 07:35 | |
| openstackgerrit | paresh sao proposed openstack/ironic master: Follow-up patch for out-of-band `erase_devices` clean step for iLO https://review.opendev.org/679927 | 07:36 |
| *** pcaruana has joined #openstack-ironic | 07:42 | |
| *** gregoryo has quit IRC | 07:43 | |
| *** lucasagomes has joined #openstack-ironic | 07:47 | |
| *** irclogbot_1 has joined #openstack-ironic | 07:50 | |
| *** irclogbot_1 has quit IRC | 07:53 | |
| *** hjensas has joined #openstack-ironic | 07:54 | |
| *** ociuhandu has joined #openstack-ironic | 07:57 | |
| *** e0ne has joined #openstack-ironic | 07:59 | |
| *** ociuhandu has quit IRC | 07:59 | |
| *** ociuhandu has joined #openstack-ironic | 08:00 | |
| *** hwoarang has quit IRC | 08:05 | |
| *** hwoarang has joined #openstack-ironic | 08:07 | |
| *** e0ne has quit IRC | 08:07 | |
| *** rachit7 has quit IRC | 08:21 | |
| *** irclogbot_2 has joined #openstack-ironic | 08:22 | |
| *** dougsz has joined #openstack-ironic | 08:23 | |
| *** ricolin_ is now known as ricolin | 08:25 | |
| *** irclogbot_2 has quit IRC | 08:25 | |
| *** e0ne has joined #openstack-ironic | 08:35 | |
| *** derekh has joined #openstack-ironic | 08:36 | |
| *** mkrai has quit IRC | 08:39 | |
| *** brtknr has joined #openstack-ironic | 08:41 | |
| brtknr | hey all, anyone out there booting up #fedora-coreos on baremetal? | 08:41 |
| brtknr | i am strugging to inject ignition metadata to the baremetal instance... which works fine on VM | 08:42 |
| *** alexmcleod has joined #openstack-ironic | 08:44 | |
| *** priteau has joined #openstack-ironic | 08:45 | |
| *** ociuhandu has quit IRC | 08:55 | |
| *** ftarasenko has joined #openstack-ironic | 08:55 | |
| *** ociuhandu has joined #openstack-ironic | 08:56 | |
| ftarasenko | Dear team, can you point me to documentation/code regarding port groups in Ironic used with Nova? Currently I get empty network_data.json in config drive. | 08:57 |
| ftarasenko | Should the config drive be created manually or I miss something in my configuration? | 08:57 |
| *** ociuhandu has quit IRC | 09:00 | |
| *** trident has quit IRC | 09:01 | |
| *** irclogbot_2 has joined #openstack-ironic | 09:08 | |
| *** trident has joined #openstack-ironic | 09:10 | |
| *** irclogbot_2 has quit IRC | 09:11 | |
| *** stendulker has quit IRC | 09:16 | |
| *** dsneddon_ has joined #openstack-ironic | 09:17 | |
| *** dsneddon has quit IRC | 09:18 | |
| *** dtantsur|afk is now known as dtantsur | 09:22 | |
| dtantsur | ftarasenko: I think nova is supposed to create network_data.json, but I don't know this code really well. hjensas, mgoddard? | 09:23 |
| dtantsur | morning ironic | 09:23 |
| xXraphXx | dtantsur, ftarasenko: hi, yep nova does it but there are some cases when it cannot fill the data, becasue they are not available at the time the configdrive is built (with neutron segments activated for example, or in any case when the port ip allocation is deferred) | 09:25 |
| dtantsur | indeed | 09:26 |
| dtantsur | morning (?) xXraphXx | 09:26 |
| ftarasenko | haha, I definetely have segments in my configuration. What should be done in this case - manual built of config drive or we need to patch nova? | 09:26 |
| ftarasenko | xXraphXx have you seen any working solution? I think, for tripleo manual built of config drives is used. | 09:29 |
| dtantsur | tripleo uses os-net-config for network configuration | 09:30 |
| xXraphXx | ftarasenko. No, not seen any upstream solution for now, unless I'm wrong :/. Not sure what the upstream solution should be: may be write a neutron custom ipam ? on my side, i've made sure the data are statically stored on the ironic node driver info. And made a nova driver to override the network metadata. But this is so ugly I did not push it upstream | 09:31 |
| dtantsur | well, there is a thing in progress for ironic-side mini-IPAM | 09:33 |
| dtantsur | maybe we could integrate it with nova as well | 09:33 |
| dtantsur | http://specs.openstack.org/openstack/ironic-specs/specs/not-implemented/L3-based-deployment.html (etingof is updating it) | 09:33 |
| * etingof is interested in real-world use-cases | 09:34 | |
| *** ricolin has quit IRC | 09:34 | |
| *** ricolin has joined #openstack-ironic | 09:35 | |
| xXraphXx | ftarasenko, thinking about some solution that might be a solution for you (or not, depending on your use case): if you can make sure that you specify a fixed ip in all your nova boot, then you will bypass the ip allocation deferred and thus bypass the problem | 09:36 |
| hjensas | sounds like nova would need to build the config-drive after the instance has been scheduled to a host, the port is bound and neutron can allocate an IP? | 09:37 |
| xXraphXx | hjensas, yep, once an ironic node is bound to the instance port (after vif_attach call to ironic, if not mistaken) | 09:38 |
| dtantsur | but we're binding a port quite late in the deployment | 09:41 |
| dtantsur | we may need a new API endpoint for nova to use | 09:41 |
| xXraphXx | yep in driver.spawn | 09:41 |
| xXraphXx | (was answering to the quite late assertion) | 09:42 |
| *** ftarasenko has quit IRC | 09:42 | |
| dtantsur | it would be great if somebody who has a clue what is going on could drive this work further | 09:43 |
| xXraphXx | my bad it's a bit before driver.spawn it's in plug_vifs | 09:44 |
| dtantsur | does port binding actually happen there? | 09:44 |
| *** ftarasenko has joined #openstack-ironic | 09:44 | |
| dtantsur | I think we attach VIFs, but I don't remember if we actually call into neutron to bind ports | 09:44 |
| dtantsur | I suspect it may only be done in ironic's boot interface | 09:45 |
| xXraphXx | checking in ironic | 09:45 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Correctly set OS_TEST_TIMEOUT in the multinode CI job https://review.opendev.org/679949 | 09:48 |
| xXraphXx | dtantsur, you're right | 09:48 |
| dtantsur | xXraphXx: we could probably make ironic rebuild network_data.json | 09:49 |
| dtantsur | After all, we're already in the business of building config drives | 09:49 |
| * dtantsur needs more opinions on ^^^ | 09:49 | |
| *** rachit7 has joined #openstack-ironic | 09:49 | |
| xXraphXx | binding is, depending on the driver, in the deployment end (def configure_tenant_networks). rebuild network_data could be a good solution, if it's planned for ironic to tamper with the configdrive | 09:51 |
| dtantsur | we can open a discussion on it. we have a valid use case now. | 09:53 |
| *** diga has quit IRC | 09:54 | |
| ftarasenko | xXraphXx dtantsur thank you for explanation, try to create workaround that will work in my setup. Not so good in python/openstack to discuss how it should be integrated with upstream, but the solution of rebuilding config drive looks good to me. | 09:55 |
| hjensas | dtantsur: won't we need to write the config-driver prior to calling configure_tenant_networks? There will be a vif attach to the provisioning network, which may trigger ml2 configuring switch etc. Once configure_tenant_networks happens the ml2 will reconfigure. We would have to be done generating and writing the config-drive at that point. | 09:55 |
| dtantsur | ouch | 09:55 |
| dtantsur | yes, you're right | 09:56 |
| dtantsur | sounds like someone should start an ML thread to collect more opinions | 09:56 |
| * hjensas wonder if we could use placement+neutron info to update the port with a fixed-ip prior to doing the bind. | 09:57 | |
| hjensas | ML ++ | 09:57 |
| *** hjensas is now known as hjensas|lunch | 09:57 | |
| dtantsur | well, when configdrive is written, we're still attached to a provisioning network with a different IP | 09:58 |
| * dtantsur wonders why the port doesn't end up being bound before nova calls us | 09:58 | |
| *** irclogbot_1 has joined #openstack-ironic | 10:28 | |
| *** irclogbot_1 has quit IRC | 10:33 | |
| *** irclogbot_1 has joined #openstack-ironic | 10:42 | |
| *** irclogbot_1 has quit IRC | 10:45 | |
| *** irclogbot_3 has joined #openstack-ironic | 10:58 | |
| *** ociuhandu has joined #openstack-ironic | 11:01 | |
| *** irclogbot_3 has quit IRC | 11:02 | |
| etingof | if rpioso o/ got a link to the recommended way of flipping boot mode for virtual media boot - that would be helpful | 11:02 |
| *** trident has quit IRC | 11:04 | |
| *** dougsz has quit IRC | 11:07 | |
| *** hjensas|lunch is now known as hjensas | 11:09 | |
| *** trident has joined #openstack-ironic | 11:13 | |
| *** alexmcleod has quit IRC | 11:13 | |
| *** alexmcleod has joined #openstack-ironic | 11:14 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: devstack: save iPXE httpd logs https://review.opendev.org/679968 | 11:16 |
| *** mkrai has joined #openstack-ironic | 11:17 | |
| *** alexmcleod has quit IRC | 11:21 | |
| *** lucasagomes has quit IRC | 11:25 | |
| *** lucasagomes has joined #openstack-ironic | 11:25 | |
| *** bfournie has quit IRC | 11:27 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: WIP ipv6 jobs https://review.opendev.org/613169 | 11:28 |
| *** rh-jelabarre has joined #openstack-ironic | 11:29 | |
| *** ftarasenko has quit IRC | 11:39 | |
| *** irclogbot_2 has joined #openstack-ironic | 11:42 | |
| *** rachit7 has quit IRC | 11:44 | |
| *** irclogbot_2 has quit IRC | 11:48 | |
| *** alexmcleod has joined #openstack-ironic | 12:03 | |
| *** tesseract has joined #openstack-ironic | 12:03 | |
| *** bfournie has joined #openstack-ironic | 12:18 | |
| dtantsur | mgoddard: are there outstanding n-g-s patches that we need to land in train? | 12:24 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-ui master: Update various links in the documentation https://review.opendev.org/617803 | 12:24 |
| dtantsur | updated, please somebody review ^^^ | 12:24 |
| *** dougsz has joined #openstack-ironic | 12:26 | |
| dtantsur | hjensas: similar question re networking-baremetal: do we need anything in train? | 12:31 |
| hjensas | dtantsur: https://review.opendev.org/673972 ? I have a couple of other changes I should pick up when I have time, but nothing we need in train afik. | 12:36 |
| patchbot | patch 673972 - networking-baremetal - Blacklist sphinx 2.1.0 (autodoc bug) - 2 patch sets | 12:36 |
| *** iurygregory has joined #openstack-ironic | 12:36 | |
| dtantsur | ok, thanks | 12:36 |
| iurygregory | good morning all | 12:36 |
| dtantsur | hi iurygregory | 12:37 |
| dtantsur | iurygregory: anything we need in the exporter for train (except for the descriptions patch)? | 12:37 |
| iurygregory | dtantsur, I think we may want to add the initial support for RedFish metrics right etingof | 12:37 |
| iurygregory | ? | 12:38 |
| etingof | iurygregory, yes, I plan to work on this hopefully by the end of this week | 12:38 |
| etingof | dtantsur, ^ | 12:40 |
| dtantsur | etingof: it really has to happen by the end of this week or it'll miss train | 12:40 |
| openstackgerrit | Dmitry Tantsur proposed openstack/bifrost master: DNM checking CI health https://review.opendev.org/679981 | 12:41 |
| etingof | dtantsur, yeah, that's tight | 12:42 |
| dtantsur | etingof: yep. do you think it's doable and you will have cycles for it? | 12:42 |
| etingof | dtantsur, is it Friday or Monday? | 12:42 |
| etingof | dtantsur, well, the plan is to add some rudimentary parser for redfish metrics | 12:42 |
| etingof | dtantsur, then improve that by bugfixes | 12:42 |
| dtantsur | I mean.. given that exporter is not used anywhere, I'm fine with not applying feature freeze to it | 12:43 |
| dtantsur | but if you need to change ironic, the patches have to be up, say, by Monday meeting | 12:43 |
| dtantsur | and if you need to change sushy, by EOD today | 12:43 |
| etingof | dtantsur, I think ironic and sushy parts are all in place for the metrics for now | 12:48 |
| dtantsur | okay, then the schedule is less tight. but let's have this work wrapped up by end of next week please. | 12:50 |
| *** ociuhandu has quit IRC | 12:50 | |
| dtantsur | etingof: what's the status of sushy-tools now? | 12:54 |
| *** ociuhandu has joined #openstack-ironic | 12:55 | |
| openstackgerrit | Merged openstack/networking-baremetal master: Blacklist sphinx 2.1.0 (autodoc bug) https://review.opendev.org/673972 | 12:55 |
| *** mkrai has quit IRC | 12:56 | |
| *** mbeierl has joined #openstack-ironic | 13:11 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Release notes cleanup for 13.0.0 https://review.opendev.org/679985 | 13:15 |
| etingof | dtantsur, I just +2 sushy-tools patches that I think are good to go | 13:15 |
| etingof | dtantsur, plus this patch of mine I'd suggest to review -- https://review.opendev.org/#/c/679456/ | 13:16 |
| patchbot | patch 679456 - sushy-tools - Fix non-functional driver configuration on Py2 - 1 patch set | 13:16 |
| dtantsur | k, thx | 13:16 |
| *** mbeierl has quit IRC | 13:17 | |
| etingof | dtantsur, I'd love to merge this patch as well, though there are some nits to address -- https://review.opendev.org/#/c/677119/ | 13:17 |
| patchbot | patch 677119 - sushy-tools - Add Volume resource support - 5 patch sets | 13:17 |
| *** rachit7 has joined #openstack-ironic | 13:19 | |
| *** mkrai has joined #openstack-ironic | 13:22 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Release notes cleanup for the next release https://review.opendev.org/679989 | 13:25 |
| brtknr | anyone out there booting up fedora coreos baremetal in openstack? | 13:26 |
| brtknr | how do you pass user-data? | 13:26 |
| *** hjensas has quit IRC | 13:29 | |
| dtantsur | brtknr: I'm aware that it's possible. If you use Nova, you may need to look at its documentation. You also need a coreos image that understands configdrives (I think it should be an openstack image). | 13:30 |
| brtknr | i have been able to boot up fedora coreos as a VM but not as a baremetal instance... | 13:31 |
| brtknr | I have tried enabling config drive but no luck so far | 13:32 |
| brtknr | in the VM, it appears to get the ignition config from `http://169.254.169.254/openstack/latest/user_data` | 13:32 |
| dtantsur | right, the metadata service | 13:32 |
| *** mbeierl has joined #openstack-ironic | 13:32 | |
| dtantsur | brtknr: you may need to tell nova to generate configdrive | 13:32 |
| *** irclogbot_2 has joined #openstack-ironic | 13:38 | |
| *** irclogbot_2 has quit IRC | 13:42 | |
| openstackgerrit | Merged openstack/sushy master: Add conditional field matching https://review.opendev.org/675073 | 13:43 |
| *** rloo has joined #openstack-ironic | 13:45 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/bifrost master: Make the opensuse-15 jobs non-voting https://review.opendev.org/679981 | 13:51 |
| dtantsur | folks, this seems needed for bifrost CI ^^ | 13:51 |
| TheJulia | ugh | 13:51 |
| dtantsur | TheJulia: morning | 13:51 |
| dtantsur | I'm not sure how long it will take to recover, but it's not the first day | 13:52 |
| TheJulia | it happens sadly | 13:52 |
| TheJulia | we've had a few weeks once where debian was having repo problems too | 13:52 |
| dtantsur | I will write something to the ML when it merges | 13:53 |
| TheJulia | brtknr: That image for fedora coreos is _supposed_ to also check the disks for a config drive partition upon boot, but the build team is very focused on single path images as opposed to checking for all possible metadata services/configuration drive formats. So as dtantsur said, enabling config drive should solve it for you | 13:55 |
| *** irclogbot_3 has joined #openstack-ironic | 13:59 | |
| brtknr | TheJulia: dtantsur: I have already tried enabling config drive but didn't get anywhere... the instance boots up okay... but I cant login because my key is not added | 14:00 |
| TheJulia | brtknr: do you have a link to the image your using? | 14:01 |
| brtknr | https://builds.coreos.fedoraproject.org/prod/streams/testing/builds/30.20190801.0/x86_64/fedora-coreos-30.20190801.0-openstack.qcow2.xz | 14:01 |
| brtknr | this one | 14:01 |
| *** irclogbot_3 has quit IRC | 14:02 | |
| TheJulia | brtknr: to you have the general download page your going from? | 14:02 |
| *** mkrai has quit IRC | 14:02 | |
| *** mkrai_ has joined #openstack-ironic | 14:03 | |
| brtknr | https://getfedora.org/en/coreos/download/ | 14:03 |
| brtknr | this one | 14:03 |
| openstackgerrit | Merged openstack/ironic-ui master: Update various links in the documentation https://review.opendev.org/617803 | 14:03 |
| brtknr | TheJulia: ^ | 14:03 |
| brtknr | I'm using the OpenStack iamge | 14:04 |
| brtknr | Not sure how it differs from the general Qemu one | 14:04 |
| *** irclogbot_3 has joined #openstack-ironic | 14:05 | |
| *** mkrai_ has quit IRC | 14:05 | |
| TheJulia | brtknr: they have specific logic on what to do to get configuration | 14:05 |
| TheJulia | baked into the image | 14:06 |
| TheJulia | Someone should tell the fedora folks that OpenStack can be baremetal as well :( | 14:06 |
| openstackgerrit | Merged openstack/sushy-tools master: Add Simple Storage resource support https://review.opendev.org/667328 | 14:06 |
| brtknr | TheJulia: so ive been able to boot up the image on a vm instance but sadly not as baremetal | 14:07 |
| TheJulia | Thank you OOMKiller | 14:07 |
| *** irclogbot_3 has quit IRC | 14:08 | |
| * TheJulia extracts the image to take apart the qcow2 | 14:08 | |
| *** mkrai has joined #openstack-ironic | 14:08 | |
| *** irclogbot_1 has joined #openstack-ironic | 14:10 | |
| brtknr | TheJulia: what do you use to extract qcow2? | 14:12 |
| TheJulia | Generally If I have the disk space, I extract it to raw and mount it with losetup | 14:13 |
| TheJulia | as a loop device, and go from there | 14:13 |
| *** irclogbot_1 has quit IRC | 14:14 | |
| dtantsur | you can mount a qcow2 using guestfish | 14:15 |
| dtantsur | (you can even modify it this way) | 14:15 |
| dtantsur | brtknr: e.g. https://github.com/openstack/metalsmith/blob/master/playbooks/integration/centos-image.yaml#L55 | 14:16 |
| dtantsur | (this file a few fun hacks) | 14:16 |
| brtknr | i've just mounted /dev/sda1 from the qcow2 image but nothing too exciting in there | 14:17 |
| brtknr | using guestfish | 14:17 |
| openstackgerrit | Iury Gregory Melo Ferreira proposed openstack/ironic-prometheus-exporter master: Support for metrics description https://review.opendev.org/675345 | 14:17 |
| TheJulia | brtknr: the secrets are in the 4th partition | 14:25 |
| rpioso | etingof: ^^^ I don't understand your request. Please elaborate. | 14:25 |
| *** edleafe has joined #openstack-ironic | 14:27 | |
| *** edleafe has left #openstack-ironic | 14:28 | |
| brtknr | TheJulia: I can mount sda1 and sda2 but not sda4 | 14:29 |
| brtknr | using guestfish | 14:30 |
| *** tesseract has quit IRC | 14:30 | |
| brtknr | using guestmount* | 14:30 |
| brtknr | libguestfs: error: mount_options: mount exited with status 32: mount: wrong fs type, bad option, bad superblock on /dev/sda4, | 14:30 |
| TheJulia | rutro... ostree/deploy/fedora-coreos/deploy/a9c8d66d3628d1b9b4c4690777e8b730d08329b4359410cb410a2003296af1ca.0/usr/lib/os-release:PLATFORM_ID="platform:f30" | 14:31 |
| TheJulia | they changed it | 14:32 |
| etingof | rpioso, is there some element in imported configuration XML to specify boot mode? -- https://github.com/etingof/sushy-oem-dellemc/blob/master/sushy_oem_dellemc/resources/manager/manager.py#L45 | 14:32 |
| TheJulia | brtknr: mount it read only xfs | 14:32 |
| brtknr | i did: guestmount --ro ./fcos/ -m /dev/sda4 -a fedora-coreos-30.20190801.0-openstack.qcow2 | 14:34 |
| rpioso | etingof: There almost certainly is, although I don't know if it works with that referenced code. | 14:34 |
| rpioso | etingof: Why must boot mode be set when the boot source is set to one-time virtual CD? Isn't the boot mode already set? | 14:35 |
| etingof | rpioso, the code is taken literally from https://github.com/dell/iDRAC-Redfish-Scripting/blob/master/Redfish%20Python/SetNextOneTimeBootVirtualMediaDeviceOemREDFISH.py | 14:35 |
| TheJulia | brtknr: I didn't use guestfish, I did raw mount | 14:35 |
| TheJulia | with a loopback | 14:36 |
| etingof | rpioso, can virtual CD image contain legacy or uefi boot loader? | 14:36 |
| TheJulia | so, if it is doing the openstack metadata, then it should also be looking for the configuration drive | 14:36 |
| *** irclogbot_2 has joined #openstack-ironic | 14:37 | |
| TheJulia | brtknr: in your nova.conf, is force_config_drive set to true or false? | 14:38 |
| rpioso | etingof: I see! I expect the boot loader of the inserted media must match the boot mode of the server. Isn't it analogous to PXE booting? | 14:38 |
| etingof | rpioso, exactly! we know our CD image, so we need to set boot mode to the server accordingly. the question is - how? | 14:39 |
| *** tesseract has joined #openstack-ironic | 14:39 | |
| etingof | rpioso, I imagine we can do that in a standard way i.e. by patching redfish system tree. however that seemingly would require pushing a bios configuration job followed by a reboot | 14:40 |
| *** irclogbot_2 has quit IRC | 14:40 | |
| openstackgerrit | Merged openstack/ironic stable/stein: Do not wait for console being started on timeout https://review.opendev.org/679025 | 14:40 |
| etingof | rpioso, we do reboot while setting boot device anyway, so I am looking to set both things (boot device and boot mode) in one shot | 14:40 |
| rpioso | etingof: The standard way sounds workable. Sounds like you're asking about an optimization. | 14:41 |
| etingof | rpioso, well... each server reboot involving bios reconfiguration takes 5 minutes for me | 14:42 |
| rpioso | etingof: Sounds about right. | 14:42 |
| rpioso | etingof: It's a one-time thing, right? | 14:43 |
| etingof | rpioso, boot mode? well, every time ironic wants to change it we would have to go through this process | 14:43 |
| *** tesseract has quit IRC | 14:44 | |
| rpioso | etingof: Soon it'll only be UEFI :-) | 14:44 |
| rpioso | etingof: I don't see ironic wanting to change it very often (never?), once it's been set. | 14:45 |
| etingof | rpioso, however... can we submit two jobs in a row? because first we push a job to change boot device, then we push a job to change boot mode. do we need to reboot in-between? | 14:45 |
| brtknr | TheJulia: force_config_drive is not set | 14:46 |
| brtknr | in the conductor config | 14:46 |
| rpioso | etingof: Perhaps exporting the "Server Configuration Profile" (SCP) would provide info on what to add to the referenced code. | 14:46 |
| rpioso | etingof: No, there can only be one pending/scheduled BIOS config job. | 14:47 |
| brtknr | TheJulia: however, we've used config drive with centos images and it works fine... by enabling it at the time of instance creation | 14:47 |
| rpioso | etingof: If memory serves, exporting can be done via the iDRAC web GUI. | 14:48 |
| TheJulia | brtknr: if you look in ironic logs for the instance, do you see if ironic is acting upon a config drive for the instance deployment? Something seems super fishy | 14:48 |
| etingof | rpioso, does this mean that we can't really set one-time boot from virtual media? otherwise we lose that option upon first reboot (after setting boot device) by the time we reach second reboot (after setting boot mode)? | 14:49 |
| brtknr | TheJulia: would this be ironic-conductor? | 14:50 |
| rpioso | etingof: Does the "standard way" you mentioned above work? | 14:50 |
| *** irclogbot_2 has joined #openstack-ironic | 14:51 | |
| TheJulia | brtknr: yes | 14:51 |
| TheJulia | brtknr: do you see "neither config drive nor metadata service were available in time. Continuing without a config..." on the screen from fcos booting? | 14:51 |
| etingof | rpioso, I need to code that... I've found an ansible playbook doing exactly that so I suppose it should work | 14:51 |
| rpioso | etingof: I suggest confirming that works and then attempt to optimize it. | 14:52 |
| brtknr | TheJulia: there is no mention of config drive in the logs | 14:52 |
| brtknr | https://seashells.io/p/eVhrvxB3 | 14:53 |
| brtknr | TheJulia: oh wait, it was capital C - Configdrive | 14:53 |
| TheJulia | yup, I was just going to say... :) | 14:53 |
| *** irclogbot_2 has quit IRC | 14:54 | |
| TheJulia | So the drive is getting written, something is just broken upon reading it or getting the user-data in place. And if you deploy a VM it just works | 14:54 |
| etingof | rpioso, should we push set_boot_mode thing to ironic/drivers/modules/drac/management.py? | 14:54 |
| brtknr | `Configdrive for node 720d9d7d-e7d9-46e7-8346-7b0dae61dc36 successfully copied onto partition` | 14:55 |
| openstackgerrit | Varsha Verma proposed openstack/sushy-tools master: Add Storage and Storage Controllers resource support https://review.opendev.org/674339 | 14:56 |
| TheJulia | brtknr: hmmmm | 14:57 |
| brtknr | TheJulia: Hmm indeed... as far as I can see, I dont see any attempt to read config drive in the serial console | 14:58 |
| TheJulia | its logic is all hidden in a go binary... | 14:59 |
| TheJulia | :\ | 14:59 |
| TheJulia | the string config-2 is in there though | 14:59 |
| openstackgerrit | Varsha Verma proposed openstack/sushy-tools master: Add Storage and Storage Controllers resource support https://review.opendev.org/674339 | 15:05 |
| openstackgerrit | Varsha Verma proposed openstack/sushy-tools master: Add Drive resource support https://review.opendev.org/675085 | 15:05 |
| brtknr | TheJulia: Here's the output of my serial console if you can spot anything useful: https://seashells.io/p/5JWKWs9r | 15:06 |
| openstackgerrit | Julia Kreger proposed openstack/ironic-python-agent master: Attempt tgtd session detach https://review.opendev.org/679332 | 15:07 |
| TheJulia | it doesn't even look like ignition fires up | 15:09 |
| brtknr | TheJulia: thats what I noticed too | 15:10 |
| *** Goneri has joined #openstack-ironic | 15:11 | |
| openstackgerrit | Merged openstack/bifrost master: Make the opensuse-15 jobs non-voting https://review.opendev.org/679981 | 15:12 |
| TheJulia | brtknr: what deploy_interface are you using? and what versio of ironic? | 15:14 |
| *** tesseract has joined #openstack-ironic | 15:14 | |
| brtknr | We are using Ironic, is iPXE a deploy interface? | 15:16 |
| brtknr | Rocky ironic* | 15:17 |
| *** trandles has joined #openstack-ironic | 15:17 | |
| rpioso | etingof: No, we shouldn't. See https://github.com/rpioso/ironic/tree/story/2004592 | 15:17 |
| *** mkrai has quit IRC | 15:17 | |
| TheJulia | brtknr: direct or iscsi, pxe is a boot interface. | 15:18 |
| TheJulia | brtknr: openstack baremetal node show $node_uuid | 15:18 |
| TheJulia | *_interface fields | 15:18 |
| rpioso | etingof: It's the implementation of https://storyboard.openstack.org/#!/story/2004592 | 15:18 |
| TheJulia | we don't have any logic to handle os-trees... specifically | 15:18 |
| TheJulia | I'm kind of stumped. I feel like I'd need to take apart the coreos system while running | 15:19 |
| *** irclogbot_0 has joined #openstack-ironic | 15:19 | |
| rpioso | etingof: I'm pushing to submit it for review and inclusion in Train. | 15:19 |
| *** ddurst has quit IRC | 15:19 | |
| etingof | rpioso, afaik, you should get it merged by the end of the week | 15:20 |
| brtknr | TheJulia: iscsi | 15:20 |
| openstackgerrit | Varsha Verma proposed openstack/sushy-tools master: Add Volume resource support https://review.opendev.org/677119 | 15:20 |
| rpioso | etingof: Understood | 15:20 |
| *** irclogbot_0 has quit IRC | 15:22 | |
| etingof | rpioso, however, does that code support setting desired boot mode? | 15:22 |
| openstackgerrit | Julia Kreger proposed openstack/ironic-python-agent master: Don't try to put a bootloader in place when bootable https://review.opendev.org/679333 | 15:25 |
| iurygregory | dtantsur, about bifrost, its already switched to zuulv3, the only thing we should do is try to move to use bionic I would say | 15:28 |
| dtantsur | iurygregory: it has stuff in playbooks/legacy using zuul-cloner. we need to change that to new-style jobs. | 15:29 |
| dtantsur | and bionic, yes (OMG) | 15:29 |
| *** e0ne has quit IRC | 15:29 | |
| *** e0ne_ has joined #openstack-ironic | 15:29 | |
| iurygregory | yeah | 15:29 |
| rpioso | etingof: It declares support for a minimal derivations of the redfish h/w type's management and power interface implementations, e.g., https://github.com/rpioso/ironic/blob/7a738da3d9b4f73b00e8d855d1eb03afdc23e1dd/ironic/drivers/modules/drac/management.py#L286-L287 | 15:32 |
| *** e0ne_ is now known as e0ne | 15:33 | |
| *** irclogbot_1 has joined #openstack-ironic | 15:33 | |
| etingof | rpioso, I can see that set_boot_mode() is not implemented there. that means that we still need to add it. | 15:34 |
| *** ociuhandu has quit IRC | 15:34 | |
| rpioso | etingof: ? | 15:34 |
| *** rpioso is now known as rpioso|afk | 15:35 | |
| * rpioso|afk will be back | 15:35 | |
| etingof | rpioso|afk, this method is missing -- https://github.com/openstack/ironic/blob/master/ironic/drivers/base.py#L906 | 15:35 |
| *** irclogbot_1 has quit IRC | 15:36 | |
| TheJulia | brtknr: I'm moderately stumped. I'm wondering if boot setup is breaking it, but I can't be sure. I'm also in no position to really test at the moment :\ | 15:36 |
| etingof | rpioso|afk, w/o it ironic can't change node boot mode | 15:36 |
| *** irclogbot_0 has joined #openstack-ironic | 15:37 | |
| *** altlogbot_3 has joined #openstack-ironic | 15:39 | |
| tzumainn | TheJulia, hi! larsks and I had a question about the spec we're writing regarding multi-tenancy with respect to an owner and a lessee - would this be the right place to ask? | 15:43 |
| *** S4ren has joined #openstack-ironic | 15:48 | |
| dtantsur | tzumainn: yes | 15:48 |
| *** lucasagomes has quit IRC | 15:48 | |
| dtantsur | (if you mean ironic, of course :) | 15:48 |
| tzumainn | hahaha | 15:49 |
| TheJulia | dtantsur: thanks :) | 15:49 |
| tzumainn | okay, so I think there may be a fairly simple solution, which involves oslo policy | 15:49 |
| brtknr | TheJulia: thanks for looking, I will park this for now | 15:49 |
| TheJulia | (The goal being to take over the world of course) | 15:49 |
| TheJulia | tzumainn: I was kind of expecting that from when we originally talked about it. | 15:50 |
| dtantsur | tzumainn: involving oslo.policy is probably the right thing to do, although I'm not sure if it'll help with filtering node listing or with allocations | 15:50 |
| tzumainn | basically, when ironic does policy.authorize - for example, https://github.com/openstack/ironic/blob/master/ironic/api/controllers/v1/node.py#L225 - the first cdict argument is actually a 'target' and can be used in a policy rule | 15:50 |
| etingof | rpioso|afk, it seems importing bios configuration into iDRAC fails without any explanation if configuration XML has insignificant whitespaces | 15:51 |
| tzumainn | okay, so my question was regarding the spec - TheJulia, you suggested just writing one spec; however, it seems that there could be an initial spec that just covers an owner, since that would require no db changes | 15:51 |
| tzumainn | and then a follow-up spec that covers the lessee, since that does involve a database change, and there could be complexities depending on how flexible we wanted that side of things to be | 15:52 |
| tzumainn | do you think it might be worth splitting the spec up like that? | 15:52 |
| dtantsur | I'd definitely split the owner and the lessee work | 15:52 |
| TheJulia | well, owner is already a field | 15:52 |
| dtantsur | the former is much simpler, the latter may end up attracting opposite opinions | 15:52 |
| tzumainn | dtantsur, yeah, that was my fear | 15:52 |
| dtantsur | TheJulia: I guess that's what tzumainn is saying: we have a field, we just need to wire it in | 15:52 |
| TheJulia | this is true, but I'm worried about the big picture not being percievable without them being viewed together | 15:53 |
| dtantsur | TheJulia: isn't working ownership quite useful on its own? | 15:53 |
| tzumainn | so this way we separate the policy discussion from how we might model the lessee | 15:53 |
| TheJulia | dtantsur: yeah, and determine how we handle defaults and all | 15:53 |
| TheJulia | maybe that does mean a separate spec migh tbe best | 15:53 |
| TheJulia | lessee or "allocated to" kind of field | 15:54 |
| *** sthussey has joined #openstack-ironic | 15:54 | |
| tzumainn | okay! I'll start with the owner/policy spec then and get that submitted for discussion first | 15:54 |
| tzumainn | thanks! | 15:54 |
| TheJulia | tzumainn: okay, thanks! | 15:55 |
| *** priteau has quit IRC | 15:55 | |
| dtantsur | also, nice to see you again, tzumainn, that's been a while :) | 15:55 |
| * dtantsur has to run now | 15:55 | |
| tzumainn | dtantsur, hahaha, yeah, it's been a bit : ) | 15:55 |
| *** dtantsur is now known as dtantsur|afk | 15:55 | |
| dtantsur|afk | have a good rest of the day folks | 15:55 |
| *** alexmcleod has quit IRC | 15:58 | |
| etingof | rpioso|afk, https://github.com/etingof/sushy-oem-dellemc/blob/master/sushy_oem_dellemc/resources/manager/manager.py#L45 | 15:58 |
| TheJulia | goodnight dtantsur|afk | 15:59 |
| *** gyee has joined #openstack-ironic | 15:59 | |
| *** ociuhandu has joined #openstack-ironic | 16:06 | |
| *** rachit7 has quit IRC | 16:08 | |
| *** ociuhandu has quit IRC | 16:11 | |
| arne_wiebalck | bye everyone o/ (tmrw is a holiday here, so back on Friday) | 16:15 |
| iurygregory | bye arne_wiebalck o/ | 16:16 |
| *** e0ne has quit IRC | 16:31 | |
| *** S4ren has quit IRC | 16:31 | |
| -openstackstatus- NOTICE: Zuul job logs stored in OVH may fail. We have updated the base job to remove OVH from our storage location. If you have POST_FAILURES a recheck should fix them at this point. | 16:52 | |
| *** dougsz has quit IRC | 16:57 | |
| *** derekh has quit IRC | 17:00 | |
| *** ociuhandu has joined #openstack-ironic | 17:22 | |
| *** ociuhandu has quit IRC | 17:26 | |
| *** jtwill98 has joined #openstack-ironic | 17:28 | |
| jtwill98 | Do you know if someone has recently updated: https://docs.openstack.org/sushy-tools/latest/user/dynamic-emulator.html#uefi-boot | 17:30 |
| jtwill98 | I don't recall it having the SimpleStorage section. | 17:30 |
| *** rpioso|afk is now known as rpioso | 17:33 | |
| rpioso | jtwill98: https://opendev.org/openstack/sushy-tools/commits/branch/master | 17:33 |
| *** ociuhandu has joined #openstack-ironic | 17:38 | |
| *** ociuhandu has quit IRC | 17:42 | |
| rpioso | etingof: | 17:44 |
| *** bdodd has joined #openstack-ironic | 17:50 | |
| rpioso | etingof: Doesn't the idrac h/w type's new DracRedfishManagement (entrypoint idrac-redfish) management interface implementation picks up that method via derivation from the standard Redfish's implementation, DracRedfishManagement. | 17:51 |
| *** ociuhandu has joined #openstack-ironic | 17:54 | |
| *** ociuhandu has quit IRC | 17:58 | |
| *** tesseract has quit IRC | 18:01 | |
| *** iurygregory has quit IRC | 18:08 | |
| *** iurygregory has joined #openstack-ironic | 18:09 | |
| etingof | rpioso, what would happen to set_boot_mode() call if a configuration job has been submitted by some other module (e.g. set_boot_device())? | 18:22 |
| rpioso | etingof: Please answer ^^ | 18:33 |
| etingof | rpioso, would intermediate reboot upon submitting a configuration job help? I think I am doing that in sushy-oem-dell... | 18:36 |
| *** hjensas has joined #openstack-ironic | 18:43 | |
| *** ricolin has quit IRC | 18:50 | |
| *** irclogbot_0 has quit IRC | 18:54 | |
| *** irclogbot_0 has joined #openstack-ironic | 18:55 | |
| *** trandles has quit IRC | 18:57 | |
| *** e0ne has joined #openstack-ironic | 18:59 | |
| *** trandles_ has joined #openstack-ironic | 18:59 | |
| *** irclogbot_0 has quit IRC | 19:02 | |
| *** irclogbot_0 has joined #openstack-ironic | 19:03 | |
| rpioso | etingof: Doesn't the idrac h/w type's new DracRedfishManagement (entrypoint idrac-redfish) management interface implementation pick up that method via derivation from the standard Redfish implementation, DracRedfishManagement? | 19:12 |
| etingof | rpioso, I do not see if DracRedfishManagement is any different from RedfishManagement, which does not do any OEM stuff like intermediate reboots. Let me know if I am missing something. | 19:41 |
| etingof | rpioso, my understanding is that, with current iDRAC, set-boot-device-to-vmedia and set-boot-mode calls must be separated by reboots or the second one will surely fail | 19:42 |
| etingof | rpioso, to avoid running into pending configuration job, may be we should add some job queue checking prior to submitting new jobs or just handle specific OEM error that indicates "queue full" condition and retry. in sushy-oem-dell I am trying out the second approach | 19:46 |
| *** ociuhandu has joined #openstack-ironic | 19:47 | |
| etingof | rpioso, conceptually, I think it would be much better to hide all these implementation details behind BMC's Redfish agent. I hope we will get there, eventually | 19:51 |
| *** jtomasek has quit IRC | 20:08 | |
| *** ociuhandu has quit IRC | 20:09 | |
| *** ociuhandu has joined #openstack-ironic | 20:10 | |
| *** trandles_ has quit IRC | 20:16 | |
| *** trandles has joined #openstack-ironic | 20:20 | |
| rpioso | etingof: Correct, DracRedfishManagement is presently functionally equivalent to RedfishManagement. Therefore, it has https://github.com/openstack/ironic/blob/master/ironic/drivers/base.py#L906 | 20:22 |
| *** trandles has quit IRC | 20:24 | |
| *** trandles has joined #openstack-ironic | 20:27 | |
| *** e0ne has quit IRC | 20:37 | |
| *** ociuhandu has quit IRC | 20:38 | |
| *** e0ne has joined #openstack-ironic | 20:48 | |
| openstackgerrit | OpenStack Release Bot proposed openstack/metalsmith stable/train: Update .gitreview for stable/train https://review.opendev.org/680218 | 20:54 |
| openstackgerrit | OpenStack Release Bot proposed openstack/metalsmith stable/train: Update TOX/UPPER_CONSTRAINTS_FILE for stable/train https://review.opendev.org/680219 | 20:54 |
| openstackgerrit | OpenStack Release Bot proposed openstack/metalsmith master: Update master for stable/train https://review.opendev.org/680220 | 20:54 |
| *** e0ne has quit IRC | 20:59 | |
| *** bfournie has quit IRC | 21:20 | |
| *** ddurst has joined #openstack-ironic | 21:39 | |
| *** dsneddon_ is now known as dsneddon | 21:52 | |
| *** iurygregory has quit IRC | 21:58 | |
| *** trident has quit IRC | 22:05 | |
| *** trident has joined #openstack-ironic | 22:14 | |
| *** jhesketh has quit IRC | 22:27 | |
| *** jhesketh has joined #openstack-ironic | 22:33 | |
| *** bfournie has joined #openstack-ironic | 22:39 | |
| *** rcernin has joined #openstack-ironic | 23:04 | |
| *** bdodd has quit IRC | 23:14 | |
| openstackgerrit | Michael Davies proposed openstack/ironic-python-agent-builder master: Allow extra python pakgs into tinyipa https://review.opendev.org/679735 | 23:51 |
| *** trident has quit IRC | 23:52 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!