| *** Goneri has quit IRC | 00:19 | |
| *** igordc has quit IRC | 00:42 | |
| *** kaifeng|afk is now known as kaifeng | 00:45 | |
| kaifeng | TheJulia: ipxe is not mandatory i guess? my current issue with vm is that after pxe failed, ifconfig -l shows the interface didn't get an IP address, but -r manually and after a few seconds it will do | 00:47 |
|---|---|---|
| TheJulia | kaifeng: i mean, you could use grub, where are you not seeing an interface ip address? | 00:49 |
| kaifeng | however during pxe, there are dhcpoffer/ack logs at dnsmasq on the host side. i will try it on a real machine. | 00:49 |
| kaifeng | TheJulia: morning :) | 00:49 |
| kaifeng | after pxe failed, the vm runs to the uefi shell, that's where i see the eth0 has no address | 00:50 |
| TheJulia | kaifeng: oh, yeah... that shell has some weird default constraints which includes no networking | 00:51 |
| kaifeng | so i need to do some thing in the uefi firmware? | 00:51 |
| *** ociuhandu has joined #openstack-ironic | 01:21 | |
| *** ociuhandu has quit IRC | 01:26 | |
| *** ociuhandu has joined #openstack-ironic | 01:32 | |
| *** ociuhandu has quit IRC | 01:37 | |
| *** ricolin has joined #openstack-ironic | 01:52 | |
| *** bobmel has joined #openstack-ironic | 02:24 | |
| *** bobmel has quit IRC | 02:29 | |
| *** trungnv has joined #openstack-ironic | 02:52 | |
| *** chenhaw has quit IRC | 03:00 | |
| *** goldyfruit_ has joined #openstack-ironic | 03:02 | |
| *** ociuhandu has joined #openstack-ironic | 03:04 | |
| *** tzumainn has quit IRC | 03:06 | |
| *** ociuhandu has quit IRC | 03:10 | |
| *** mkrai has joined #openstack-ironic | 03:17 | |
| *** trungnv has quit IRC | 03:18 | |
| *** rloo has quit IRC | 03:35 | |
| *** goldyfruit_ has quit IRC | 04:01 | |
| *** ociuhandu has joined #openstack-ironic | 04:05 | |
| *** ociuhandu has quit IRC | 04:10 | |
| *** hwoarang has quit IRC | 04:26 | |
| *** ociuhandu has joined #openstack-ironic | 05:08 | |
| *** ociuhandu has quit IRC | 05:13 | |
| mkrai | gm ironic o/ | 05:25 |
| *** ociuhandu has joined #openstack-ironic | 05:30 | |
| *** ociuhandu has quit IRC | 05:35 | |
| *** threestrands has quit IRC | 05:51 | |
| mkrai | Hi, I tried to boot a baremetal server and it was succesful but I can't ping or ssh the node. | 06:18 |
| mkrai | Can someone help debug this issue? | 06:19 |
| mkrai | I am using networking-baremetal | 06:19 |
| mkrai | hjensas, Hi o/ | 06:24 |
| *** chenhaw has joined #openstack-ironic | 06:27 | |
| *** verdurin has quit IRC | 06:34 | |
| *** pcaruana has joined #openstack-ironic | 06:34 | |
| *** jtomasek has joined #openstack-ironic | 06:39 | |
| *** verdurin has joined #openstack-ironic | 06:47 | |
| *** hjensas has quit IRC | 07:00 | |
| openstackgerrit | Madhuri Kumari proposed openstack/ironic-ui master: Drop python 2.7 support and testing https://review.opendev.org/695801 | 07:01 |
| *** ociuhandu has joined #openstack-ironic | 07:20 | |
| *** ociuhandu has quit IRC | 07:20 | |
| *** ociuhandu has joined #openstack-ironic | 07:21 | |
| arne_wiebalck | Good morning, ironic! | 07:24 |
| *** ociuhandu has quit IRC | 07:26 | |
| mkrai | arne_wiebalck, gm | 07:27 |
| mkrai | Ironic gate is broken it seems | 07:28 |
| mkrai | I am working on it | 07:28 |
| arne_wiebalck | Hey mkrai o/ | 07:28 |
| mkrai | arne_wiebalck, Did you see the pep8 failures? | 07:30 |
| arne_wiebalck | mkrai: for which change? | 07:30 |
| arne_wiebalck | mkrai: "no" is the answer to your question ;) | 07:31 |
| mkrai | ironic gate is broken due to pep8 failure | 07:31 |
| mkrai | all the patches has -1 | 07:31 |
| mkrai | arne_wiebalck, other question, how the ironic node get the network config, by config-drive or the metadata service? | 07:32 |
| arne_wiebalck | mkrai: tbh, I am somewhat out of my depth when it comes to ironic networking (we are completely bypassing this in our deployment) ... but from what I understand it would work in a similar fashion as it does for virtual instances, e.g. with neutron creating the ports and networking-baremetal providing the corresponding integration | 07:41 |
| arne_wiebalck | hjensas mgoddard dtantsur|afk TheJulia can certainly give more insights | 07:42 |
| mkrai | arne_wiebalck, Ok thanks | 07:42 |
| mkrai | arne_wiebalck, I have provisioned a node successfully but can't login | 07:42 |
| mkrai | so debugging the issue | 07:43 |
| arne_wiebalck | mkrai: so the node is up? | 07:43 |
| mkrai | yes | 07:43 |
| mkrai | arne_wiebalck, but can't ping or login | 07:43 |
| *** belmoreira has joined #openstack-ironic | 07:43 | |
| arne_wiebalck | mkrai: can you get in via the console? | 07:43 |
| arne_wiebalck | mkrai: I had this exact situation the other day | 07:44 |
| mkrai | arne_wiebalck, Yes with BMC console | 07:44 |
| arne_wiebalck | mkrai: right, so you can log in? | 07:44 |
| mkrai | arne_wiebalck, I used centos image, it has no default login. So I can't | 07:45 |
| mkrai | I am going to try cirros image | 07:45 |
| arne_wiebalck | mkrai: wait | 07:45 |
| mkrai | arne_wiebalck, Ok | 07:45 |
| arne_wiebalck | mkrai: you can add some user data to configure the log in | 07:46 |
| arne_wiebalck | mkrai: this is what I did the other dayx | 07:46 |
| mkrai | arne_wiebalck, Ok I will check that first | 07:46 |
| arne_wiebalck | mkrai: this way you can stick with the centos image which I guess is what you want to use eventually | 07:47 |
| mkrai | arne_wiebalck, Yes | 07:47 |
| arne_wiebalck | mkrai: look for cloud-init password | 07:48 |
| arne_wiebalck | mkrai: sth like http://paste.openstack.org/show/786770/ | 07:50 |
| mkrai | ok use the same file with --user-data option? | 07:51 |
| mkrai | arne_wiebalck, ^ | 07:51 |
| arne_wiebalck | mkrai: this was for ubuntu (and the ubuntu user which is privileged) | 07:51 |
| arne_wiebalck | mkrai: I guess you need to do this for the root user on centos | 07:52 |
| arne_wiebalck | mkrai: but, yes, pass such a file to the --user-data option when creating the instance | 07:52 |
| mkrai | arne_wiebalck, Ok I am trying it | 07:52 |
| arne_wiebalck | mkrai: or you pass an ssh key and use the centos user (and sudo) ... maybe worth to try both :) | 07:53 |
| mkrai | arne_wiebalck, the ssh key is passed using keypair already | 07:54 |
| arne_wiebalck | mkrai: yeah, but you don't have network (I forgot) | 07:55 |
| mkrai | arne_wiebalck, yes ;) | 07:55 |
| arne_wiebalck | mkrai: I'd go for cloud-init | 07:56 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add `network_data` field to ports and portgroups https://review.opendev.org/687910 | 07:58 |
| arne_wiebalck | mkrai: I'd think sth like http://paste.openstack.org/show/786771/ should do it | 08:00 |
| mkrai | arne_wiebalck, Yes I am using similar file. Thanks :) | 08:00 |
| *** tesseract has joined #openstack-ironic | 08:04 | |
| *** iurygregory has joined #openstack-ironic | 08:08 | |
| iurygregory | morning Ironic | 08:09 |
| kaifeng | good morning o/ | 08:25 |
| kaifeng | TheJulia: i think the issue i have is related with the firmware shipped with ubuntu, have tested on phsical arm64 machine and worked. | 08:26 |
| *** ociuhandu has joined #openstack-ironic | 08:28 | |
| kaifeng | mkrai: if you are using dib with cloudimage, there are two elements: devuser and dynamic-login can be useful in diagnosing. | 08:36 |
| openstackgerrit | raphael.glon proposed openstack/ironic master: Feature: finalize_rescue, send sha512 hashed password, not clear ones https://review.opendev.org/695649 | 08:40 |
| *** tssurya has joined #openstack-ironic | 08:41 | |
| *** hjensas has joined #openstack-ironic | 08:46 | |
| mkrai | kaifeng, I am using the centos cloud image | 08:52 |
| *** ociuhandu has quit IRC | 08:52 | |
| *** bobmel has joined #openstack-ironic | 08:52 | |
| *** ociuhandu has joined #openstack-ironic | 08:52 | |
| *** lucasagomes has joined #openstack-ironic | 08:55 | |
| *** ociuhandu has quit IRC | 08:57 | |
| *** rpittau|afk is now known as rpittau | 08:57 | |
| rpittau | good morning ironic! o/ | 08:57 |
| rpittau | happy hump day! | 08:57 |
| *** ociuhandu has joined #openstack-ironic | 08:58 | |
| iurygregory | morning rpittau kaifeng mkrai o/ | 08:58 |
| mkrai | iurygregory, o/ | 08:58 |
| rpittau | hey iurygregory :) | 08:59 |
| iurygregory | I hate the CI failures on my patch .-. I'm starting to think is related to pyghmi =X | 09:03 |
| * etingof is yet to encounter folks enjoying CI failures | 09:07 | |
| *** ociuhandu has quit IRC | 09:08 | |
| *** dougsz has joined #openstack-ironic | 09:12 | |
| kaifeng | iurygregory :) | 09:12 |
| openstackgerrit | raphael.glon proposed openstack/ironic-python-agent-builder master: Add dosfstools to tinyipa https://review.opendev.org/696180 | 09:13 |
| rpittau | iurygregory: do you have any logs at hand ? | 09:20 |
| iurygregory | rpittau, nvm the ones on ipmi jobs are basically a network issue Err:1 http://mirror.mtl01.inap.opendev.org/ubuntu bionic/main amd64 libharfbuzz0b amd64 1.7.2-1ubuntu1 | 09:22 |
| iurygregory | 404 Not Found [IP: 198.72.125.4 80] | 09:22 |
| iurygregory | but standalone is a little crazy with the random failures | 09:23 |
| rpittau | iurygregory: ok, I was starting to feel guilty as I vouched for pyghmi being py3 ready :D | 09:23 |
| iurygregory | the failures in standalone are fun ... http://zuul.openstack.org/builds?job_name=ironic-standalone | 09:23 |
| iurygregory | the best part is that is different tests that are failling everytime bassicaly yay | 09:24 |
| rpittau | looks like 50-50 ? roughly ? | 09:24 |
| iurygregory | nope | 09:26 |
| iurygregory | i saw one with 8/14 | 09:26 |
| *** mkowalski has quit IRC | 09:27 | |
| *** csmart has quit IRC | 09:27 | |
| *** ajya[m] has quit IRC | 09:27 | |
| iurygregory | oh Im not looking the Count Pass Fail Error Skip values only the amount of "rows" | 09:28 |
| iurygregory | and also there are some network issues on the CI | 09:28 |
| iurygregory | Err:1 http://mirror.mtl01.inap.opendev.org/ubuntu bionic/main amd64 libharfbuzz0b amd64 1.7.2-1ubuntu1 404 Not Found [IP: 198.72.125.4 80] | 09:29 |
| *** openstackstatus has quit IRC | 09:30 | |
| rpittau | well that doesn't help | 09:30 |
| iurygregory | yeah | 09:30 |
| iurygregory | I just asked infra, since I saw more than 3 times already | 09:30 |
| iurygregory | probably something wrong on a specific server we are using I would say | 09:31 |
| *** ociuhandu has joined #openstack-ironic | 09:31 | |
| iurygregory | ok, yesterday there was a problem and it's already fixed, now we just need to pray for the tempest tests XD | 09:36 |
| *** derekh has joined #openstack-ironic | 09:39 | |
| * rpittau needs coffee | 09:53 | |
| *** alexmcleod_ has joined #openstack-ironic | 09:58 | |
| *** alexmcleod has quit IRC | 09:59 | |
| etingof | iurygregory's prayer seems successful \o/ | 10:01 |
| iurygregory | etingof, I will only consider successful after the patch merges XD | 10:02 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add `get_node_network_data` method to NetworkInterface https://review.opendev.org/688184 | 10:03 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add `get_node_network_data` to non-Neutron NetworkInterface https://review.opendev.org/688266 | 10:04 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Make redfish boot interface using network interface metadata https://review.opendev.org/688272 | 10:04 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add `get_node_network_data` to Neutron NetworkInterface https://review.opendev.org/688896 | 10:04 |
| openstackgerrit | Ilya Etingof proposed openstack/ironic master: Add L3 boot section to the docs https://review.opendev.org/689844 | 10:04 |
| * etingof is asking too much from mighty Zuul | 10:05 | |
| *** rcernin has quit IRC | 10:31 | |
| *** ociuhandu has quit IRC | 10:34 | |
| *** ociuhandu has joined #openstack-ironic | 10:34 | |
| openstackgerrit | Arne Wiebalck proposed openstack/ironic-inspector master: Get details when listing ports upon inspection of active nodes https://review.opendev.org/696133 | 10:37 |
| *** ociuhandu has quit IRC | 10:39 | |
| *** rcernin has joined #openstack-ironic | 10:46 | |
| *** csmart has joined #openstack-ironic | 10:46 | |
| *** mkowalski has joined #openstack-ironic | 10:46 | |
| *** ajya[m] has joined #openstack-ironic | 10:46 | |
| *** sshnaidm|afk is now known as sshnaidm | 10:47 | |
| *** tkajinam has joined #openstack-ironic | 10:50 | |
| *** dtantsur|afk is now known as dtantsur | 10:56 | |
| mgoddard | mkrai: did you fix your network issue? Does the node's port have a physical_network defined? That is necessary for the flat network driver. It needs to match the physical network of the neutron net you're attaching to | 10:58 |
| *** ociuhandu has joined #openstack-ironic | 10:59 | |
| mkrai | mgoddard, Hi, yes it has | 10:59 |
| mgoddard | mkrai: has the port been bound? | 11:00 |
| mkrai | mgoddard, yes it is | 11:00 |
| mgoddard | mkrai: can you ping? | 11:00 |
| mgoddard | mkrai: maybe your image needs dhcp-all-interfaces DIB element? | 11:01 |
| mkrai | mgoddard, No I can't ping nor ssh | 11:01 |
| mkrai | mgoddard, i am using centos cloud image | 11:01 |
| mkrai | as my guest OS image | 11:01 |
| arne_wiebalck | rpittau: dtantsur: TheJulia: We managed to manually inspect an instance from within a container \o/ | 11:02 |
| mgoddard | mkrai: if you have multiple NICs then cloud-init might choose the wrong one to bring up. You can build an image using DIB with the dhcp-all-interfaces element to try DHCP on all interfaces | 11:02 |
| dtantsur | arne_wiebalck: w00t!! any hacks or workarounds? | 11:03 |
| arne_wiebalck | dtantsur: yes | 11:03 |
| rpittau | arne_wiebalck: great! | 11:03 |
| arne_wiebalck | dtantsur: rpittau: it failed b/c of the pxe_enabled field in the ports table | 11:03 |
| arne_wiebalck | dtantsur: rpittau: it was apparently set by the initial inspection (to 1) | 11:04 |
| arne_wiebalck | dtantsur: rpittau: but on the instance it is 0 | 11:04 |
| *** ociuhandu has quit IRC | 11:04 | |
| arne_wiebalck | https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L2683 raises an exception | 11:04 |
| rpittau | mmm and just stayed there | 11:04 |
| dtantsur | arne_wiebalck: I lost a part of scrollback, could you repeat everything after "yes"? | 11:04 |
| dtantsur | sorry | 11:04 |
| arne_wiebalck | it failed b/c of the pxe_enabled field in the ports table | 11:05 |
| arne_wiebalck | it was apparently set by the initial inspection (to 1) | 11:05 |
| arne_wiebalck | but on the instance it is 0 | 11:05 |
| arne_wiebalck | https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L2683 raises an exception | 11:05 |
| dtantsur | hmmm | 11:05 |
| arne_wiebalck | dtantsur: ^ that's all | 11:05 |
| arne_wiebalck | :) | 11:05 |
| dtantsur | I see. Interesting. | 11:05 |
| dtantsur | We should disable pxe_enabled handling for active nodes introspection, it makes no sense. | 11:05 |
| dtantsur | maybe even: if we don't have ANY PXE information, don't try to touch it | 11:06 |
| arne_wiebalck | dtantsur: not sureif the node does not get it or does not have it ... | 11:07 |
| arne_wiebalck | dtantsur: we can see if we can figure that out | 11:07 |
| dtantsur | arne_wiebalck: PXE information is based on a certain kernel parameter that our PXE templates set | 11:08 |
| dtantsur | I don't think it's possible to fetch it from a normal running instance | 11:08 |
| arne_wiebalck | hmm, so you think the manual inspection simply does not have that info at all? | 11:08 |
| arne_wiebalck | and uses 0 as a default? | 11:08 |
| dtantsur | I think so, yes | 11:09 |
| dtantsur | I need to check the inspector code to be sure | 11:09 |
| arne_wiebalck | does it make sense to update the port info a all for active instances? | 11:09 |
| arne_wiebalck | at all | 11:09 |
| arne_wiebalck | I mean any port info. | 11:09 |
| mkrai | mgoddard, Ok I will try that. Thanks! | 11:10 |
| dtantsur | arne_wiebalck: this is a good question | 11:10 |
| *** ociuhandu has joined #openstack-ironic | 11:10 | |
| dtantsur | for the reference, that's the problematic place: https://opendev.org/openstack/ironic-inspector/src/branch/master/ironic_inspector/plugins/standard.py#L298-L299 | 11:10 |
| arne_wiebalck | dtantsur: the code allows for MAC changes of active nodes in maintenance | 11:10 |
| arne_wiebalck | dtantsur: that's probably h/w replacements? | 11:11 |
| dtantsur | we won't be able to change MAC like this because we find ports by MAC | 11:11 |
| dtantsur | I mean.. we may try to delete a port and create it anew | 11:11 |
| dtantsur | mmmmm | 11:11 |
| dtantsur | I think we should avoid updating pxe_enabled or deleting ports. I'm not so sure about creating ports and updating LLDP information. | 11:13 |
| dtantsur | arne_wiebalck: ^^ | 11:13 |
| arne_wiebalck | dtantsur: hmm, yeah | 11:13 |
| arne_wiebalck | dtantsur: esp if we do not have valid information :-D | 11:14 |
| openstackgerrit | raphael.glon proposed openstack/ironic-python-agent master: Software RAID: make it bootable with uefi https://review.opendev.org/696156 | 11:14 |
| dtantsur | right | 11:14 |
| *** ociuhandu has quit IRC | 11:15 | |
| arne_wiebalck | dtantsur: ok, let me wrap this up in a story as a starting point for discussion how this should be handled ... sounds ok? | 11:16 |
| rpittau | in general I also think it doesn't make sense to have active inspection updates port info, or handle that anyway | 11:16 |
| arne_wiebalck | rpittau: I guess there is a reason this makes sense as we explicitly allow this | 11:17 |
| dtantsur | welll.. | 11:17 |
| dtantsur | I think we just did not think about it :) | 11:17 |
| rpittau | heh :) | 11:17 |
| dtantsur | but I can see value in updating LLDP information | 11:17 |
| arne_wiebalck | ha! :) | 11:17 |
| dtantsur | assuming we can collect LLDP from a container on a running machine... | 11:18 |
| arne_wiebalck | https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L2659 | 11:18 |
| dtantsur | well, yes | 11:18 |
| dtantsur | I think we should short-circuit this logic after https://opendev.org/openstack/ironic-inspector/src/branch/master/ironic_inspector/plugins/standard.py#L268 | 11:19 |
| arne_wiebalck | dtantsur: done by you even :) | 11:19 |
| * dtantsur looks around innocently | 11:19 | |
| arne_wiebalck | lol | 11:20 |
| arne_wiebalck | dtantsur: you mean short circuit for active nodes? | 11:20 |
| dtantsur | exactly | 11:20 |
| *** jtomasek has quit IRC | 11:23 | |
| *** ociuhandu has joined #openstack-ironic | 11:24 | |
| *** jtomasek has joined #openstack-ironic | 11:25 | |
| arne_wiebalck | dtantsur: ok, we'll give that a try | 11:29 |
| arne_wiebalck | dtantsur: thanks! | 11:29 |
| *** rcernin has quit IRC | 11:31 | |
| arne_wiebalck | dtantsur: https://storyboard.openstack.org/#!/story/2006956 not sure if this is a bug or an RFE ... | 11:39 |
| dtantsur | arne_wiebalck: I think it's a bug | 11:40 |
| dtantsur | etingof: hey, how far are we from redfish virtual media working on devstack? | 11:46 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Implement managed in-band inspection boot for redfish-virtual-media https://review.opendev.org/696291 | 11:48 |
| dtantsur | etingof: I would like to test ^^^ somehow | 11:48 |
| *** mkrai has quit IRC | 11:51 | |
| *** tzumainn has joined #openstack-ironic | 11:53 | |
| *** tkajinam has quit IRC | 12:11 | |
| openstackgerrit | Merged openstack/sushy-tools master: Fix defaults to storage and drive resources https://review.opendev.org/687539 | 12:11 |
| openstackgerrit | Merged openstack/sushy-tools master: Fix emulator socket binding options https://review.opendev.org/684685 | 12:11 |
| *** sshnaidm is now known as sshnaidm|afk | 12:14 | |
| *** tkajinam has joined #openstack-ironic | 12:15 | |
| *** tkajinam has quit IRC | 12:16 | |
| dtantsur | I guess we at least need all this ^^ | 12:20 |
| openstackgerrit | Merged openstack/sushy-tools master: Fix `BootSourceOverrideMode` enumeration https://review.opendev.org/688458 | 12:26 |
| *** ociuhandu has quit IRC | 12:32 | |
| *** ociuhandu has joined #openstack-ironic | 12:33 | |
| *** ociuhandu has quit IRC | 12:37 | |
| *** dtantsur is now known as dtantsur|brb | 12:48 | |
| etingof | dtantsur|brb, devstack is the best place to exercise virtual media boot! | 12:55 |
| etingof | dtantsur|brb, once devstack understands your intention to do vmedia boot, it should set up everything for you | 12:57 |
| *** rh-jelabarre has joined #openstack-ironic | 13:03 | |
| openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: Fix libvirt CD image insertion https://review.opendev.org/693561 | 13:07 |
| openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: Always log via Flask application https://review.opendev.org/693639 | 13:08 |
| *** iurygregory has left #openstack-ironic | 13:11 | |
| *** iurygregory has joined #openstack-ironic | 13:11 | |
| *** rloo has joined #openstack-ironic | 13:38 | |
| *** Goneri has joined #openstack-ironic | 13:58 | |
| *** gryf has quit IRC | 14:01 | |
| openstackgerrit | Merged openstack/ironic master: Drop python 2.7 support and testing https://review.opendev.org/695213 | 14:10 |
| iurygregory | YAY | 14:15 |
| iurygregory | etingof, my prayer was successful \o/ | 14:27 |
| *** ociuhandu has joined #openstack-ironic | 14:28 | |
| etingof | could you pray for my patch please, iurygregory | 14:30 |
| iurygregory | etingof, doing now | 14:32 |
| iurygregory | its only one or the whole chain? | 14:33 |
| etingof | let it be a Mass | 14:36 |
| *** ociuhandu has quit IRC | 14:40 | |
| *** priteau has joined #openstack-ironic | 14:50 | |
| rpittau | hey iurygregory, did you include also the ironic-staging-drivers in the py27 dropping patches ? | 14:50 |
| *** sshnaidm|afk is now known as sshnaidm | 14:52 | |
| *** goldyfruit_ has joined #openstack-ironic | 14:55 | |
| iurygregory | rpittau, I can =) | 15:00 |
| * iurygregory didn't found this repo on governance under ironic .-. | 15:00 | |
| *** ociuhandu has joined #openstack-ironic | 15:07 | |
| TheJulia | it is very much not under our governance, but we can be kind and maybe fix drivers? | 15:11 |
| *** belmoreira has quit IRC | 15:13 | |
| *** sziviani has quit IRC | 15:17 | |
| *** gudrutis has quit IRC | 15:17 | |
| iurygregory | Done =) https://review.opendev.org/696322 | 15:19 |
| patchbot | patch 696322 - x/ironic-staging-drivers - Drop python 2.7 support and testing - 1 patch set | 15:19 |
| *** hjensas has quit IRC | 15:22 | |
| *** ricolin_ has joined #openstack-ironic | 15:26 | |
| *** belmoreira has joined #openstack-ironic | 15:26 | |
| *** dtantsur|brb is now known as dtantsur | 15:26 | |
| dtantsur | etingof: right, but I see there are sushy-tools patches to merge, right? | 15:27 |
| etingof | dtantsur, you can try it with sushy-tools master. the only potentially useful patch is this -- https://review.opendev.org/#/c/693561/ | 15:28 |
| patchbot | patch 693561 - sushy-tools - Fix libvirt CD image insertion - 10 patch sets | 15:28 |
| *** ricolin has quit IRC | 15:28 | |
| dtantsur | okay, let's see if we can merge them soon | 15:28 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Add an ironic-inspector job with managed boot https://review.opendev.org/696117 | 15:28 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Prevent localhost from being used as ironic-inspector callback URL https://review.opendev.org/693564 | 15:28 |
| dtantsur | oh, yes, we can merge this one | 15:29 |
| etingof | \o/ | 15:29 |
| *** ricolin_ is now known as ricolin | 15:29 | |
| *** ociuhandu has quit IRC | 15:32 | |
| *** ociuhandu has joined #openstack-ironic | 15:32 | |
| *** ricolin has quit IRC | 15:34 | |
| dtantsur | etingof: I'm starting wondering if we can move the L3 spec faster if we schedule a call between interested parties to go over the most contentious bits | 15:36 |
| etingof | meeting++ | 15:36 |
| * etingof can doodle the meeting | 15:38 | |
| iurygregory | TheJulia, your idea to verify if the bits on the disk image would be using hexdump? | 15:39 |
| etingof | dtantsur, may be we can raise L3 thing at ironic upstream mtg? | 15:40 |
| TheJulia | iurygregory: for master boot records, and it does work fun enough. For UEFI we'd likely need to look for a boot file system and then check contents. The conundrum is there are standard locations for the bootloader and such. Then some OSes do stupid things like hardcode different paths into the binaries | 15:42 |
| TheJulia | On a plus side, we only need to get the initial loader right, and it will be in the filesystem... | 15:42 |
| *** ricolin has joined #openstack-ironic | 15:43 | |
| *** sziviani has joined #openstack-ironic | 15:43 | |
| TheJulia | Given the holiday in the states, adding it to the meeting might be a good idea for the L3 stuff. | 15:44 |
| iurygregory | let me see if I got right, so ipa would try to search and check the disk if the system is using UEFI if yes run the necessary commands to setup using efibootmgr? | 15:47 |
| dtantsur | etingof: upstream meetings don't tend to have enough time to solve difficult issues | 15:48 |
| iurygregory | ^ truth | 15:49 |
| dtantsur | etingof: doodle++ | 15:50 |
| dtantsur | we'll have a bit of a challenge to accommodate both Julia and Kaifeng.. | 15:50 |
| *** hoonetorg has quit IRC | 15:51 | |
| TheJulia | What is the simplest happy path? | 15:53 |
| TheJulia | where can we reach as an MVP and where can we iterate | 15:53 |
| TheJulia | The whole "this can be used for instances" concepts makes me want to scream | 15:53 |
| TheJulia | personally | 15:53 |
| TheJulia | or cry, I'm not quite sure | 15:53 |
| dtantsur | ++ | 15:54 |
| TheJulia | possibly drink although I've been trying not to do that | 15:54 |
| dtantsur | exactly, I think it could be substantially simpler if we didn't account for future enhancements that nobody so far has asked about | 15:55 |
| dtantsur | ++ to "not to do" :) | 15:55 |
| TheJulia | ++ It makes sense for some stubbing possibly, or maybe basic functionality, and the ramdisk concept in the boot interface actually provides that. But the ramdisk interface is basically a special interface for HPC folks with ephemeral workers | 15:56 |
| TheJulia | and vmedia interfaces eventually disconnect too.... | 15:56 |
| dtantsur | and then there is a problem of merging configdrives (user-provided and ours). brrrr | 15:57 |
| TheJulia | please, lets not do that | 15:57 |
| TheJulia | the ephemeral users expect that they are special and they are NOT nova driven | 15:58 |
| * TheJulia suspects the disagreement between dtantsur and herself is the classic "i want this implemented and no longer a headache" verses "sanity" | 15:59 | |
| dtantsur | that's why I don't quite like supporting this work for final instances :) | 15:59 |
| * TheJulia wonders if she can have some sort of santiy award sent to dtantsur | 15:59 | |
| dtantsur | :D | 16:00 |
| * etingof is not sure what "this can be used for instances" concept is all about | 16:00 | |
| * TheJulia wonders how to work through disconnecting the concepts | 16:00 | |
| dtantsur | etingof: where you mention cloud-init | 16:01 |
| dtantsur | i.e. the same thing applied to running instances, not IPA | 16:01 |
| etingof | that's the coolest part indeed | 16:01 |
| TheJulia | but cool is not mvp | 16:01 |
| TheJulia | dtantsur: call if you have spoons | 16:01 |
| dtantsur | ouch, sorry | 16:02 |
| TheJulia | no worries! | 16:02 |
| * dtantsur shakes fist at his calendar | 16:02 | |
| etingof | hmm, but will it be still entirely dhcpless w/o that? | 16:02 |
| TheJulia | dhcpless deploy != full dhcpless operation | 16:02 |
| TheJulia | we embed the config drives on disk for that reason | 16:02 |
| dtantsur | right | 16:03 |
| dtantsur | people can do dhcp-less instances right now (and I'm sure they do) | 16:03 |
| dtantsur | the deploy/inspection/cleaning time is a problem | 16:03 |
| etingof | right | 16:03 |
| etingof | so what's the better way? | 16:03 |
| dtantsur | let's try to strip down the spec to the bare minimum? | 16:04 |
| dtantsur | I'd still suggest delaying port groups | 16:04 |
| *** hoonetorg has joined #openstack-ironic | 16:04 | |
| dtantsur | (I don't insist, just testing them will be a pain) | 16:04 |
| dtantsur | then exclude cloud-init from the discussion | 16:05 |
| dtantsur | I'd personally avoid building configdrives, rather just putting the JSON file on the CD | 16:05 |
| *** gryf has joined #openstack-ironic | 16:07 | |
| dtantsur | actually, we can even skip the "link" declaration, only leave "network". we need IP addresses, after all. | 16:07 |
| etingof | links links IPs to NICs | 16:08 |
| dtantsur | etingof: you place the records on ports, you don't need that | 16:08 |
| dtantsur | or rather: ironic can generate it. let's not put stuff in the API that we already know. | 16:08 |
| etingof | so how'd ramdisk match NICs to IPs? | 16:09 |
| dtantsur | etingof: the ramdisk will have this information. it won't be provided by the operator though. | 16:09 |
| dtantsur | mmm, maybe I need to prototype an example, gimme a few minutes | 16:09 |
| etingof | ah, that would work | 16:09 |
| TheJulia | dtantsur: we can't skip that, see my comments on the latest, there is an id association expected out there | 16:10 |
| dtantsur | please wait for an example | 16:10 |
| TheJulia | example of the data is in the discussion thread from kaifeng | 16:10 |
| *** gryf has quit IRC | 16:12 | |
| * etingof feels that inventing our own JSON for network config could be suboptimal | 16:12 | |
| etingof | even if it's based on nova metadata | 16:12 |
| dtantsur | as suboptimal as writing our own tool to parse it? ;) | 16:13 |
| dtantsur | TheJulia, etingof https://etherpad.openstack.org/p/hlOiTjhsQd | 16:14 |
| dtantsur | I think this answers your questions | 16:14 |
| dtantsur | I don't know how DNS fits in, but that's the problem with the current spec too | 16:14 |
| etingof | will we still need to parse our own format? | 16:15 |
| dtantsur | etingof: this transformation happens on ironic side from the information we hve | 16:15 |
| dtantsur | * have | 16:15 |
| dtantsur | note that you already need some transformation since you need to merge the data from different ports | 16:15 |
| etingof | that ^ is eliminated in the latest revision | 16:16 |
| etingof | by moving from ports to node | 16:16 |
| dtantsur | oh. I haven't noticed that. | 16:16 |
| etingof | please, read 'Alternatives' | 16:16 |
| dtantsur | etingof: these are ideas that will NOT be implemented | 16:17 |
| dtantsur | the current proposal puts information on the ports | 16:17 |
| dtantsur | and this is what I'm trying to simplify | 16:17 |
| dtantsur | putting information on a node will solve the transformation part but will make data duplication much worse | 16:18 |
| etingof | I just wanted to offer the redesigned spec that way | 16:18 |
| * dtantsur is confused | 16:18 | |
| etingof | re parsing tool - it runs in ramdisk and it might be needed regardless of the format to convey network config there | 16:19 |
| etingof | if we stick with some standard format, then there would be an option to use some existing parser | 16:19 |
| dtantsur | the ramdisk side is not changed by any of the discussed proposals | 16:20 |
| dtantsur | we're discussing the ironic side and *only* it | 16:20 |
| * etingof thinks that ramdisk changes are noted in the spec | 16:21 | |
| dtantsur | etingof: I'm not sure what we're discussing any more. I'm talking about API. | 16:22 |
| * etingof assumes we are talking about L3 spec | 16:22 | |
| dtantsur | right, the API part of it | 16:23 |
| * etingof can do a deconfusing write-up in gerrit | 16:23 | |
| dtantsur | etingof: that's really what my recent comments have been about | 16:24 |
| dtantsur | etingof: please consider https://etherpad.openstack.org/p/hlOiTjhsQd for the API design, unless you really want to move network_data to node | 16:26 |
| openstackgerrit | Merged openstack/sushy-tools master: Fix libvirt CD image insertion https://review.opendev.org/693561 | 16:28 |
| etingof | yeah | 16:29 |
| dtantsur | etingof: okay, so is sushy-tools ready now? ^^^ I've seen that the CI job still fails for some reason that seems a real bug. | 16:30 |
| etingof | dtantsur, yes! I am on it. I am suspecting another sushy-tools bug \o/ | 16:31 |
| dtantsur | w00t | 16:31 |
| *** tssurya has quit IRC | 16:32 | |
| dtantsur | iurygregory: do you know if anybody is handling python 2 removal for staging-drivers? | 16:33 |
| iurygregory | dtantsur, https://review.opendev.org/696322 | 16:34 |
| patchbot | patch 696322 - x/ironic-staging-drivers - Drop python 2.7 support and testing - 1 patch set | 16:34 |
| iurygregory | But CI is not happy =) | 16:34 |
| dtantsur | thanks! do you know why? | 16:34 |
| * TheJulia needs to read this discussion in detail, but other humans first need to be prodded | 16:34 | |
| TheJulia | s/prodded/reached out to and provided information/ | 16:34 |
| iurygregory | dtantsur, still looking at the logs to find something useful =) | 16:35 |
| dtantsur | iurygregory: >=https://storage.gra1.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_d74/696322/1/check/ironic-staging-drivers-dsvm-all-drivers/d749683/logs/devstacklog.txt.gz | 16:40 |
| dtantsur | we need ironic-python-agent-builder added to required-projects | 16:40 |
| *** lucasagomes has quit IRC | 16:40 | |
| dtantsur | OMG it's a legacy job | 16:40 |
| dtantsur | then to PROJECTS | 16:40 |
| iurygregory | dtantsur, YES! | 16:40 |
| iurygregory | =D | 16:40 |
| rpittau | was that even working before ? :/ | 16:40 |
| iurygregory | legacy ftw - NO! | 16:40 |
| * iurygregory is not sure ... | 16:41 | |
| dtantsur | it was, the IPA-builder dependency is recent | 16:41 |
| iurygregory | ok i will add | 16:41 |
| rpittau | yeah | 16:41 |
| iurygregory | dtantsur, done | 16:46 |
| dtantsur | thx! | 16:46 |
| iurygregory | np! | 16:46 |
| iurygregory | bifrost need a lot of work before we can drop py27 =( | 16:47 |
| TheJulia | iurygregory: just break it | 16:48 |
| iurygregory | O.o | 16:48 |
| TheJulia | bifrost is a consumer, it doesn't always need to be working on master branch | 16:48 |
| iurygregory | aka non-voting jobs? | 16:48 |
| donnyd | has anyone has any CI related issues with fortnebula ? I haven't heard any registered complaints from ironic in a while and I wanted to check and make sure its working for you | 16:49 |
| TheJulia | iurygregory: yeah... | 16:49 |
| donnyd | you as in ironic - I know there were some bumps | 16:49 |
| TheJulia | donnyd: afaik once we change the MTU settings, things have seemingly been okay | 16:49 |
| iurygregory | donnyd, I know yesterday there was some problem but I didn't dig to check if was on fortnebula http://paste.openstack.org/show/786781/ =) | 16:50 |
| donnyd | TheJulia: awesome - I am happy we were able to figure it out | 16:50 |
| donnyd | inap.opendev.org | 16:50 |
| donnyd | If there is anything you need FN to do, please feel free to reach out... also if your jobs are failing on FN please let me know so I can get to the bottom of it. :) | 16:52 |
| *** _gryf has joined #openstack-ironic | 16:53 | |
| iurygregory | donnyd, tks for your support =) | 16:53 |
| iurygregory | TheJulia, https://review.opendev.org/#/c/695800/ this the patch to drop opensuse seem to be the problem when trying to drop, but also I'm not sure bifrost is supporting full py3 yet | 16:54 |
| patchbot | patch 695800 - bifrost - Drop python 2.7 support and testing - 1 patch set | 16:54 |
| *** _gryf is now known as gryf | 16:55 | |
| dtantsur | does anyone like removing code? I have a lot of code removal for you :) https://review.opendev.org/#/c/695041/ | 16:57 |
| patchbot | patch 695041 - ironic-python-agent - Replace WSME and Pecan with Werkzeug - 10 patch sets | 16:57 |
| TheJulia | dtantsur: I <3 removing code | 16:57 |
| TheJulia | iurygregory: they look like they are broken when zipper tries to look at repos... :\ | 16:57 |
| iurygregory | dtantsur, remove code is fun, but first i will try to make ipa work for the uefi secure boot XD | 16:58 |
| TheJulia | iurygregory: that bug +++ please | 16:58 |
| iurygregory | TheJulia, if I got right, so ipa would try to search and check the disk if the system is using UEFI if yes run the necessary commands to setup using efibootmgr? | 16:59 |
| TheJulia | iurygregory: correct | 16:59 |
| TheJulia | that way on disk signed files are not trashed | 17:00 |
| iurygregory | TheJulia, ack! time to start writting some code for this and think how to test o/ | 17:00 |
| iurygregory | something else would be necessary in other repos? | 17:00 |
| TheJulia | iurygregory: other repos, was there a context switch from your ack message? | 17:01 |
| *** dtantsur is now known as dtantsur|brb | 17:01 | |
| iurygregory | TheJulia, nope | 17:02 |
| iurygregory | i was just wondering since the have ipa-builder and maybe we would need something | 17:02 |
| *** ociuhandu has quit IRC | 17:05 | |
| openstackgerrit | Vishal Manchanda proposed openstack/ironic-ui master: Drop python 2.7 support and testing https://review.opendev.org/695801 | 17:05 |
| iurygregory | yay horizon help =D | 17:06 |
| TheJulia | \o/ | 17:06 |
| *** dhellmann_ has joined #openstack-ironic | 17:14 | |
| *** dhellmann has quit IRC | 17:14 | |
| *** dhellmann_ is now known as dhellmann | 17:15 | |
| rpioso | TheJulia: I'm deploying devstack. How can I specify a specific commit in the ironic repo? | 17:45 |
| rpioso | TheJulia: This fails -- enable_plugin ironic https://opendev.org/openstack/ironic aba282a7fc23124c2fe6a733925cafe5de59fee4 | 17:45 |
| openstackgerrit | Vishal Manchanda proposed openstack/ironic-ui master: Drop python 2.7 support and testing https://review.opendev.org/695801 | 17:49 |
| clarkb | rpioso: it may need to be a named ref. Try checking out that revision onto a branch then use the branch name in enable_plugin | 17:51 |
| iurygregory | enable_plugin <NAME> <GITURL> [GITREF] | 17:51 |
| iurygregory | name - an arbitrary name. (ex: glusterfs, docker, zaqar, congress) | 17:52 |
| iurygregory | giturl - a valid git url that can be cloned | 17:52 |
| iurygregory | gitref - an optional git ref (branch / ref / tag) that will be cloned. Defaults to master. | 17:52 |
| clarkb | rpioso: I think what it is trying to do is clone from that remote location, then fetch the gitref value from that location. You cannot fetch revisions they have to be a named ref when fetching | 17:52 |
| clarkb | eg git fetch origin my_branch is valid but git fetch origin aba282a7fc23124c2fe6a733925cafe5de59fee4 is not | 17:52 |
| *** tesseract has quit IRC | 17:53 | |
| rpioso | clarkb, iurygregory: Thank you! I'll give that a try. | 17:53 |
| iurygregory | you can also try using upstream reference you need https://review.opendev.org/#/q/I33e51f064daa0a0d70dd87e163d401c0e7b922bc | 17:54 |
| *** priteau has quit IRC | 17:55 | |
| * rpioso thinks iurygregory is reading his mind ;-) | 17:55 | |
| iurygregory | I just found the upstream reference for the commit you used XD | 17:56 |
| *** naved002 has joined #openstack-ironic | 17:56 | |
| *** dhellmann_ has joined #openstack-ironic | 17:57 | |
| *** yoctozepto5 has joined #openstack-ironic | 17:57 | |
| *** dhellmann has quit IRC | 17:57 | |
| *** naved001 has quit IRC | 17:57 | |
| *** dhellmann_ is now known as dhellmann | 17:58 | |
| *** yoctozepto has quit IRC | 17:58 | |
| *** yoctozepto5 is now known as yoctozepto | 17:58 | |
| *** dougsz has quit IRC | 17:58 | |
| * rpioso wonders if that applies to LIBS_FROM_GIT, xxx_REPO, and xxx_BRANCH, too | 17:59 | |
| rpioso | I'll soon-ish know :-) | 18:00 |
| *** derekh has quit IRC | 18:01 | |
| *** rpioso is now known as rpioso|afk | 18:01 | |
| * rpioso|afk goes to lunch | 18:01 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Replace netaddr dependency with stdlib ipaddress https://review.opendev.org/696378 | 18:09 |
| rpittau | good night! o/ | 18:13 |
| *** rpittau is now known as rpittau|afk | 18:14 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Remove the unused dependency on iso8601 https://review.opendev.org/696379 | 18:14 |
| *** ricolin has quit IRC | 18:16 | |
| *** dtantsur|brb is now known as dtantsur | 18:20 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Remove the unused dependency on iso8601 https://review.opendev.org/696379 | 18:22 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: Replace netaddr dependency with stdlib ipaddress https://review.opendev.org/696378 | 18:22 |
| *** rnoriega- has quit IRC | 18:23 | |
| *** ijw has joined #openstack-ironic | 18:23 | |
| *** xXraphXx has quit IRC | 18:23 | |
| *** dtantsur has quit IRC | 18:27 | |
| *** dtantsur has joined #openstack-ironic | 18:29 | |
| *** rnoriega_ has joined #openstack-ironic | 18:29 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent-builder master: Start building CentOS 8 images https://review.opendev.org/696383 | 18:31 |
| *** ociuhandu has joined #openstack-ironic | 19:22 | |
| *** rpioso|afk is now known as rpioso | 19:22 | |
| *** dtantsur is now known as dtantsur|afk | 19:46 | |
| *** _gryf has joined #openstack-ironic | 20:05 | |
| *** gryf is now known as _gryf_ | 20:05 | |
| *** _gryf is now known as gryf | 20:05 | |
| *** ociuhandu has quit IRC | 20:09 | |
| *** ijw_ has joined #openstack-ironic | 20:10 | |
| *** ijw has quit IRC | 20:13 | |
| *** pooja_pf9 has joined #openstack-ironic | 20:25 | |
| pooja_pf9 | Hi everyone.. I was trying to use direct deploy interface with Ironic/Nova on Rocky, but its not clear from docs if swift is mandatory or not. At one place, it says you can use custom HTTP(s) for image | 20:26 |
| pooja_pf9 | Is it correct that swift is optional? | 20:27 |
| openstackgerrit | Julia Kreger proposed openstack/ironic-python-agent master: RAID 5/6 https://review.opendev.org/694705 | 20:30 |
| *** alexmcleod_ has quit IRC | 20:37 | |
| *** ociuhandu has joined #openstack-ironic | 20:43 | |
| *** iurygregory has quit IRC | 20:48 | |
| *** ociuhandu has quit IRC | 20:48 | |
| *** tzumainn has quit IRC | 20:59 | |
| *** rh-jelabarre has quit IRC | 21:37 | |
| pooja_pf9 | Is swift optional for direct deploy only after upgrading to Stein release? | 21:42 |
| *** sshnaidm is now known as sshnaidm|afk | 21:43 | |
| *** rcernin has joined #openstack-ironic | 21:56 | |
| *** ociuhandu has joined #openstack-ironic | 22:02 | |
| *** zaneb has quit IRC | 22:04 | |
| *** pcaruana has quit IRC | 22:06 | |
| *** ociuhandu has quit IRC | 22:07 | |
| *** zaneb has joined #openstack-ironic | 22:10 | |
| *** zaneb has quit IRC | 22:23 | |
| *** ociuhandu has joined #openstack-ironic | 22:29 | |
| * etingof found libvirt being root of all evil | 22:36 | |
| *** ociuhandu has quit IRC | 22:44 | |
| TheJulia | etingof: \o/ | 22:48 |
| TheJulia | etingof: don't forget backports :) | 22:48 |
| TheJulia | pooja_pf9: I believe only optional after upgrading to stein. Direct deploy has been available for a "standalone" or user directly interacting with ironic for some time with just a url and checksum being stored, but that does not integrate with nova. | 22:49 |
| *** pooja_pf9 has quit IRC | 22:51 | |
| TheJulia | sadness | 22:51 |
| * TheJulia goes and begins "The cookening" | 22:51 | |
| *** ociuhandu has joined #openstack-ironic | 23:07 | |
| *** ociuhandu has quit IRC | 23:11 | |
| *** Lucas_Gray has joined #openstack-ironic | 23:13 | |
| *** Lucas_Gray has quit IRC | 23:19 | |
| *** ijw has joined #openstack-ironic | 23:21 | |
| *** Lucas_Gray has joined #openstack-ironic | 23:23 | |
| *** ijw_ has quit IRC | 23:23 | |
| *** rh-jelabarre has joined #openstack-ironic | 23:33 | |
| *** ociuhandu has joined #openstack-ironic | 23:49 | |
| *** goldyfruit_ has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!