*** tosky has quit IRC | 00:18 | |
*** paras333 has joined #openstack-ironic | 00:24 | |
*** paras333 has quit IRC | 00:28 | |
openstackgerrit | Jacob Anders proposed openstack/ironic-inspector master: Use OOB inspection to fetch MACs for IB inspection https://review.opendev.org/758994 | 00:51 |
---|---|---|
janders | ^^ finally I figured out where I stuffed up the unit tests \o/ | 00:51 |
janders | still WIP though | 00:51 |
janders | will be good to see if CI shares my enthusiasm regarding the fix | 00:52 |
*** gyee has quit IRC | 01:01 | |
stevebaker | janders: I just had a quick look at that change | 01:32 |
stevebaker | janders: I'm just wondering, is there a reason the lookup value is a string which is parsed into structured data and the end of the code-path? Could it be structured data all the way through? (a list of dicts or somesuch) | 01:35 |
* janders is opening the story | 01:36 | |
janders | good point stevebaker | 01:38 |
janders | my initial understanding of the story was that's the suggested approach, but now that I looked into all the components involved a little more this makes sense | 01:40 |
janders | it is also in line with earlier comments (which I'm going to work on now, the current revision of the patch was aiming to get all the tests passing again after tweaking parameter lists in different places) | 01:41 |
janders | let me see what the Redfish response with the MACs looks like, I wonder if it would make sense to pass that from the management interface to inspector | 01:42 |
stevebaker | janders: where does the lookup value come from? | 01:45 |
janders | stevebaker in the previous revision I was doing it here: https://review.opendev.org/#/c/750943/8/ironic/drivers/modules/redfish/management.py | 01:47 |
patchbot | patch 750943 - ironic - Use OOB inspection to fetch MACs for IB inspection - 10 patch sets | 01:47 |
janders | reworking this now | 01:48 |
stevebaker | janders: ah cool, I'll have a proper look later | 01:48 |
janders | no worries :) | 01:49 |
janders | this change covers ironic, inspector and sdk | 01:50 |
janders | https://review.opendev.org/#/c/762603/ | 01:50 |
patchbot | patch 762603 - openstacksdk - Use OOB inspection to fetch MACs for IB inspection - 1 patch set | 01:50 |
janders | that's the last bit (SDK) | 01:50 |
janders | I'll likely be tweaking all three a fair bit today | 01:51 |
janders | thanks for looking into this and good hints stevebaker | 01:54 |
janders | will get some lunch now | 01:54 |
*** thiagop has quit IRC | 01:57 | |
*** thiagop has joined #openstack-ironic | 01:57 | |
*** k_mouza has joined #openstack-ironic | 02:05 | |
*** k_mouza has quit IRC | 02:09 | |
*** Qianbiao has joined #openstack-ironic | 02:21 | |
*** jhesketh has quit IRC | 02:27 | |
*** tzumainn has quit IRC | 02:30 | |
*** jhesketh has joined #openstack-ironic | 02:33 | |
*** ociuhandu has joined #openstack-ironic | 02:36 | |
*** ociuhandu has quit IRC | 02:41 | |
*** thiagop has quit IRC | 02:59 | |
*** mkrai has joined #openstack-ironic | 03:14 | |
*** rcernin has quit IRC | 03:22 | |
*** rcernin has joined #openstack-ironic | 03:23 | |
*** iurygregory has quit IRC | 03:42 | |
openstackgerrit | Steve Baker proposed openstack/ironic master: Improve object_to_dict arguments https://review.opendev.org/763501 | 04:05 |
openstackgerrit | Jacob Anders proposed openstack/ironic master: Use OOB inspection to fetch MACs for IB inspection https://review.opendev.org/750943 | 04:32 |
*** ayoung has quit IRC | 04:32 | |
janders | stevebaker I will work on the Inspector side next - and check with Dmitry how he'd like to see this info passed on | 04:35 |
janders | (he left some comments on that in https://review.opendev.org/#/c/758994/3/ironic_inspector/main.py@346 ) | 04:35 |
patchbot | patch 758994 - ironic-inspector - Use OOB inspection to fetch MACs for IB inspection - 4 patch sets | 04:35 |
janders | heading off for some time, back in an hour or so... have a great weekend stevebaker :) few hours away for me still | 04:36 |
*** tkajinam has quit IRC | 04:40 | |
*** tkajinam has joined #openstack-ironic | 04:40 | |
stevebaker | janders: see you! | 04:48 |
*** zzzeek has quit IRC | 05:01 | |
*** zzzeek has joined #openstack-ironic | 05:07 | |
*** ociuhandu has joined #openstack-ironic | 05:29 | |
*** ociuhandu has quit IRC | 05:34 | |
*** Qianbiao has quit IRC | 05:34 | |
*** zzzeek has quit IRC | 05:37 | |
*** zzzeek has joined #openstack-ironic | 05:39 | |
*** rh-jlabarre has quit IRC | 05:40 | |
*** zzzeek has quit IRC | 05:47 | |
*** zzzeek has joined #openstack-ironic | 05:51 | |
*** zzzeek has quit IRC | 05:55 | |
*** zzzeek has joined #openstack-ironic | 05:58 | |
*** zzzeek has quit IRC | 06:12 | |
*** zzzeek has joined #openstack-ironic | 06:14 | |
*** zzzeek has quit IRC | 06:35 | |
*** k_mouza has joined #openstack-ironic | 06:36 | |
*** zzzeek has joined #openstack-ironic | 06:38 | |
*** k_mouza has quit IRC | 06:41 | |
*** rpittau|afk is now known as rpittau | 06:52 | |
rpittau | good morning ironic! o/ | 06:52 |
rpittau | happy friday! | 06:52 |
arne_wiebalck | Good morning rpittau and ironic! | 06:52 |
rpittau | hey arne_wiebalck :) | 06:52 |
*** rcernin has quit IRC | 06:55 | |
arne_wiebalck | jslater: as you suggest already, we add our downstream h/w manager as an additional dib element to a local copy of the upstream IPAB repo, then we have some additional convenience scripts which set the branches for the h/w manager and the IPA, and build the image | 06:56 |
*** Qianbiao has joined #openstack-ironic | 06:57 | |
arne_wiebalck | jslater: (in fact our h/w manager is just a shell which git clones the real h/w manager only when the IPA starts: this way we can use the same image and change the manager via git commits, which is handy for quick turn-around time during development) | 06:57 |
*** QianbiaoNG has joined #openstack-ironic | 06:59 | |
arne_wiebalck | rpittau: this mechanism^^ is maybe also sth we could add to the IPAB at some point, so users would only need to set the URL of a git repo to point to their downstream h/w manager (and it would be downloaded and started when the IPA starts) | 07:00 |
*** Qianbiao has quit IRC | 07:02 | |
*** mkrai has quit IRC | 07:02 | |
*** mkrai_ has joined #openstack-ironic | 07:02 | |
rpittau | arne_wiebalck: yep, that's interesting, sounds very useful, will make life easier for integrating custom h/w managers in the ipa ramdisk | 07:03 |
arne_wiebalck | rpittau: it was very handy over the past years: building an image still takes ~10mins and if you code like I do ... :-D | 07:03 |
*** rcernin has joined #openstack-ironic | 07:04 | |
rpittau | arne_wiebalck: I totally get that :D | 07:04 |
arne_wiebalck | rpittau: "this time the code is correct ..." ;) | 07:05 |
rpittau | restarting ipa building in the middle of the process because you realized you forgot a comma or a bracket, check! | 07:06 |
arne_wiebalck | lol, yeah that one | 07:08 |
*** mkrai_ has quit IRC | 07:20 | |
*** fmuyassarov has joined #openstack-ironic | 07:29 | |
*** iurygregory has joined #openstack-ironic | 07:39 | |
iurygregory | good morning Ironic! | 07:40 |
*** mkrai_ has joined #openstack-ironic | 07:46 | |
rpittau | hey iurygregory :) | 07:49 |
iurygregory | rpittau, o/ | 07:50 |
*** zzzeek has quit IRC | 07:51 | |
*** rcernin has quit IRC | 07:53 | |
*** zzzeek has joined #openstack-ironic | 07:53 | |
*** zzzeek has quit IRC | 08:17 | |
*** zzzeek has joined #openstack-ironic | 08:18 | |
arne_wiebalck | hey iurygregory o/ | 08:39 |
iurygregory | arne_wiebalck, o/ | 08:39 |
*** mgoddard has joined #openstack-ironic | 08:43 | |
*** zzzeek has quit IRC | 09:02 | |
*** zzzeek has joined #openstack-ironic | 09:05 | |
*** derekh has joined #openstack-ironic | 09:15 | |
openstackgerrit | Riccardo Pittau proposed openstack/sushy-tools master: [WIP] Support processors schema https://review.opendev.org/761909 | 09:26 |
*** rcernin has joined #openstack-ironic | 09:33 | |
QianbiaoNG | Good morning my friends and Ironic | 09:34 |
QianbiaoNG | Hello, anyone knows what IPA means in this project https://opendev.org/x/novajoin | 09:34 |
QianbiaoNG | this maybe? https://www.freeipa.org/page/Main_Page | 09:35 |
rpittau | QianbiaoNG: possibly, maybe ask in #openstack-nova ? | 09:37 |
*** k_mouza has joined #openstack-ironic | 09:37 | |
QianbiaoNG | ok sure. | 09:37 |
janders | good morning rpittau arne_wiebalck iurygregory QianbiaoNG and Ironic o/ | 09:37 |
rpittau | hey janders :) | 09:38 |
QianbiaoNG | Hello janders | 09:38 |
QianbiaoNG | you are late today :) | 09:38 |
janders | haha in a way :) | 09:38 |
rpittau | janders: it was my turn to get access denied to gerrit after fedora upgrade :) | 09:38 |
janders | started early, finished early, reappeared late | 09:39 |
janders | rpittau ouch! | 09:39 |
janders | rpittau is it under control, or still troubleshooting? | 09:39 |
QianbiaoNG | janders nice, i like to work at midnight when i am alone. | 09:40 |
rpittau | janders: it's ok now, it didn't take me too long luckily, just added an option to ssh config | 09:40 |
janders | rpittau good on you | 09:40 |
arne_wiebalck | hey janders QianbiaoNG o/ | 09:40 |
janders | it looks like my car is back to roadworthy... will park it on a steep uphill overnight see if it starts easily | 09:41 |
arne_wiebalck | QianbiaoNG: yes, FreeIPA | 09:41 |
QianbiaoNG | hey arne_wiebalck | 09:42 |
QianbiaoNG | ty | 09:42 |
QianbiaoNG | arne_wiebalck did you use vendor metadata service in your cloud? | 09:42 |
janders | I had an extra fuel filter fitted (wasnt cheap), so that if there's any fuel contamination, or water gets into the fuel system while water crossing, it protects the fuel system | 09:42 |
janders | turns out that while it *maybe* does that it's got an air leak somewhere and it's draining the fuel system from any diesel overnight | 09:42 |
janders | now the extra filter is bypassed, lets see if the car is back to normal :) | 09:43 |
janders | technology... gotta love it | 09:43 |
*** rcernin has quit IRC | 09:49 | |
iurygregory | hey QianbiaoNG and janders o/ | 09:55 |
QianbiaoNG | hey iurygregory \0/ | 09:55 |
*** tosky has joined #openstack-ironic | 09:57 | |
*** dtantsur|afk is now known as dtantsur | 09:57 | |
dtantsur | morning ironic | 09:57 |
janders | SPUC time? | 09:57 |
janders | hey dtantsur | 09:57 |
dtantsur | SPUC, SPUC \o/ | 09:58 |
dtantsur | yeah, gimme a couple of minutes | 09:58 |
iurygregory | SPUC \o/ | 09:58 |
dtantsur | FileNotFoundError: [Errno 2] No such file or directory: '/dev/disk/by-label' | 09:59 |
dtantsur | jesus, tinyIPA!! | 09:59 |
janders | now, for a change, my chromium is exploding when I try to fire up BJeans | 10:01 |
janders | FF? | 10:01 |
iurygregory | we have 2 janders on SPUC =O | 10:02 |
rpittau | that's probablt better | 10:02 |
rpittau | I don't have the url though :/ | 10:02 |
janders | yeah and they are both dead :/ | 10:02 |
iurygregory | =/ | 10:02 |
janders | chromium explosions killed them | 10:02 |
dtantsur | webrtc is broken in chromium atm | 10:03 |
janders | okay, trying FF, fingers crossed! | 10:03 |
* dtantsur has uses FF since forever | 10:03 | |
rpittau | can anyone pass the SPUC url pls ? | 10:03 |
dtantsur | anyone else up for SPUC? rpittau, arne_wiebalck? | 10:03 |
dtantsur | https://bluejeans.com/643711802 | 10:03 |
rpittau | thanks | 10:03 |
*** mkrai_ has quit IRC | 10:10 | |
*** belmoreira has joined #openstack-ironic | 10:15 | |
*** openstackgerrit has quit IRC | 10:25 | |
*** mgoddard has quit IRC | 10:28 | |
*** ociuhandu has joined #openstack-ironic | 10:32 | |
*** ociuhandu has quit IRC | 10:33 | |
*** mgoddard has joined #openstack-ironic | 10:42 | |
*** k_mouza has quit IRC | 10:43 | |
*** uzumaki has joined #openstack-ironic | 10:43 | |
*** k_mouza has joined #openstack-ironic | 10:44 | |
*** rcernin has joined #openstack-ironic | 10:51 | |
*** rcernin has quit IRC | 10:56 | |
*** uzumaki has quit IRC | 10:57 | |
*** QianbiaoNG has quit IRC | 11:08 | |
*** QianbiaoNG has joined #openstack-ironic | 11:11 | |
arne_wiebalck | thanks for the ping dtantsur, was fun to SPUC with you :) | 11:17 |
dtantsur | glad to hear :) | 11:20 |
dtantsur | arne_wiebalck: it's happening every Friday until x-mas | 11:21 |
*** ociuhandu has joined #openstack-ironic | 11:26 | |
*** openstackgerrit has joined #openstack-ironic | 11:26 | |
openstackgerrit | Merged openstack/ironic master: Expunge the internal version of WSME https://review.opendev.org/750861 | 11:26 |
dtantsur | \o/ | 11:27 |
dtantsur | stevebaker++ | 11:27 |
*** k_mouza has quit IRC | 11:41 | |
*** k_mouza has joined #openstack-ironic | 11:43 | |
*** k_mouza has quit IRC | 11:45 | |
*** k_mouza has joined #openstack-ironic | 11:45 | |
*** ociuhandu has quit IRC | 11:58 | |
*** ociuhandu has joined #openstack-ironic | 11:59 | |
*** ociuhandu has quit IRC | 12:01 | |
*** ociuhandu has joined #openstack-ironic | 12:02 | |
*** k_mouza has quit IRC | 12:11 | |
QianbiaoNG | hello, if port-group is used, will nova auto add metadata to config bonding in baremetal os | 12:14 |
dtantsur | I think so | 12:15 |
QianbiaoNG | nice | 12:15 |
QianbiaoNG | will try test it on lab env. | 12:15 |
janders | dtantsur regarding comment https://review.opendev.org/#/c/758994/3/ironic_inspector/main.py@346 would you be happy if I just grab the MAC address info off sushy in get_mac_addresses in the management interface ( say in this way: https://review.opendev.org/#/c/750943/11/ironic/drivers/modules/redfish/management.py@1022 ) and pass that to Inspector | 12:16 |
janders | and process this further there? | 12:16 |
patchbot | patch 758994 - ironic-inspector - Use OOB inspection to fetch MACs for IB inspection - 4 patch sets | 12:16 |
patchbot | patch 750943 - ironic - Use OOB inspection to fetch MACs for IB inspection - 11 patch sets | 12:16 |
dtantsur | janders: not sure I fully grasp the question | 12:18 |
dtantsur | do you want to change the API format? | 12:18 |
janders | hmm :) | 12:19 |
janders | that doesn't sound like a good idea | 12:19 |
janders | Steve pointed out I'm grabbing the MAC info, converting it into a string, sending it to inspector and parsing that again | 12:20 |
janders | so I figured I might just pass it as-is to avoid that | 12:21 |
dtantsur | wdym as is? | 12:21 |
dtantsur | I imagined ?lookup=mac:11:22:33:44:55:66,mac:aa:bb:cc:dd:ee:ff etc | 12:21 |
* janders is grabbing references to the code | 12:22 | |
*** zzzeek has quit IRC | 12:22 | |
janders | I am grabbing the MAC address info here: https://review.opendev.org/#/c/750943/11/ironic/drivers/modules/redfish/management.py@1022 | 12:24 |
patchbot | patch 750943 - ironic - Use OOB inspection to fetch MACs for IB inspection - 11 patch sets | 12:24 |
*** zzzeek has joined #openstack-ironic | 12:24 | |
janders | previously I was converting this to a string, get_mac_addresses would return this string and it would get passed on to here: https://review.opendev.org/#/c/758994/3/ironic_inspector/main.py@346 | 12:24 |
patchbot | patch 758994 - ironic-inspector - Use OOB inspection to fetch MACs for IB inspection - 4 patch sets | 12:25 |
dtantsur | hmm, okay, so it's not a simple list, but rather some structure? | 12:25 |
janders | I was wondering if it would be a good idea to skip converting mac lookup info into a string and pass on the MAC as returned from sushy | 12:26 |
*** fmuyassarov has quit IRC | 12:26 | |
dtantsur | how would it look in the URL? | 12:26 |
janders | but having said that I naively haven't through through the mechanics of how that traverses APIs | 12:26 |
dtantsur | right :) | 12:26 |
janders | and the reality is I dont think the sushy info on what ports seem enabled matters - we want all, correct? | 12:27 |
* dtantsur thinks | 12:27 | |
dtantsur | I wonder what their definition of enabled is | 12:28 |
dtantsur | but yeah, for the purpose of lookup we can live with disabled (in any sense) NICs | 12:28 |
janders | that is a very good question :) | 12:28 |
janders | if it's better to use only the enabled ones that's a good case for moving the code that's identifying those into redfish_utils | 12:29 |
janders | as per your other comment | 12:30 |
dtantsur | I guess using only enabled NICs may be less confusing when debugging, but otherwise I don't care much | 12:33 |
*** fmuyassarov has joined #openstack-ironic | 12:33 | |
janders | that is a fair point - and this is what RF OOB inspection does too | 12:34 |
janders | it seems like it's better if I stick with that | 12:34 |
janders | okay! that gives me a good pointer for Monday :) thank you dtantsur | 12:35 |
dtantsur | cool | 12:35 |
*** ociuhandu has quit IRC | 12:38 | |
*** ociuhandu has joined #openstack-ironic | 12:39 | |
*** k_mouza has joined #openstack-ironic | 12:40 | |
*** fmuyassarov has quit IRC | 12:42 | |
janders | wrapping up for the day (night), have a great weekend everyone, see you next week | 12:43 |
janders | weeks_left_to_xmas-- | 12:43 |
dtantsur | o/ | 12:43 |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent master: [WIP] Copy any configuration from the virtual media https://review.opendev.org/763207 | 12:43 |
*** ociuhandu has quit IRC | 12:43 | |
-openstackstatus- NOTICE: The Gerrit service at review.opendev.org will be offline starting at 15:00 UTC (roughly two hours from now) for a weekend upgrade maintenance: http://lists.opendev.org/pipermail/service-announce/2020-October/000012.html | 13:02 | |
*** ociuhandu has joined #openstack-ironic | 13:09 | |
*** zzzeek has quit IRC | 13:13 | |
*** fmuyassarov has joined #openstack-ironic | 13:14 | |
openstackgerrit | Merged openstack/ironic-python-agent stable/victoria: Fix vendor info retrieval for some versions of lshw https://review.opendev.org/763406 | 13:15 |
*** zzzeek has joined #openstack-ironic | 13:17 | |
*** ociuhandu has quit IRC | 13:19 | |
*** rh-jlabarre has joined #openstack-ironic | 13:21 | |
*** ociuhandu has joined #openstack-ironic | 13:23 | |
*** ociuhandu has quit IRC | 13:23 | |
openstackgerrit | Riccardo Pittau proposed openstack/sushy-tools master: [WIP] Support processors schema https://review.opendev.org/761909 | 13:29 |
*** QianbiaoNG has quit IRC | 13:34 | |
*** QianbiaoNG has joined #openstack-ironic | 13:38 | |
*** ociuhandu has joined #openstack-ironic | 13:38 | |
*** QianbiaoNG has quit IRC | 13:39 | |
*** QianbiaoNG has joined #openstack-ironic | 13:39 | |
*** fmuyassarov has quit IRC | 13:42 | |
*** zzzeek has quit IRC | 13:56 | |
*** zzzeek has joined #openstack-ironic | 13:57 | |
dtantsur | lovely https://bugzilla.redhat.com/show_bug.cgi?id=950760 | 13:58 |
openstack | bugzilla.redhat.com bug 950760 in grub2 "double quotes in kernel parameter" [High,New] - Assigned to pjones | 13:58 |
-openstackstatus- NOTICE: The Gerrit service at review.opendev.org will be offline starting at 15:00 UTC (roughly one hour from now) for a weekend upgrade maintenance: http://lists.opendev.org/pipermail/service-announce/2020-October/000012.html | 14:01 | |
iurygregory | dtantsur, dammm | 14:01 |
* dtantsur is testing https://review.opendev.org/763566 | 14:02 | |
patchbot | patch 763566 - diskimage-builder - Fix dynamic-login with grub2 - 1 patch set | 14:02 |
dtantsur | because of this dynamic-login does not work with virtual media | 14:02 |
iurygregory | damm it =( | 14:03 |
dtantsur | oh heck, I've never seen this bit https://opendev.org/openstack/diskimage-builder/src/branch/master/diskimage_builder/elements/simple-init#user-content-networkmanager | 14:08 |
TheJulia | hmm... gerrit down for the weekend | 14:10 |
dtantsur | TheJulia: they're forcing us to take a break! | 14:10 |
dtantsur | and good mornign | 14:10 |
TheJulia | dtantsur: I know... I don't know how I feel about it | 14:10 |
dtantsur | you can always contribute to something on github :) | 14:10 |
TheJulia | dtantsur: I've got a lot of unit tests I could write | 14:11 |
rpittau | oh wow the entire weekend | 14:11 |
dtantsur | there's never a shortage of unit tests to write or refactor | 14:11 |
rpittau | I read 1 hour, but that's just the start | 14:11 |
dtantsur | rpittau: it's a minor upgrade, followed by a major upgrade, followed by a minor upgrade | 14:11 |
rpittau | I see | 14:11 |
TheJulia | I guess I could always go back to sleep... | 14:13 |
*** k_mouza has quit IRC | 14:14 | |
dtantsur | +++ | 14:14 |
dtantsur | I only woke up because of SPUC | 14:14 |
*** rloo has joined #openstack-ironic | 14:15 | |
*** zzzeek has quit IRC | 14:17 | |
*** zzzeek has joined #openstack-ironic | 14:17 | |
TheJulia | gah, when is the spuc? | 14:18 |
TheJulia | 1700 utc? | 14:18 |
dtantsur | TheJulia: the first instance is over, the next is at 5pm UTC | 14:18 |
TheJulia | okay | 14:18 |
*** nam-est has joined #openstack-ironic | 14:18 | |
TheJulia | well plenty of time for me to CAFFINATE | 14:18 |
dtantsur | (the first was at 10am UTC - way too early for you and a bit too early for today's me :) | 14:18 |
nam-est | Hi folks, | 14:18 |
nam-est | I am trying to run Ironic-api as a wsgi application on top of an Apache server | 14:19 |
nam-est | the problem is that in apache conf file, there is no way to add parameters to run the ironic-api | 14:20 |
nam-est | However, it allows to set up environment variable | 14:21 |
TheJulia | sadly dalek caffinate gifs seem to be few and far between | 14:21 |
* TheJulia goes and caffinates | 14:21 | |
nam-est | and the parameter I want to pass is --config-file to choose the ironic configuration file | 14:22 |
nam-est | so my question is: are there anyways to pick the configuration file based on environment variables? | 14:22 |
arne_wiebalck | nam-est: the config file for the ironic virtual host? | 14:22 |
dtantsur | I think it's about the location of ironic.conf.. a good question indeed | 14:23 |
*** ociuhandu has quit IRC | 14:23 | |
nam-est | arne_wiebalck: No, config file for the ironic-api itself | 14:23 |
*** mgoddard has quit IRC | 14:25 | |
dtantsur | nam-est: if you use uwsgi or gunicorn, they seem to be able to do that natively | 14:25 |
dtantsur | now, oslo.config actually have a way to pass any (?) option via environment variables | 14:26 |
dtantsur | I don't remember the exact format, something like OS_DEFAULT__CONFIG_FILE | 14:26 |
dtantsur | nam-est: maybe using uwsgi as a backend (https://uwsgi-docs.readthedocs.io/en/latest/WSGIquickstart.html#putting-behind-a-full-webserver) is a better option though | 14:28 |
iurygregory | good morning TheJulia | 14:31 |
TheJulia | https://specs.openstack.org/openstack/oslo-specs/specs/rocky/config-from-environment.html <-- looks like it was implemented | 14:31 |
dtantsur | yeah, I'm pretty sure metal3 uses this feature :) | 14:32 |
nam-est | dtantsur: I plan to use Apache for handling the TLS part of the connection. | 14:32 |
nam-est | I checked this: https://docs.openstack.org/oslo.config/latest/reference/configuration-files.html but it seems like there is no other option expect the commandline parameter | 14:34 |
*** ociuhandu has joined #openstack-ironic | 14:34 | |
dtantsur | nam-est: you can use apache for HTTPS and use uwsgi with ironic as a backend (instead of mod_wsgi) | 14:35 |
nam-est | dtantsur: I see. Good suggestion. I will take a look at it. Thanks | 14:37 |
*** zzzeek has quit IRC | 14:40 | |
nam-est | TheJulia: dtantsur: Oh I get the idea you mentioned above. So it is possible to set one option per environment variable. This is also a good option. Thanks | 14:42 |
*** tzumainn has joined #openstack-ironic | 14:42 | |
TheJulia | ++ | 14:42 |
dtantsur | yep | 14:42 |
*** zzzeek has joined #openstack-ironic | 14:43 | |
*** mkrai has joined #openstack-ironic | 14:50 | |
*** Goneri has joined #openstack-ironic | 14:51 | |
rpittau | bye everyone, have a great weekend! o/ | 14:53 |
*** rpittau is now known as rpittau|afk | 14:53 | |
-openstackstatus- NOTICE: The Gerrit service at review.opendev.org is offline for a weekend upgrade maintenance, updates will be provided once it's available again: http://lists.opendev.org/pipermail/service-announce/2020-October/000012.html | 15:04 | |
dtantsur | bye bye gerrit, have a good weekend :D | 15:04 |
* dtantsur is curious how it will look on Monday | 15:04 | |
iurygregory | hahaha | 15:07 |
TheJulia | ++ | 15:07 |
iurygregory | monday gerrit will be sleeping | 15:07 |
TheJulia | although last time they did a super major upgrade, they kept a lot of the prior skinning | 15:07 |
iurygregory | and we will send emails so it will wake up haha | 15:07 |
TheJulia | so users were happier | 15:07 |
TheJulia | or something like that | 15:07 |
dtantsur | the announcements says not everything is possible to keep | 15:08 |
* dtantsur is desperately trying to make dhcp-less work.. | 15:09 | |
*** paras333_ has joined #openstack-ironic | 15:13 | |
TheJulia | :( | 15:13 |
*** paras333_ is now known as Paras | 15:13 | |
* TheJulia is trying to figure out how much pain the secure rbac effort is going to be | 15:14 | |
*** rloo has quit IRC | 15:16 | |
*** rloo has joined #openstack-ironic | 15:17 | |
clarkb | it requires significant rework to keep the skinning. We decided it was prudent to upgrade and sort that out after the upgrade since the new ui is functional as is if not as awesome as it could be | 15:22 |
dtantsur | fair enough | 15:22 |
dtantsur | on the dhcp-less note: | 15:23 |
dtantsur | remember I was arguing with Ilya that we should automatically popular MAC addresses because people will make mistakes? | 15:23 |
*** k_mouza has joined #openstack-ironic | 15:23 | |
dtantsur | I've just made a mistake | 15:23 |
TheJulia | doh | 15:25 |
dtantsur | A lot of usability improvements ahead, assuming I can make it work at all | 15:26 |
*** k_mouza has quit IRC | 15:27 | |
*** k_mouza_ has joined #openstack-ironic | 15:27 | |
dtantsur | all I could do so far is to kill any network connectivity | 15:34 |
iurygregory | people can always make mistakes .-. | 15:34 |
dtantsur | I wonder if glean works with NM operating systems at all | 15:34 |
TheJulia | ohhhh | 15:34 |
TheJulia | I bet it doesn't | 15:34 |
TheJulia | it bet it is only doing dhcp in that case | 15:34 |
dtantsur | I mean, it's supposed to, and someone from infra told it did, but I see no way it could | 15:35 |
TheJulia | maybe support means the dhcp only case? | 15:35 |
TheJulia | That, afaik, is their use case | 15:35 |
dtantsur | well, DHCP works in NM by default anyway, so what's the point | 15:37 |
TheJulia | is this where we go back to "NetworkManager was a mistake"?' | 15:38 |
dtantsur | probably? at least for servers | 15:38 |
*** mgoddard has joined #openstack-ironic | 15:41 | |
clarkb | we use glean on all the centos and I think fedora VMs that weuse for testing | 15:44 |
clarkb | with network manager | 15:44 |
dtantsur | DHCP? static IP quite clearly does not work | 15:44 |
clarkb | both, rackspace is static only | 15:44 |
dtantsur | clarkb: please tell me how you make it work.. | 15:45 |
dtantsur | right now it looks like this: http://paste.openstack.org/show/800261/ | 15:45 |
dtantsur | (--use-nm doesn't make any difference to me) | 15:45 |
clarkb | I dont have it all paged in now. ianw added that support and may be able to help (sorry doing gerrit things) | 15:46 |
dtantsur | okie, no worries | 15:46 |
dtantsur | it's good to know it works, it's sad that just using simple-init element doesn't seem to be the path | 15:46 |
dtantsur | okay, this sheds some light: https://mail.gnome.org/archives/networkmanager-list/2014-January/msg00025.html | 15:49 |
dtantsur | oh funny. networkmanager creates a new inactive connection instead of updating the active one.. | 15:57 |
nam-est | In Ironic, does parameter --config-dir have default value? | 16:07 |
dtantsur | nam-est: I don't think so | 16:07 |
nam-est | dtantsur: I see. Thanks | 16:09 |
dtantsur | only --config-file defaults to /etc/<project>/<project>.conf where <project> is ironic in our case | 16:09 |
*** mkrai has quit IRC | 16:11 | |
*** ociuhandu has quit IRC | 16:23 | |
*** ociuhandu has joined #openstack-ironic | 16:25 | |
*** ociuhandu has quit IRC | 16:25 | |
*** gyee has joined #openstack-ironic | 16:32 | |
*** QianbiaoNG has quit IRC | 16:37 | |
*** k_mouza_ has quit IRC | 16:40 | |
*** k_mouza has joined #openstack-ironic | 16:51 | |
dtantsur | reminder: SPUC in 8 minutes! | 16:52 |
TheJulia | brraaainnnnsss | 16:54 |
*** tosky has quit IRC | 16:56 | |
JayF | same bluejeans, right? | 16:59 |
JayF | I'm the only one in right now :D | 16:59 |
dtantsur | I'm on my way | 16:59 |
TheJulia | SPUC!!!! | 17:03 |
*** belmoreira has quit IRC | 17:16 | |
NobodyCam | Good Morning Ironic folks, Happy Friday! | 17:27 |
dtantsur | hey-hey NobodyCam | 17:27 |
dtantsur | wanna join SPUC? | 17:28 |
NobodyCam | Sorry; to many fires going today :( | 17:28 |
dtantsur | ouch. good luck then! | 17:28 |
*** rloo has quit IRC | 17:30 | |
*** rloo has joined #openstack-ironic | 17:30 | |
*** k_mouza has quit IRC | 17:32 | |
*** k_mouza has joined #openstack-ironic | 17:38 | |
*** k_mouza has quit IRC | 17:39 | |
*** k_mouza has joined #openstack-ironic | 17:45 | |
*** mgoddard has quit IRC | 17:51 | |
*** nam-est has quit IRC | 17:58 | |
*** tosky has joined #openstack-ironic | 18:10 | |
*** derekh has quit IRC | 18:12 | |
*** k_mouza has quit IRC | 18:20 | |
*** ociuhandu has joined #openstack-ironic | 18:20 | |
*** k_mouza has joined #openstack-ironic | 18:23 | |
iurygregory | I lost the SPUC probably (just got back home) | 18:29 |
dtantsur | yep, it's over already, sorry | 18:29 |
iurygregory | no worries | 18:29 |
iurygregory | I had to go out to buy masks =( | 18:29 |
iurygregory | 5 wwouldn't be enough for the whole travel | 18:29 |
*** ociuhandu has quit IRC | 18:30 | |
*** ociuhandu has joined #openstack-ironic | 18:31 | |
dtantsur | true | 18:31 |
*** ociuhandu has quit IRC | 18:36 | |
*** yoctozepto has quit IRC | 18:37 | |
*** ociuhandu has joined #openstack-ironic | 18:38 | |
*** ociuhandu has quit IRC | 18:38 | |
*** yoctozepto has joined #openstack-ironic | 18:38 | |
*** ociuhandu has joined #openstack-ironic | 18:40 | |
*** ociuhandu has quit IRC | 18:44 | |
TheJulia | the real question will be, how many people bring like whole headgear things. | 18:51 |
*** zzzeek has quit IRC | 18:52 | |
*** zzzeek has joined #openstack-ironic | 18:53 | |
dtantsur | a good question! | 18:55 |
*** jamesdenton has quit IRC | 18:57 | |
*** jamesden_ has joined #openstack-ironic | 18:57 | |
*** zzzeek has quit IRC | 19:00 | |
*** ociuhandu has joined #openstack-ironic | 19:03 | |
*** zzzeek has joined #openstack-ironic | 19:03 | |
JayF | Hey; who is leading up the conversations with the DMTF, if someone downstream here wanted to get involved with it. | 19:05 |
dtantsur | JayF: maybe rpioso? | 19:07 |
*** rloo has quit IRC | 19:10 | |
*** rloo has joined #openstack-ironic | 19:10 | |
JayF | ack; I'll reach out to them when I figure out who is going to want to have that chat here | 19:11 |
dtantsur | the last thing I've heard was that only vendors have direct access to DMTF (rpioso is from Dell) | 19:11 |
dtantsur | maybe I misunderstood that.. | 19:11 |
*** k_mouza has quit IRC | 19:12 | |
JayF | I don't know, but I am aware that there are folks here who are having those kind of conversations with our vendors, and may have insight into things we want to ensure are in the ironic profile | 19:12 |
dtantsur | I see | 19:13 |
*** tosky has quit IRC | 19:13 | |
*** dtantsur is now known as dtantsur|afk | 19:15 | |
dtantsur|afk | have a nice weekend folks! | 19:15 |
arne_wiebalck | JayF: there is a change proposed with a seth of redfish interop profiles; these profile reflect the Ironic use of redfish | 19:19 |
arne_wiebalck | JayF: https://review.opendev.org/#/c/754061/ | 19:20 |
patchbot | No data found for patch 754061 | 19:20 |
arne_wiebalck | JayF: I have started to test these with some of our hardware; more testers/feedback would be great | 19:20 |
arne_wiebalck | JayF: there are also some practical issue I raised on the change as well | 19:21 |
arne_wiebalck | JayF: once there is more feedback, I guess we will have another discussion how to proceed | 19:22 |
TheJulia | gmann: w/r/t nova's policy testing, is there any reason it is just not built against the policy in code. It seems a little weird that there is the whole fixture substrate built in. | 19:22 |
arne_wiebalck | bye everyone, have a good weekend o/ | 19:22 |
TheJulia | you too arne_wiebalck | 19:23 |
buhman | I've thought for years it might be "cool" to have per-machine-type kernels (with compiled-in firmware, via CONFIG_EXTRA_FIRMWARE) from both a minimalism and hardware qualification perspective. The dream would be that whatever team is responsible for qualifying new hardware platforms is also responsible for defining the minimal Linux kernel configuration for that platform (so that buhman doesn't have to do that) | 19:24 |
buhman | I've never seen a hardware team actually operate this way though. | 19:24 |
rpioso | dtantsur, JayF: Members of the DMTF have full access. Google joined relatively recently: https://www.dmtf.org/about/list. Note that both IBM and Verizon are Leadership members. | 19:24 |
rpioso | arne_wiebalck: \o Hope you have an enjoyable weekend :) | 19:24 |
*** ociuhandu has quit IRC | 19:27 | |
rpioso | JayF: I, along with arne_wiebalck and rpittau, are developing Redfish Interoperability Profiles for the vendor-independent redfish hardware type: https://storyboard.openstack.org/#!/story/2007345. | 19:29 |
rpioso | JayF: bdodd does work for the DMTF Redfish Forum. He's been implementing a RAID hardware interface implementation which uses Redfish. It's up on Gerrit. He's typically on this channel. | 19:34 |
rpioso | JayF: Also, Mike Raineri, Dell and Jeff Autor, HPE, co-chairs of the DMTF Redfish Forum, have actively engaged with the ironic community here and during meetings, such as the PTG. The Redfish Forum is responsible for the Redfish standard. Mike is often on this channel. His nick is mraineri. | 19:34 |
JayF | These links are great; I've passed them on. Thanks! | 19:35 |
*** ociuhandu has joined #openstack-ironic | 19:36 | |
*** ociuhandu has quit IRC | 19:58 | |
*** ociuhandu has joined #openstack-ironic | 19:59 | |
gmann | TheJulia: we have unit test now specific to testing the policy in code without fixture. But yes policy fixture for other unit/functional tests is not good and I am planning to remove those fixture and run all tests against policy in code by default. | 20:02 |
gmann | how much test changes are required for that, that is one thing but plan is to get rid of those policy fixture | 20:02 |
*** ociuhandu has quit IRC | 20:05 | |
*** rloo has quit IRC | 20:07 | |
*** rloo has joined #openstack-ironic | 20:08 | |
TheJulia | gmann: got a good example to one that is independent of the fixture code handy? | 20:25 |
rpioso | JayF: Glad to help. | 20:25 |
*** ociuhandu has joined #openstack-ironic | 20:30 | |
*** iurygregory has quit IRC | 20:42 | |
*** iurygregory has joined #openstack-ironic | 20:59 | |
gmann | TheJulia: all these policy tests run with actual policy defaults in code - https://github.com/openstack/nova/tree/master/nova/tests/unit/policies | 20:59 |
TheJulia | gmann: okay, I guess common_policy_check will continue to be used then, just seems like rule name passing could be dropped today if it is really representitive testing of the actual encoded policy | 21:11 |
gmann | TheJulia: that is used only for error message checks in case of expected 403. | 21:12 |
*** k_mouza has joined #openstack-ironic | 21:13 | |
gmann | this is another set of code I need to push in oslotest and make it more generic for reusing it. | 21:13 |
* TheJulia wonders why we would intentionally leak policy details | 21:13 | |
TheJulia | oh, it is not going over actual HTTP, it is going to the internal method | 21:15 |
TheJulia | and I would then hope the exception would be cached | 21:15 |
gmann | yeah, internal, actual HTTP would not have these rule name | 21:15 |
TheJulia | err, catched | 21:15 |
TheJulia | I'm really worrying that few may be able to re-use this style of testing without rewriting their apis | 21:15 |
TheJulia | fwiw | 21:16 |
gmann | at least basic framework can be reused and each service can write per policy tests based on common_policy_check | 21:17 |
*** k_mouza has quit IRC | 21:18 | |
TheJulia | Well, even then, you'd have to support additional validator methods because trying to hunt for a rule name or text string wouldn't fly if we need to run over http to exercise the policy enforcement | 21:18 |
gmann | i mean just for unit tests and with mock. | 21:20 |
TheJulia | uhh, even then | 21:20 |
TheJulia | well | 21:20 |
TheJulia | to clarify, we're calling over http in many cases and mock the backend rpc so we can see what happens on the back-end | 21:21 |
gmann | i see your point now, yes that depends on how API/url controller/methods are implemented in services | 21:23 |
TheJulia | exactly | 21:26 |
*** jamesden_ is now known as jamesdenton | 21:29 | |
TheJulia | gmann: looks like in ironic, we mock the memcache cache, does any of the changes actually structurally change any of that. i.e. would the fields/data be expected to be the same because if so it should be fairly easy to run with that pattern in these cases | 21:31 |
gmann | TheJulia: it depends case by case, If policy controls the fields/data like in nova many policies does, then we need to specifically assert on fields returned | 21:33 |
TheJulia | gmann: indeed, and that will have to be encoded at the test level I think | 21:33 |
gmann | yeah tests level. | 21:33 |
gmann | if it is API access level then it is easy | 21:34 |
gmann | sometime single API might have multiple policy checks, these policy tests then need to open up the policy to assert on all those multiple policy | 21:35 |
gmann | example this one - https://github.com/openstack/nova/blob/master/nova/tests/unit/policies/test_instance_actions.py#L132 | 21:36 |
TheJulia | Yeah, iterate through the policy and expected behavior, luckilly our deviations are rare | 21:36 |
gmann | ok, in nova we have many | 21:37 |
TheJulia | yeah, I need to build a real matrix of it since we're still mostly an admin API. Inspector luckilly is pure admin usage only | 21:38 |
gmann | i see | 21:39 |
* TheJulia thinks it is time to call it a day and let maybe go write words() | 21:46 | |
jamesdenton | TheJulia quick question for you if you're still around. this is re: NGS | 21:48 |
TheJulia | jamesdenton: what's up? | 21:56 |
TheJulia | Also o/ | 21:56 |
jamesdenton | hi! | 21:56 |
jamesdenton | little tough without Gerrit, but some of the feedback (IIRC) was to try and implement some checks to ensure the switch *and* the physnet were valid. | 21:57 |
jamesdenton | So, if i have a port group with two links, and one of the links references a switch that's not configured, what should the action be? | 21:58 |
jamesdenton | any of this ringing a bell? L:D | 21:59 |
TheJulia | kind of yeah :) | 22:00 |
TheJulia | hm.. the bind should fail, ideally | 22:00 |
TheJulia | oh, two links | 22:00 |
TheJulia | hmmmmm | 22:00 |
TheJulia | Likely still bindfail I think | 22:01 |
jamesdenton | so, looking at the logs, i have a portgroup but during *provisioning* each port of the port group is handled independently. Of the two, i have a valid one and an invalid one. The valid one binds and the invalid one fails (expected) | 22:02 |
TheJulia | yeah, I don't think we abort on binding failure, it just reflects in port state in neutron | 22:04 |
jamesdenton | right | 22:05 |
jamesdenton | putting together a longer response, one sec | 22:06 |
TheJulia | ok | 22:06 |
jamesdenton | http://paste.openstack.org/show/800267/ | 22:11 |
jamesdenton | sorry, had to wait for the output | 22:11 |
*** rh-jlabarre has quit IRC | 22:12 | |
jamesdenton | post-provisioning, you'll see the VIF port fail to bind due to one of the links being invalid. | 22:12 |
jamesdenton | but of the two links, one was actually configured and the other not (because the switch is fake). Anyway, i'm guessing the goal would be to not configure either of them if one is invalid, right? | 22:13 |
jamesdenton | all of this is to say... during provisioning, if the ports are processed individually, we might actually deploy an OS on the host, but when it comes time to configure the ports after the reboot (to take them off provisioning network and move to tenant network) that could fail | 22:15 |
jamesdenton | and is that acceptable | 22:16 |
*** ociuhandu has quit IRC | 22:17 | |
TheJulia | Eh... | 22:18 |
*** ociuhandu has joined #openstack-ironic | 22:18 | |
TheJulia | I think it is acceptable even though it doesn't feel great | 22:18 |
TheJulia | It is ultimately a misconfiguration, it would be great if somewhere that were visible though as if there was a warning or "hey, this might be a problem | 22:18 |
TheJulia | I think also depending on switch config, the switch may isolate a link and things may still work on the one link | 22:20 |
jamesdenton | i mean, it's certainly possible to log a warning that the switch or physnet isn't configured, and ultimately it wouldn't bind | 22:20 |
TheJulia | but that is entirely dependent on what they encoded in the behavior | 22:21 |
TheJulia | but the other port would? or not at all? | 22:21 |
jamesdenton | well, i have a patch that will not configure any link in the group if any of them are invalid | 22:22 |
jamesdenton | i need to see if the server moves off the provisioning vlan onto the parking vlan | 22:22 |
jamesdenton | before moving to the tenant vlan | 22:22 |
*** ociuhandu has quit IRC | 22:23 | |
jamesdenton | but that would kinda be a silent failure, as far as the user is concerned | 22:23 |
jamesdenton | anyway, it's late and i'm sure we have better things to do :) | 22:24 |
TheJulia | hmm... | 22:25 |
TheJulia | It might need to be admin configurable behavior, in the end | 22:25 |
jamesdenton | i'll revisit this next week and bounce it off of you again | 22:28 |
jamesdenton | thanks for your time and have a good weekend | 22:28 |
*** rloo has quit IRC | 22:30 | |
*** ociuhandu has joined #openstack-ironic | 22:48 | |
TheJulia | jamesdenton: sounds good, have a wonderful weekend | 22:49 |
*** knikolla has quit IRC | 22:50 | |
*** bfournie has quit IRC | 22:52 | |
*** knikolla has joined #openstack-ironic | 22:52 | |
*** ociuhandu has quit IRC | 23:13 | |
*** iurygregory has quit IRC | 23:22 | |
*** ociuhandu has joined #openstack-ironic | 23:40 | |
*** ociuhandu has quit IRC | 23:44 | |
*** ociuhandu has joined #openstack-ironic | 23:57 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!