| *** jawad_axd has joined #openstack-ironic | 00:01 | |
| *** jawad_axd has quit IRC | 00:05 | |
| *** ociuhandu has joined #openstack-ironic | 00:12 | |
| *** jawad_axd has joined #openstack-ironic | 00:22 | |
| *** igordc has quit IRC | 00:23 | |
| *** jawad_axd has quit IRC | 00:26 | |
| *** astnbomb has joined #openstack-ironic | 00:29 | |
| *** astnbomb has quit IRC | 00:33 | |
| *** astnbomb has joined #openstack-ironic | 00:33 | |
| *** astnbomb has quit IRC | 00:38 | |
| *** ociuhandu has quit IRC | 00:51 | |
| *** astnbomb has joined #openstack-ironic | 00:55 | |
| *** k_mouza has joined #openstack-ironic | 01:01 | |
| *** k_mouza has quit IRC | 01:06 | |
| *** diurnalist has quit IRC | 01:14 | |
| *** diurnalist has joined #openstack-ironic | 01:24 | |
| *** stevebaker has joined #openstack-ironic | 01:27 | |
| *** astnbomb has quit IRC | 01:28 | |
| *** diurnalist has quit IRC | 01:31 | |
| *** gyee has quit IRC | 01:52 | |
| *** rloo has quit IRC | 02:48 | |
| *** mkrai has joined #openstack-ironic | 03:09 | |
| *** jdandrea has quit IRC | 03:37 | |
| *** tzumainn has quit IRC | 03:56 | |
| *** stevebaker has quit IRC | 04:03 | |
| *** mkrai has quit IRC | 04:05 | |
| *** mkrai_ has joined #openstack-ironic | 04:05 | |
| *** rh-jelabarre has quit IRC | 05:08 | |
| *** ijw has quit IRC | 05:46 | |
| *** kozhukalov has joined #openstack-ironic | 06:15 | |
| *** ijw has joined #openstack-ironic | 06:15 | |
| *** ijw has quit IRC | 06:20 | |
| *** ijw has joined #openstack-ironic | 06:47 | |
| *** ijw has quit IRC | 06:52 | |
| *** jawad_axd has joined #openstack-ironic | 07:08 | |
| *** ijw has joined #openstack-ironic | 07:19 | |
| *** ijw has quit IRC | 07:23 | |
| arne_wiebalck | Good morning, ironic! | 07:37 |
|---|---|---|
| arne_wiebalck | diurnalist: I find the easiest way to get into a RAM disk is to put a node into maintenance and trigger cleaning. The node will boot, but then stop and wait. If you have your keys embedded in the RAM disk image, this will allow you to log in. | 07:39 |
| *** ijw has joined #openstack-ironic | 07:51 | |
| *** ijw has quit IRC | 07:55 | |
| *** tesseract has joined #openstack-ironic | 08:05 | |
| *** ociuhandu has joined #openstack-ironic | 08:07 | |
| *** ijw has joined #openstack-ironic | 08:22 | |
| *** rpittau|afk is now known as rpittau | 08:22 | |
| rpittau | good morning ironic! o/ | 08:22 |
| *** ijw has quit IRC | 08:26 | |
| rpittau | TGIF! | 08:28 |
| *** ociuhandu has quit IRC | 08:29 | |
| *** ociuhandu has joined #openstack-ironic | 08:29 | |
| *** ociuhandu has quit IRC | 08:38 | |
| *** ociuhandu has joined #openstack-ironic | 08:42 | |
| *** ociuhandu has quit IRC | 08:42 | |
| *** amoralej|off is now known as amoralej | 08:54 | |
| *** rcernin has quit IRC | 08:56 | |
| *** diga has joined #openstack-ironic | 09:08 | |
| diga | Hi Ironic! | 09:08 |
| diga | This patch is getting error for sushy package | 09:09 |
| diga | https://review.opendev.org/#/c/710327/1 | 09:09 |
| patchbot | patch 710327 - x/sushy-oem-idrac - Export System Configuration Added code for exporti... - 1 patch set | 09:09 |
| diga | https://www.irccloud.com/pastebin/tlpMGEGa/ | 09:09 |
| *** lucasagomes has joined #openstack-ironic | 09:10 | |
| diga | I think issues is in existing Auth module of sushy and that's the reason its breaking the sushy package | 09:10 |
| diga | Only tox-py27 is failing | 09:10 |
| *** maelk has joined #openstack-ironic | 09:11 | |
| *** k_mouza has joined #openstack-ironic | 09:21 | |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic master: Default IRONIC_RAMDISK_TYPE to dib https://review.opendev.org/710278 | 09:22 |
| *** ijw has joined #openstack-ironic | 09:24 | |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic master: Default IRONIC_RAMDISK_TYPE to dib https://review.opendev.org/710278 | 09:24 |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic master: Rename ironic-tox-unit-with-driver-libs-python3 https://review.opendev.org/710298 | 09:25 |
| openstackgerrit | M V P Nitesh proposed openstack/ironic-python-agent master: Fixes error handling if efibootmgr is not present in ramdisk https://review.opendev.org/709107 | 09:26 |
| *** ijw has quit IRC | 09:28 | |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic master: General cleaning of jobs https://review.opendev.org/710415 | 09:31 |
| *** alexmcleod has joined #openstack-ironic | 09:37 | |
| openstackgerrit | Riccardo Pittau proposed x/sushy-oem-idrac master: Cap sushy version for python 2 https://review.opendev.org/710416 | 09:37 |
| rpittau | diga: the version of sushy used there is Python3-only compatible, that's why the error. This ^ should help for now | 09:37 |
| diga | Okay, then in that case, Py27 job should be removed for sushy right ? | 09:38 |
| rpittau | diga: well, that depends, if we want to keep supporting Python 2 then no. You can depend rebase the patch that is failing on mine for now and it will pass. | 09:39 |
| rpittau | Then we need to discuss if keeping python 2 support or not | 09:39 |
| diga | ok | 09:39 |
| rpittau | I guess we need to involve etingof and rpioso at least | 09:40 |
| *** derekh has joined #openstack-ironic | 09:40 | |
| etingof | so why would we want py2 support? | 09:41 |
| rpittau | I don't :D | 09:45 |
| rpittau | etingof: I just wonder if we should merge the final support for python 2, cut a release, and then remove the support | 09:46 |
| rpittau | latest release of sushy-oem-idrac is from september 2019 | 09:47 |
| etingof | but we do not have py2 support in sushy master already, right? | 09:47 |
| rpittau | etingof: exactly, that's why capping the version | 09:47 |
| rpittau | so we can give a final python2 "compatible" version | 09:49 |
| etingof | ah, you mean cut a finaly py2 release for sushy-oem-idrac? | 09:49 |
| rpittau | yeah | 09:49 |
| rpittau | sorry :) | 09:49 |
| etingof | right! | 09:49 |
| etingof | makes sense to me, but we should summon rpioso for approval | 09:49 |
| openstackgerrit | Riccardo Pittau proposed x/sushy-oem-idrac master: Cap sushy version for python 2 https://review.opendev.org/710416 | 09:51 |
| rpioso | etingof, rpittau: Sounds good to me. Why not!? | 09:56 |
| * rpioso tends to be very agreeable while on vaca :) | 09:56 | |
| rpittau | :) | 10:01 |
| rpittau | ok then we can add the cap first, cut a release, and then remove python2 support entirely | 10:01 |
| openstackgerrit | Arne Wiebalck proposed openstack/ironic-python-agent master: [doc] How to pause the IPA for debugging https://review.opendev.org/710419 | 10:05 |
| *** tesseract has quit IRC | 10:11 | |
| *** tesseract has joined #openstack-ironic | 10:12 | |
| *** k_mouza has quit IRC | 10:16 | |
| diga | rpioso, rpittau I will ask Mahendra not to update the patch till above patch merges | 10:16 |
| *** k_mouza_ has joined #openstack-ironic | 10:16 | |
| rpittau | diga: ack | 10:16 |
| *** tesseract has quit IRC | 10:16 | |
| *** tesseract has joined #openstack-ironic | 10:16 | |
| openstackgerrit | Riccardo Pittau proposed x/sushy-oem-idrac master: Cap sushy version for python 2 https://review.opendev.org/710416 | 10:20 |
| *** ijw has joined #openstack-ironic | 10:27 | |
| *** mkrai_ has quit IRC | 10:29 | |
| *** trident has quit IRC | 10:31 | |
| *** mkrai_ has joined #openstack-ironic | 10:31 | |
| *** ijw has quit IRC | 10:32 | |
| *** trident has joined #openstack-ironic | 10:34 | |
| *** khansa has joined #openstack-ironic | 10:39 | |
| *** khansa has quit IRC | 10:41 | |
| rpittau | fantastic I can't reproduce the error in py27 | 10:41 |
| arne_wiebalck | heisenbug | 10:50 |
| rpittau | fun fact: changing version of tox will change the result of the test | 10:57 |
| *** ijw has joined #openstack-ironic | 10:58 | |
| *** ijw has quit IRC | 11:03 | |
| *** ijw has joined #openstack-ironic | 11:30 | |
| *** rloo has joined #openstack-ironic | 11:33 | |
| *** ijw has quit IRC | 11:34 | |
| *** khansa has joined #openstack-ironic | 11:35 | |
| *** khansa has quit IRC | 11:58 | |
| *** ijw has joined #openstack-ironic | 12:01 | |
| *** ijw has quit IRC | 12:06 | |
| *** khansa has joined #openstack-ironic | 12:10 | |
| *** ijw has joined #openstack-ironic | 12:32 | |
| *** tkajinam has quit IRC | 12:34 | |
| *** ijw has quit IRC | 12:37 | |
| *** jdandrea has joined #openstack-ironic | 12:51 | |
| *** jdandrea has quit IRC | 12:55 | |
| *** diga has quit IRC | 12:59 | |
| *** rh-jelabarre has joined #openstack-ironic | 13:00 | |
| rpittau | etingof, rpioso, I have the feeling that never worked with python2 | 13:01 |
| etingof | did not we have py2 job? | 13:02 |
| *** rgp has left #openstack-ironic | 13:04 | |
| *** ijw has joined #openstack-ironic | 13:04 | |
| rpittau | yeah, we did, I see the CI config was added 2 months ago, so I tried with different sushy versions but I'm always getting an error with py27 and not with py36 | 13:04 |
| etingof | so some newer sushy version broke sushy-oem-idrac? | 13:05 |
| etingof | on py2 | 13:05 |
| rpittau | I tried all the versions from 2.0.0 to 3.0.0 | 13:07 |
| rpittau | 3.1.0 removes six so can't be used anymore | 13:07 |
| *** ijw has quit IRC | 13:09 | |
| rpittau | it *might* be an issue with the CI, it installs tox using pip on the system and runs the test on py27 using tox under py36, but I don't see how that could influence the tests | 13:11 |
| * etingof is confused | 13:12 | |
| etingof | how CI installs tox on py3? | 13:12 |
| rpittau | I'm also very confused, but if you look at the CI logs, you'll see tox installed using pip and it runs with: using tox-3.14.5 from /usr/local/lib/python3.6/dist-packages/tox/__init__.py (pid 1413) | 13:14 |
| rpittau | even for py27 | 13:14 |
| etingof | is this how the problem manifests itself? -- https://zuul.opendev.org/t/openstack/build/2291c46feabf45f88b274b6c1b1bc4a7 | 13:16 |
| rpittau | etingof: yeah | 13:17 |
| rpittau | the funny thing is that if I use system tox from python3.7 works fine, if I use tox from a py27 virtualenv it gives the error | 13:21 |
| etingof | is it reproducible locally? | 13:22 |
| rpittau | yeah | 13:22 |
| rpittau | I can reproduce it on ubuntu and fedeora | 13:22 |
| etingof | let me dig in a bit | 13:23 |
| rpittau | etingof: if it can help, this is the run with the system tox under python3.7 http://paste.openstack.org/show/790106/ | 13:24 |
| etingof | running tox against your patch just works for me, hmmm | 13:25 |
| rpittau | etingof: this is the paste using a python 2.7 virtualenv and tox running there http://paste.openstack.org/show/790107/ | 13:25 |
| rpittau | first: success! second: fail! | 13:26 |
| *** Lucas_Gray has joined #openstack-ironic | 13:26 | |
| *** mkrai_ has quit IRC | 13:26 | |
| arne_wiebalck | etingof: is there something like a list of things that Ironic would need from a redfish endpoint to work properly? | 13:28 |
| arne_wiebalck | etingof: background is that if we want to use redfish on new hardware with ironic we would need to specify what we actually need from the hardware | 13:29 |
| etingof | arne_wiebalck, that depends on the interfaces being used | 13:29 |
| arne_wiebalck | etingof: so there are interface based lists :-P | 13:30 |
| etingof | arne_wiebalck, this is difficult... because even if your vendors confirm that they support certain endpoints, they still might be broken... | 13:30 |
| arne_wiebalck | etingof: in which case we would send things back I guess :) | 13:30 |
| arne_wiebalck | etingof: srsly, I'd think everything that we do now with ipmi would be a candidate for redfish | 13:31 |
| *** Lucas_Gray has quit IRC | 13:32 | |
| arne_wiebalck | etingof: power, vendor, management | 13:32 |
| etingof | yes, but the mapping is not always 1-to-1 | 13:32 |
| arne_wiebalck | I see. | 13:32 |
| etingof | for example, we have bios mgmt | 13:32 |
| arne_wiebalck | ok, so I guess the we could do is to look what we request for ipmi to work (in our specification) and require this for redfish | 13:33 |
| *** astnbomb has joined #openstack-ironic | 13:34 | |
| etingof | ideally, you should have: power mgmt, boot mode and boot device mgmt, vmedia boot, more than one vmedia device (e.g. + floppy or USB thumb drive) to host configs, exposed fans, temps, power metrics | 13:34 |
| *** ijw has joined #openstack-ironic | 13:35 | |
| etingof | + console redirection, may be some storage management... | 13:36 |
| *** zaneb has quit IRC | 13:36 | |
| arne_wiebalck | that's quite a list already, thanks! | 13:36 |
| arne_wiebalck | from your experience is that list usually implement or does this vary siginificantly from vendor to vendor? | 13:37 |
| arne_wiebalck | implemented | 13:37 |
| arne_wiebalck | manufacturerer, not vendor | 13:38 |
| arne_wiebalck | manufacturer | 13:38 |
| *** astnbomb has quit IRC | 13:38 | |
| * arne_wiebalck blames Friday | 13:38 | |
| etingof | I do not have that much experience yet, but based on how permissive redfish spec is in part of what needs to be implemented, I fear there can be significant differences... | 13:39 |
| *** astnbomb has joined #openstack-ironic | 13:39 | |
| *** ijw has quit IRC | 13:39 | |
| etingof | another completely orthogonal complication that we might be facing is how the vendor interprets the details of the standard | 13:40 |
| *** tesseract has quit IRC | 13:40 | |
| etingof | and finally, we've seen many silly bugs with some vendors (like typos in JSON docs) that break things apart | 13:40 |
| *** tesseract has joined #openstack-ironic | 13:42 | |
| *** astnbomb has quit IRC | 13:43 | |
| arne_wiebalck | sounds like a lot of fun | 13:45 |
| arne_wiebalck | "fun" | 13:45 |
| etingof | may be we could ask the vendor to web-scrape their BMC to see at least what's present...? | 13:45 |
| arne_wiebalck | that's a nice idea ... or even provide a test program that will check what is offered | 13:46 |
| arne_wiebalck | the main things we are interested in ofc are the calls ironic would make | 13:46 |
| etingof | right, the light way would be to just scrape their jsons. we would have a more precise analysis if we'd try to run some commands that are part of ironic workflow (e.g. set boot params) | 13:50 |
| arne_wiebalck | one could ask a test node and perform the whole lifecycle | 13:50 |
| etingof | if we host such tool publicly, the vendors could self-certify against ironic - expose their BMC on the internet and point our tool to it | 13:52 |
| arne_wiebalck | uh, nice | 13:52 |
| etingof | that might also ease life for BMC devs if they could quickly see how their implementation works with ours | 13:53 |
| arne_wiebalck | it could also have multiple levels of certification, depending on which functionality can be performed via redfish | 13:53 |
| etingof | yes, like scenarios | 13:53 |
| arne_wiebalck | exactly | 13:53 |
| arne_wiebalck | you can do power mgmt, but now have a second vmedia device | 13:54 |
| arne_wiebalck | not | 13:54 |
| etingof | right | 13:54 |
| etingof | it's not that difficult to test so far as we do not require a lot yet | 13:55 |
| arne_wiebalck | is there an easy way to capture what calls a redfish endpoint receives? | 13:55 |
| arne_wiebalck | I'm thinking of a trace which we could use to know the calls per scenario. | 13:55 |
| arne_wiebalck | Rather than reading the code. | 13:56 |
| * arne_wiebalck coughs | 13:56 | |
| etingof | well, if you enable conductor debugging, run the node through the entire lifecycle having as much redfish interfaces enabled as your hw can afford - you will get all redfish/HTTP calls in the log | 13:57 |
| etingof | may be running ironic against sushy-tools emulator would be easier because it's being developed together with ironic they have a feature balance | 13:58 |
| *** Goneri has joined #openstack-ironic | 14:00 | |
| *** khansa has quit IRC | 14:01 | |
| arne_wiebalck | ok | 14:01 |
| *** khansa63 has joined #openstack-ironic | 14:01 | |
| arne_wiebalck | how about we put all this into a story? this may be helpful once we find the resources to actually do it. | 14:01 |
| etingof | +1 from me | 14:03 |
| etingof | I can do it unless you do it | 14:03 |
| *** mkrai_ has joined #openstack-ironic | 14:06 | |
| *** ijw has joined #openstack-ironic | 14:07 | |
| khansa63 | I am In if there is something I can help with my modest experience :D I very liked the idea (y) (y) | 14:07 |
| *** khansa63 is now known as khansa | 14:07 | |
| etingof | rpittau, I will look into the tox mystery later today | 14:07 |
| *** mkrai_ has quit IRC | 14:08 | |
| rpittau | etingof: ok, I'll get back to it also if I have some time | 14:08 |
| etingof | khansa, having sushy-cli as functional as ironic would probably help as we could use sushy-cli for poking at BMCs | 14:08 |
| etingof | that is to say: the more redfish commands sushy-cli implements - the better | 14:09 |
| *** ijw has quit IRC | 14:11 | |
| *** ociuhandu has joined #openstack-ironic | 14:19 | |
| *** ociuhandu has quit IRC | 14:20 | |
| *** ociuhandu has joined #openstack-ironic | 14:21 | |
| khansa | etingof Great! would it be an inspiration for new command implementation? :D | 14:22 |
| *** ociuhandu has quit IRC | 14:24 | |
| *** ociuhandu has joined #openstack-ironic | 14:24 | |
| etingof | khansa, indeed! I've left some thoughts on Zulip yesterday | 14:26 |
| *** amoralej is now known as amoralej|lunch | 14:30 | |
| rpittau | what's going on with the CI today..... | 14:30 |
| arne_wiebalck | etingof: how about I start and you add your thoughts? | 14:31 |
| openstackgerrit | Riccardo Pittau proposed openstack/ironic master: Default IRONIC_RAMDISK_TYPE to dib https://review.opendev.org/710278 | 14:36 |
| *** ijw has joined #openstack-ironic | 14:38 | |
| *** ijw has quit IRC | 14:42 | |
| etingof | arne_wiebalck, works for me! | 14:45 |
| etingof | should we title it "The Golden Fish Certification Initiative"? | 14:46 |
| *** jawad_axd has quit IRC | 14:48 | |
| *** jawad_axd has joined #openstack-ironic | 14:53 | |
| *** ociuhandu has quit IRC | 14:54 | |
| arne_wiebalck | lol | 14:55 |
| * arne_wiebalck is in a meeting now, will put sth up after | 14:57 | |
| *** jawad_axd has quit IRC | 14:57 | |
| *** jdandrea has joined #openstack-ironic | 14:58 | |
| *** tzumainn has joined #openstack-ironic | 15:01 | |
| *** zaneb has joined #openstack-ironic | 15:01 | |
| *** ociuhandu has joined #openstack-ironic | 15:03 | |
| *** khansa has quit IRC | 15:03 | |
| *** tesseract has quit IRC | 15:05 | |
| *** zaneb has quit IRC | 15:06 | |
| *** zaneb has joined #openstack-ironic | 15:06 | |
| *** tesseract has joined #openstack-ironic | 15:07 | |
| *** astnbomb has joined #openstack-ironic | 15:09 | |
| *** ijw has joined #openstack-ironic | 15:09 | |
| *** ijw has quit IRC | 15:13 | |
| *** amoralej|lunch is now known as amoralej | 15:13 | |
| *** bnemec is now known as beekneemech | 15:16 | |
| *** zaneb has quit IRC | 15:23 | |
| *** astnbomb has quit IRC | 15:30 | |
| *** astnbomb has joined #openstack-ironic | 15:31 | |
| *** ociuhandu has quit IRC | 15:31 | |
| etingof | rpittau, btw, it seems version capping is not respected - we have the same sushy (3.1.0) on Py2 -- https://zuul.opendev.org/t/openstack/build/2291c46feabf45f88b274b6c1b1bc4a7/log/tox/py27-1.log#259 | 15:32 |
| etingof | may be we should use all three digits in versions? | 15:32 |
| rpittau | no, wait, it is respected, otherwise it would give a different error | 15:33 |
| rpittau | if you look in the test execution it will uninstall 3.1.0 and install 3.0.0 | 15:34 |
| rpittau | etingof: https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_229/710416/3/check/openstack-tox-py27/2291c46/tox/py27-2.log | 15:36 |
| etingof | hmm | 15:37 |
| etingof | rpittau, regarding tox from py3 - may be that's OK because tox should be running prior to any venv | 15:38 |
| *** ociuhandu has joined #openstack-ironic | 15:38 | |
| rpittau | etingof: I agree, although that should not influence the execution, but it does | 15:39 |
| *** ijw has joined #openstack-ironic | 15:40 | |
| *** jcoufal has joined #openstack-ironic | 15:42 | |
| rpittau | the main thing I'm seeing here is that running tox under python 2.7 and 3.6 fails, while running under 3.7 works | 15:43 |
| etingof | right | 15:43 |
| *** ijw has quit IRC | 15:44 | |
| rpittau | etingof: I made it fail with python 3.7 too, so I can say that the py27 test doesn't work at all | 15:48 |
| *** jdandrea has quit IRC | 15:52 | |
| arne_wiebalck | khansa63: for the "golden fish initiative" you mean? | 15:56 |
| *** zaneb has joined #openstack-ironic | 15:57 | |
| *** jdandrea has joined #openstack-ironic | 16:00 | |
| *** jawad_axd has joined #openstack-ironic | 16:01 | |
| etingof | rpittau, so if sushy is not capped, does it change anything in the way how tox is invoked? | 16:01 |
| rpittau | etingof: it won't remove 3.1.0 to install 3.0.0 | 16:02 |
| etingof | can it be that reinstallation of sushy breaks something in the entry points...? | 16:02 |
| rpittau | maybe, but I did a test capping the version for py 2.7, 3.6 and 3.7 and only 2.7 fails | 16:04 |
| *** jawad_axd has quit IRC | 16:05 | |
| *** diurnalist has joined #openstack-ironic | 16:06 | |
| etingof | perhaps setuptools takes part in the whole entrypoint business. so if we have different setuptools on py2/3, may be that has something to do with that | 16:06 |
| etingof | what's surprising is that just one (of many) entry point fails | 16:07 |
| arne_wiebalck | etingof: khansa63: https://storyboard.openstack.org/#!/story/2007345 | 16:10 |
| *** ijw has joined #openstack-ironic | 16:11 | |
| etingof | looks good to me | 16:12 |
| etingof | why would a vendor take the trouble of certifying their sw, unless they acquire the 'golden fish' trophy...? | 16:16 |
| *** ijw has quit IRC | 16:16 | |
| *** gyee has joined #openstack-ironic | 16:16 | |
| rpittau | should we start printing stickers with "ironic-compliant" ? | 16:17 |
| *** tesseract has quit IRC | 16:17 | |
| arne_wiebalck | rpittau: sounds like a good idea, who can resist stickers? | 16:18 |
| *** astnbomb has quit IRC | 16:18 | |
| *** astnbomb has joined #openstack-ironic | 16:19 | |
| *** astnbomb has quit IRC | 16:19 | |
| *** astnbomb has joined #openstack-ironic | 16:19 | |
| etingof | rpittau, I confused sushy-oem-idrac with something else - the former has just one entry point, not many | 16:21 |
| rpittau | etingof: I tried downgrading setuptools to match the system one, as it's working with that, but I still get the same error :/ | 16:27 |
| etingof | how do you reproduce it locally? | 16:29 |
| rpittau | I built virtualenv with different python version (2.7,3.6,3.7,3.8) and invoking tox from there. To impose a specific setuptools version I cap it in the testenv entry in tox.ini under deps | 16:31 |
| *** ijw has joined #openstack-ironic | 16:43 | |
| rpioso | arne_wiebalck, etingof: Couldn’t 3d Party CIs be used to test fishy support? | 16:43 |
| rpioso | arne_wiebalck, etingof: Dell EMC offers a sushy 3d Party CI. | 16:45 |
| etingof | my understanding is that we are looking to have a way to quickly evaluate a Redfish implementation at hand | 16:46 |
| etingof | like press a button and get a result - fishy/non-fishy | 16:47 |
| *** ijw has quit IRC | 16:47 | |
| etingof | ideally, that button should be available to anyone on the internet e.g. BMC coders | 16:47 |
| etingof | having said that, how would a CI help here? | 16:48 |
| *** ociuhandu has quit IRC | 16:49 | |
| *** lucasagomes has quit IRC | 16:53 | |
| *** ociuhandu has joined #openstack-ironic | 16:54 | |
| arne_wiebalck | rpioso: the background is that I'd like to check if I can use certain hardware with Redfish, in particular when I look for sth to purchase | 16:54 |
| arne_wiebalck | rpioso: make that: if I can use certain hardware with Ironic & Redfish | 16:55 |
| arne_wiebalck | rpioso: basically: "to which level does this hardware and its Redfish implementation allow me to use Redfish based interfaces for the nodes" | 16:56 |
| rpioso | etingof: As Ironic features are added via Redfish, the interoperability needs will increase. Not sure how a one-time button push could “certify” a BMC. As cdearborn suggested at the midcycle, we’re experiencing growing/adaptation challenges. | 16:56 |
| rpioso | arne_wiebalck: ^^ | 16:58 |
| arne_wiebalck | rpioso: it could find things like "this hardware does not have a 'system power' entry in its tree" | 16:58 |
| arne_wiebalck | rpioso: which will then help with our choice of what to purchase :) | 16:59 |
| arne_wiebalck | rpioso: the main issue we have at the moment is to specify what we actually want | 17:01 |
| rpioso | arne_wiebalck, etingof: The sushy CI test suite could be enhanced to offer more coverage. It seems very light presently. | 17:03 |
| rpioso | And relies on an ironic workflow. | 17:04 |
| etingof | rpioso, the thing is that we want to take vendor X hardware and run ironic workflow against it to see where it chokes | 17:05 |
| arne_wiebalck | rpioso: but how do I run this against a sample node I got from a vendor? | 17:05 |
| arne_wiebalck | etingof: ++ | 17:05 |
| etingof | so we are looking to have ironic compatibility validation as a service | 17:06 |
| rpioso | arne_wiebalck, etingof: For vendor without 3d Party CIs? | 17:06 |
| arne_wiebalck | rpioso: yes | 17:06 |
| rpioso | arne_wiebalck: I see. | 17:07 |
| *** ijw has joined #openstack-ironic | 17:07 | |
| rpittau | have a nice weekend! o/ | 17:09 |
| *** rpittau is now known as rpittau|afk | 17:09 | |
| *** ociuhandu has quit IRC | 17:13 | |
| etingof | we can take redfish schemas, flip the fields that we require into being 'required', pull JSON documents from BMC endpoints and validate them against this more restrictive schema... that should catch missing fields. | 17:13 |
| etingof | however, to make sure that "write" Redfish operations actually work, we will probably need to actually execute them | 17:14 |
| etingof | that is: set boot mode/device, insert virtual CD, power cycle etc... | 17:15 |
| etingof | it would not be as good as actually deploying a node with ironic, but an approximation... | 17:16 |
| arne_wiebalck | etingof: There are probably two levels here: theoretical and practical compliance. | 17:16 |
| arne_wiebalck | etingof: Validating the schemas is already quite good I would think. | 17:16 |
| arne_wiebalck | etingof: But then we could also have real hardware, e.g. a sample server. pass a practical test :) | 17:17 |
| arne_wiebalck | etingof: I updated the story with your schema validation idea. | 17:21 |
| arne_wiebalck | bye everyone, have a good week-end o/ | 17:26 |
| etingof | arne_wiebalck, right! I did not mean schema validation does it all. it's just an easy way to catch missing bits | 17:26 |
| etingof | arne_wiebalck, o/ | 17:26 |
| arne_wiebalck | etingof: I think it is a very nice start and should probably catch everything but interpretation issues. | 17:27 |
| etingof | perhaps, except for the commands | 17:27 |
| arne_wiebalck | etingof: what do you mean? | 17:27 |
| etingof | like we discussed lately when one implementation exposes boot-target-override=continue, but then it rejects to accept the same value on POST | 17:28 |
| arne_wiebalck | ah, right | 17:29 |
| arne_wiebalck | yes | 17:29 |
| arne_wiebalck | thanks for explaining this once more | 17:29 |
| etingof | that's the second "active" part where sushy-cli can be utilized perhaps | 17:29 |
| *** k_mouza_ has quit IRC | 17:29 | |
| etingof | or may be it could be an ansible playbook? | 17:29 |
| arne_wiebalck | I think we should probably keep the dependencies to do the validation as low as possible. | 17:30 |
| * arne_wiebalck has never used ansible | 17:30 | |
| etingof | the good thing about a playbook is that it captures the workflow i.e the sequence of commands to execute | 17:31 |
| *** alexmcleod has quit IRC | 17:31 | |
| etingof | there used to be some degree of redfish support in upstream ansible | 17:31 |
| *** ociuhandu has joined #openstack-ironic | 17:32 | |
| arne_wiebalck | ah, ok | 17:33 |
| arne_wiebalck | etingof: rpioso: thanks for the thoughts and the discussion, bye o/ | 17:34 |
| *** ociuhandu has quit IRC | 17:35 | |
| *** ociuhandu has joined #openstack-ironic | 17:36 | |
| *** diurnalist has quit IRC | 17:37 | |
| *** ociuhandu has quit IRC | 17:38 | |
| *** ociuhandu has joined #openstack-ironic | 17:39 | |
| *** ociuhandu has quit IRC | 17:44 | |
| *** derekh has quit IRC | 18:02 | |
| *** kozhukalov has quit IRC | 18:10 | |
| *** kozhukalov has joined #openstack-ironic | 18:20 | |
| *** amoralej is now known as amoralej|off | 18:22 | |
| *** ociuhandu has joined #openstack-ironic | 18:34 | |
| *** diurnalist has joined #openstack-ironic | 18:37 | |
| *** ociuhandu has quit IRC | 18:39 | |
| *** astnbomb has quit IRC | 18:59 | |
| *** astnbomb has joined #openstack-ironic | 18:59 | |
| *** astnbomb has quit IRC | 18:59 | |
| *** astnbomb has joined #openstack-ironic | 19:00 | |
| *** diurnalist has quit IRC | 19:14 | |
| *** ijw has quit IRC | 19:19 | |
| *** diurnalist has joined #openstack-ironic | 19:24 | |
| *** astnbomb has quit IRC | 19:35 | |
| *** astnbomb has joined #openstack-ironic | 19:36 | |
| *** astnbomb has quit IRC | 19:39 | |
| *** astnbomb has joined #openstack-ironic | 19:39 | |
| *** astnbomb has quit IRC | 19:57 | |
| *** astnbomb has joined #openstack-ironic | 19:57 | |
| *** astnbomb_ has joined #openstack-ironic | 19:58 | |
| *** jawad_axd has joined #openstack-ironic | 20:06 | |
| *** jawad_axd has quit IRC | 20:10 | |
| *** astnbomb_ has quit IRC | 20:11 | |
| *** astnbomb has joined #openstack-ironic | 20:12 | |
| *** astnbomb has quit IRC | 20:13 | |
| *** astnbomb_ has joined #openstack-ironic | 20:13 | |
| *** naved001 has quit IRC | 20:27 | |
| *** k_mouza has joined #openstack-ironic | 20:36 | |
| *** k_mouza has quit IRC | 20:41 | |
| *** khansa has joined #openstack-ironic | 20:46 | |
| *** ociuhandu has joined #openstack-ironic | 20:48 | |
| *** ociuhandu has quit IRC | 20:52 | |
| *** khansa has quit IRC | 21:12 | |
| *** astnbomb_ has quit IRC | 21:13 | |
| *** astnbomb has joined #openstack-ironic | 21:13 | |
| *** astnbomb has quit IRC | 21:15 | |
| *** astnbomb_ has joined #openstack-ironic | 21:15 | |
| openstackgerrit | Merged openstack/ironic-python-agent master: Ignore None md5 checksum field https://review.opendev.org/709053 | 21:31 |
| *** ijw has joined #openstack-ironic | 21:44 | |
| *** astnbomb_ has quit IRC | 22:23 | |
| *** rloo has quit IRC | 22:27 | |
| *** kozhukalov has quit IRC | 22:28 | |
| openstackgerrit | Merged openstack/ironic master: Pre-shared agent token https://review.opendev.org/692609 | 22:29 |
| *** stevebaker has joined #openstack-ironic | 22:50 | |
| *** jdandrea has quit IRC | 23:09 | |
| *** kozhukalov has joined #openstack-ironic | 23:18 | |
| *** kozhukalov has quit IRC | 23:28 | |
| *** kozhukalov has joined #openstack-ironic | 23:34 | |
| *** stevebaker has quit IRC | 23:40 | |
| *** diurnalist has quit IRC | 23:41 | |
| *** diurnalist has joined #openstack-ironic | 23:42 | |
| *** jcoufal has quit IRC | 23:42 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!