| opendevreview | SONG SHUKUN proposed openstack/ironic master: [iRMC] Add SNMPv3 authentication functionality https://review.opendev.org/c/openstack/ironic/+/845347 | 02:53 |
|---|---|---|
| TheJulia | stevebaker[m]: ironic-tempest-wholedisk-bios-snmp-pxe has been my reliable reproducer, the conundrum is if a node CI lands on a host without vmx flags, we ust tinyipa instead | 03:21 |
| TheJulia | because... yay footprints | 03:21 |
| *** akahat|out is now known as akahat|ruck | 05:15 | |
| jssfr | Good morning #ironic. | 06:45 |
| ajya | kubajj: avoid rebooting devstack VM. Or if do, re-stack. devstack is not really supporting reboots, there could be more things that no longer work after reboots. Probably, could be fixed, but hasn't been focus for devstack. | 07:19 |
| kubajj | ajya: thanks | 07:33 |
| iurygregory | good morning Ironic | 11:05 |
| sshnaidm | Hi, all! I need help with Redfish and Dell, maybe someone dealt with this before. I'd like to mount VirtualMedia to Dell Idrac, then set Boot to "VirtualMedia" and restart Idrac. I suppose it's exactly what Ironic does with inspector agent. I need to it with Ansible or with python script. Do you have any examples or code where I can look at? (I can't make Ansible dell module to set boot to VirtualMedia :( ) | 11:32 |
| ajya | Hi sshnaidm which version of iDRAC are you using? | 11:33 |
| sshnaidm | ajya, mostly 9 | 11:35 |
| sshnaidm | some old might have 8 I suppose | 11:36 |
| sshnaidm | afaiu it's different scenario for old ones, but don't have the details | 11:36 |
| ajya | in older iDRACs standard Redfish did not work with virtual media, this is now fixed in newer ones for iDRAC 9. For iDRAC8 have to use OEM Redfish. | 11:37 |
| ajya | Which iDRAC firmware version do you have for iDRAC9? | 11:38 |
| ajya | The virtual media boot was fixed in 4.40.10.00 for Intel system, and in 6.00.00.00 for AMD | 11:39 |
| ajya | If you have older than those, I suggest to upgrade firmware first and try again. | 11:40 |
| ajya | And then have OEM version for iDRAC8 | 11:40 |
| ajya | ... or try OEM for both iDRAC 8, 9 | 11:43 |
| ajya | I can dig out the Redfish API calls for that | 11:44 |
| opendevreview | Verification of a change to openstack/ironic master failed: Fix Redfish RAID to update raid_config https://review.opendev.org/c/openstack/ironic/+/842979 | 11:55 |
| iurygregory | sshnaidm, any reason on why not use ironic for it? virtual media deployments just work in most scenarios | 12:03 |
| sshnaidm | ajya, thanks, checking now versions of idracs.. | 12:17 |
| sshnaidm | iurygregory, using ironic for one such task is a total overkill :) | 12:18 |
| sshnaidm | I'd prefer to borrow a code from ironic that does it actually | 12:18 |
| iurygregory | you can just use the redfish API in this case =) | 12:22 |
| sshnaidm | ajya, I have on 9: 5.00.00.00(Build 52), so should be fine | 12:22 |
| sshnaidm | iurygregory, that's exactly what I'm trying to do :) | 12:22 |
| sshnaidm | I need to understand which calls should I do | 12:23 |
| opendevreview | Verification of a change to openstack/ironic master failed: Fix Redfish RAID to update raid_config https://review.opendev.org/c/openstack/ironic/+/842979 | 12:23 |
| sshnaidm | and what is the procedure, should I reboot the idrac right after a boot config, or its it something to do (saw in some manual jobs clear) | 12:23 |
| sshnaidm | at least I can fix redfish/idrac Ansible modules with it | 12:24 |
| ajya | Ironic does something like: 1. Turns system off 2. Waits for it to be off 3. Mount ISO 4. Configure boot device to virtual CD 5. Turns system on | 12:32 |
| ajya | in Redfish jobs clearing is not absolutely necessary | 12:33 |
| ajya | it's useful when there are pending, previously not executed jobs in the queue that could block new jobs created, but in Redfish iDRAC does not create jobs here (unlike WS-Man) | 12:33 |
| ajya | when in doubt - clear | 12:33 |
| sshnaidm | ajya, ack, thanks | 12:44 |
| sshnaidm | are there redfish calls for set boot to virtualmedia somewhere in ironic code? | 12:49 |
| ajya | sshnaidm: it starts here - https://opendev.org/openstack/ironic/src/commit/8aaf2e08c074e786767a24adbbe3afec5ff622c3/ironic/drivers/modules/redfish/management.py#L228 | 12:57 |
| ajya | I can grep from logs the calls itself, as I was testing virtual media incidentally | 12:57 |
| sshnaidm | ajya, would very appreciate it | 13:02 |
| sshnaidm | ajya, log of which service is it? | 13:03 |
| ajya | sshnaidm: conductor with debug enabled | 13:04 |
| sshnaidm | ack, cool | 13:04 |
| TheJulia | good morning | 13:08 |
| TheJulia | sshnaidm: for what it is worth, your describing our managed introspection feature. Not inspector itself on it's own with the agent. | 13:09 |
| ajya | sshnaidm: from the logs https://paste.opendev.org/show/bVBlTqR7RyKcKWqfEN2a/ for iDRAC. Ironic goes to sushy that does more processing for other vendors, but this is what it is for iDRAC. | 13:11 |
| sshnaidm | TheJulia, yeah, right | 13:11 |
| ajya | morning, TheJulia | 13:11 |
| sshnaidm | ajya, amazing, thanks! | 13:12 |
| opendevreview | Aija Jauntēva proposed openstack/ironic master: Fix redfish-virtual-media for newer iDRACs https://review.opendev.org/c/openstack/ironic/+/848107 | 13:12 |
| opendevreview | Baptiste Jonglez proposed openstack/networking-generic-switch master: Try fixing a non-deterministic test failure on the CI https://review.opendev.org/c/openstack/networking-generic-switch/+/848161 | 17:01 |
| opendevreview | Mark Goddard proposed openstack/ironic stable/wallaby: deploy steps: restrict priority of in-band steps https://review.opendev.org/c/openstack/ironic/+/848163 | 17:04 |
| opendevreview | Mark Goddard proposed openstack/ironic master: deploy steps: restrict priority of in-band steps https://review.opendev.org/c/openstack/ironic/+/848164 | 17:06 |
| opendevreview | Julia Kreger proposed openstack/ironic master: DNM: Discover what is going on with dhcp https://review.opendev.org/c/openstack/ironic/+/847601 | 17:48 |
| TheJulia | iurygregory: fyi, network manager folks have asked for more debug logging | 17:51 |
| iurygregory | TheJulia, ack | 17:54 |
| *** undefined_ is now known as Guest3684 | 17:55 | |
| *** Guest3684 is now known as rcastillo_ | 17:55 | |
| *** rcastillo_ is now known as rcastillo | 17:57 | |
| gmann | iurygregory: rpittau TheJulia dtantsur as you know after the operator feedback we are moving towards the approach of dropping the 'scope' from RBAC goal. John brought a good point on keeping scope implementation in ironic - https://review.opendev.org/c/openstack/governance/+/847418/5..6/goals/selected/consistent-and-secure-rbac.rst#b224 | 21:34 |
| gmann | please check and provide your opinion on that, thanks | 21:35 |
| iurygregory | gmann, ack | 22:08 |
| TheJulia | gmann, I'm going to be honest here, and appoligize for not keeping close track of exactly what ya'll are doing... but some of that is reading like the entire access model is being redeisnged by committee... again. | 22:39 |
| TheJulia | gmann: commented | 22:51 |
| TheJulia | gmann: so, is the goal the TC is now operating under to just excise scope, because some discussion seems like excising/eliminating scope is the preference, where as other text is saying holding off for now, and then I'm wondering what drove it becuase the operators at the operators meetup seemed more like "it would be nice to have, but we like... need something generally usable overall" | 22:53 |
| opendevreview | Julia Kreger proposed openstack/ironic master: DNM: Discover what is going on with dhcp https://review.opendev.org/c/openstack/ironic/+/847601 | 23:13 |
| gmann | TheJulia: we are going as per operator feedback where they wanted to postponed it and finish project reader first. project reader are stuck with scope things which has all the complication with breaking the heat and NFV users | 23:26 |
| TheJulia | gmann: it sounds like the the underlying purpose and for lack of a better word to describe it, but the "scope of view" has kind of dissolved | 23:27 |
| TheJulia | I mean, it makes sense to focus on the needful, but I feel like things are just getting... for really lack of a better way to put it, kind of mixed up | 23:28 |
| TheJulia | at least, that is my I'm insanely busy quick takeaway reading some of the comments and some of the text changes | 23:28 |
| TheJulia | And that may be totally just me too. | 23:28 |
| TheJulia | could the end purpose still have value? I suspect so, but seems like people want a project scoped reader level of access to have equivelent end functionality based on some comments. | 23:29 |
| TheJulia | I don't know. I just feel confused if that makes sense at all | 23:29 |
| TheJulia | Maybe I just got the original intent hard wired in my brain and did the needful in ironic and I'm now looking on from the sidelines going "what?!?" | 23:30 |
| TheJulia | Actually, that is exactly it. :) | 23:30 |
| gmann | we had same impression in nova when we finished it all in ussuri and again decoupled the scope from check_str in yoga(so that scope disable can work) but how much complex it is for operator and breaking NFV users we need to go back one step. ultimately goal is here is do something that operator need and can use | 23:34 |
| TheJulia | how is it now "breaking" for nfv users? | 23:35 |
| TheJulia | Granted, flavor is likely the best example. It *should* be one thing if we started with the scoped model, history has resulted in it being another... and then having a tool like heat being used for both end usage and systems management is kind of the ultimate "well, that doesn't quite work exactly the way we hoped. | 23:36 |
| TheJulia | " | 23:36 |
| gmann | this one https://review.opendev.org/c/openstack/governance/+/847418/6/goals/selected/consistent-and-secure-rbac.rst#159 | 23:39 |
| gmann | from that only we started re-thinking on the scope. | 23:39 |
| TheJulia | ahh yeah, so because we're trying to enable everything to be done with heat, the new model is ultimately prevented by not encouraging evolution of the usage | 23:40 |
| TheJulia | I think the saying for this that comes to mind is that this is a "catch-22" | 23:40 |
| TheJulia | It feels like because we've not had a delineated service use boundrary | 23:43 |
| TheJulia | and what I mean by that, is while we have a service that enables much of that, it doesn't mean that is the right path to take, and we should be actively working towards the optimal usage. | 23:44 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!