| vanou | good morning ironic | 01:37 |
|---|---|---|
| opendevreview | Merged openstack/ironic bugfix/19.0: Add SNMPv3 authentication functionality https://review.opendev.org/c/openstack/ironic/+/864465 | 04:32 |
| opendevreview | Merged openstack/ironic bugfix/18.1: Add SNMPv3 authentication functionality https://review.opendev.org/c/openstack/ironic/+/864466 | 05:24 |
| opendevreview | SONG SHUKUN proposed openstack/ironic bugfix/19.0: Add support auth protocols for iRMC https://review.opendev.org/c/openstack/ironic/+/865946 | 07:36 |
| opendevreview | SONG SHUKUN proposed openstack/ironic stable/xena: Add support auth protocols for iRMC https://review.opendev.org/c/openstack/ironic/+/865947 | 07:42 |
| opendevreview | SONG SHUKUN proposed openstack/ironic bugfix/18.1: Add support auth protocols for iRMC https://review.opendev.org/c/openstack/ironic/+/865950 | 07:47 |
| opendevreview | SONG SHUKUN proposed openstack/ironic stable/wallaby: Add support auth protocols for iRMC https://review.opendev.org/c/openstack/ironic/+/865951 | 07:49 |
| arne_wiebalck | Good morning, Ironic! | 07:55 |
| rpittau | good morning ironic! o/ | 08:31 |
| janders | hey vanou arne_wiebalck rpittau and Ironic o/ | 08:39 |
| rpittau | hey janders :0 | 08:46 |
| kubajj | Good morning everyone | 09:34 |
| rpittau | hey kubajj :) | 09:39 |
| dtantsur | TheJulia: wdyt about silencing oslo_policy logs by default? (cc janders) | 09:47 |
| dtantsur | janders: we can actually override logs filtering in ironic-image, but I wonder what TheJulia thinks about it | 09:47 |
| janders | dtantsur ACK. And TY. Makes sense! | 09:47 |
| vanou | Hi arne_wiebalck rpittau janders kubajj dtantsur o/ | 09:52 |
| rpittau | hey vanou :) | 09:52 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Use ansible 6.x https://review.opendev.org/c/openstack/bifrost/+/865969 | 10:21 |
| dtantsur | janders, TheJulia, hmm, apparently I thought about it already: https://opendev.org/openstack/ironic/commit/7748b154b0477e42cd0937c9f52cbed33e7b7c8e | 10:29 |
| dtantsur | janders: don't we have this patch ^^? | 10:29 |
| dtantsur | it should go as far as 18.1.. | 10:30 |
| dtantsur | I have some local logs from a bugzilla that don't have oslo_policy debug logging | 10:31 |
| dtantsur | looking at the bug, it's OCP 4.7, which is ancient and not supported any more | 10:32 |
| opendevreview | Merged openstack/ironic-inspector stable/ussuri: Fix port id vs uuid in ValidateInterfacesHook https://review.opendev.org/c/openstack/ironic-inspector/+/863626 | 11:07 |
| opendevreview | kamlesh chauvhan proposed openstack/ironic-tempest-plugin master: OOB inspection timeout fix for iDRAC https://review.opendev.org/c/openstack/ironic-tempest-plugin/+/865803 | 11:46 |
| iurygregory | good morning Ironic o/ | 11:48 |
| janders | hey iurygregory | 13:22 |
| janders | ajya I have a quick iDRAC question - are you aware on any limits on number of characters in the hostname (or the URL) of the virtual media source http(s) server? stevebaker[m] and I spent a fair bit of time troubleshooting something related today and it seems that the difference between a URL that works and the one that does not work is the length | 13:23 |
| janders | It's an R640 with iDRAC 6.00.30.00 | 13:24 |
| janders | stevebaker[m] we may want re-test on 5.10.x iDRAC firmware on rpittau's advice - it seems more stable | 13:30 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Use ansible 6.x https://review.opendev.org/c/openstack/bifrost/+/865969 | 13:33 |
| ajya | janders: What length did you try? Not aware if there is a limit, but it's likely that internally there is a data type restricting it. What errors do you see? | 13:34 |
| janders | sample error: https://paste.openstack.org/show/bsGMi4kPonUH2njQQmNk/ | 13:35 |
| janders | looks like up to 64 chars worked (just for the hostname, not the whole URL) | 13:36 |
| janders | beyond that did not | 13:36 |
| janders | will PM you some details (not sharing here due to hostnames) | 13:36 |
| opendevreview | Harald Jensås proposed openstack/ironic master: benchmark: Create allocations for nodes https://review.opendev.org/c/openstack/ironic/+/865988 | 13:37 |
| opendevreview | Harald Jensås proposed openstack/ironic master: Use association_proxy for ports node_uuid https://review.opendev.org/c/openstack/ironic/+/862933 | 13:45 |
| opendevreview | Harald Jensås proposed openstack/ironic master: Use association_proxy for port groups node_uuid https://review.opendev.org/c/openstack/ironic/+/864781 | 13:45 |
| opendevreview | Harald Jensås proposed openstack/ironic master: Use association_proxy for node chassis_uuid https://review.opendev.org/c/openstack/ironic/+/864802 | 13:45 |
| opendevreview | Harald Jensås proposed openstack/ironic master: Use association_proxy for node allocation_uuid https://review.opendev.org/c/openstack/ironic/+/865989 | 13:45 |
| ajya | janders, stevebaker I'm suspecting this will be same on all iDRAC versions, especially, in older versions if still observed in newer. Indeed there could be some data type limiting host to 64 chars. Why 64 chars? Maybe because hostnames often can't be more than 64 chars, and here it does not distinguish between subdomains. | 14:01 |
| ajya | What is happening - it probably gets silently truncated and then can't find the file as it's not the same URL anymore. I will check with firmware team if they can fix this. Are there any BZs or Jiras I can link to? Also I found this - https://bugzilla.redhat.com/show_bug.cgi?id=1956421 that looks similar issue | 14:02 |
| hjensas | TheJulia: I added chassis and allocation to benchmark and added patches to use association_proxy SQL join for chassis_uuid and allocation_uuid. benchmark stats here: https://paste.opendev.org/show/bR0EaLEjG5NVr8Kp1f9r/ | 14:04 |
| hjensas | TheJulia: it is ever so slightly slower when fields are used, but also quite a bit of savings when no fields are used. | 14:06 |
| TheJulia | a 100th of a second is... in the territory of a standard deviation | 14:13 |
| TheJulia | not bad at all | 14:13 |
| TheJulia | hjensas: it is 66% faster, which is still not bad at all | 14:15 |
| hjensas | TheJulia: ~35% faster-ish, it taks ~66% of the previous time. No? | 14:19 |
| hjensas | still, not bad. | 14:19 |
| TheJulia | it could be 66ish of the previous time | 14:20 |
| TheJulia | I'm still waking up and just starting to sip coffee | 14:20 |
| TheJulia | fixing my ratio compare is 40% | 14:21 |
| TheJulia | at least in terms of nodes per second | 14:21 |
| opendevreview | Harald Jensås proposed openstack/ironic master: benchmark: Create allocations for nodes https://review.opendev.org/c/openstack/ironic/+/865988 | 15:03 |
| * dtantsur is wondering at which point of the ongoing benchmarking we get to "rewrite the conductor in rust" :D | 15:09 | |
| JayF | dtantsur: when we get past the low hanging fruit. Right now it's like we have a whole harvest on the ground waiting to be picked up :P | 15:14 |
| TheJulia | I think the rules require that we take a break between those two actions, and *then* enjoy a delicous cup of pumpkin spiced coffee | 15:15 |
| dtantsur | a very good idea! | 15:25 |
| JayF | is disliking "pumpkin spice *" an unpopular opinion/ | 15:26 |
| dtantsur | seems so | 15:26 |
| TheJulia | heh | 15:26 |
| TheJulia | I left the pumpkin spiced coffee at the RV which I won't see for like a month | 15:27 |
| TheJulia | stevebaker[m]: numerous small fixes posted to https://review.opendev.org/q/project:x%252Fproliantutils hopefully that should shed some clarity. If you want to kick that downstream bug over to me, feel free | 15:28 |
| rpittau | TheJulia: we may want to merge this first https://review.opendev.org/c/x/proliantutils/+/865956 | 15:30 |
| TheJulia | stevebaker[m]: absolutely *no idea* when they will be merged | 15:30 |
| TheJulia | rpittau: on master, sure | 15:30 |
| TheJulia | This stuff goes all the way back to wallaby | 15:30 |
| rpittau | ah right | 15:30 |
| opendevreview | Jay Faulkner proposed openstack/ironic-lib master: No longer override install_command in tox.ini https://review.opendev.org/c/openstack/ironic-lib/+/866032 | 15:38 |
| opendevreview | Jay Faulkner proposed openstack/python-ironicclient master: No longer override install_command in tox.ini https://review.opendev.org/c/openstack/python-ironicclient/+/866033 | 15:39 |
| JayF | ^^ there was a post on the mail list about how this is a misconfiguration | 15:41 |
| JayF | so I fixed it | 15:41 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Fix deprecated module ansible lint error https://review.opendev.org/c/openstack/bifrost/+/866040 | 16:03 |
| rpittau | need to split, see ya all tomorrow! o/ | 16:03 |
| TheJulia | do I get to feel old because I remember the last time we did fixes to install_command in tox.ini? | 16:07 |
| dtantsur | then please welcome me in the old club :) | 16:08 |
| * TheJulia starts a club, makes hot chocolate and passes it around | 16:13 | |
| kubajj | Can I join the club even if I don't remember it? | 16:13 |
| TheJulia | yes! | 16:13 |
| kubajj | I'll come for the hot chocolate then | 16:14 |
| dtantsur | yummy! | 16:32 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: DB & Object layer for of shards & node.shard_key https://review.opendev.org/c/openstack/ironic/+/864236 | 16:35 |
| TheJulia | Has anyone thought of ironic specific topics for Open Infra Summit? | 16:41 |
| dtantsur | Summit specifically? not really (esp. since my participation is a very wide open question) | 16:42 |
| dtantsur | you mean, Forum? | 16:42 |
| TheJulia | Summit | 16:42 |
| TheJulia | Maybe a better question: What topics might be interesting? | 16:42 |
| dtantsur | As a track chair, I should probably not provide specific suggestions :) | 16:43 |
| TheJulia | oh! Congrats! I was not aware! | 16:43 |
| dtantsur | (I'm also the only track chair, apparently, since my peer withdrew) | 16:44 |
| TheJulia | ugh | 16:44 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: DB & Object layer for node.shard https://review.opendev.org/c/openstack/ironic/+/864236 | 16:57 |
| TheJulia | JayF: anything else on the shard key spec that you think needs to change based upon your work so far? | 16:58 |
| TheJulia | beyond the 6 comments on there? | 16:58 |
| JayF | I've been trying to put comments on there as I go | 17:00 |
| TheJulia | ack ack | 17:01 |
| JayF | TheJulia: while you're working on shards; mind looking @ something w/me? | 17:04 |
| JayF | TheJulia: Trying to get my sharding PR fixed up (no more Shard table/object) and I'm getting head-scratching unit test failures | 17:04 |
| JayF | like I broke the whole node object in the API | 17:04 |
| JayF | https://review.opendev.org/c/openstack/ironic/+/864236 broken code is here | 17:04 |
| JayF | only hint I have it | 17:05 |
| JayF | **is | 17:05 |
| JayF | webtest.app.AppError: Bad response: 500 Internal Server Error (not 200 OK or 3xx redirect for http://localhost/v1/nodes) | 17:05 |
| JayF | b'{"error_message": "{\\"faultcode\\": \\"Server\\", \\"faultstring\\": \\"\'Node\' object has no attribute \'shardstorage_interface\'\\", \\"debuginfo\\": null}"}' | 17:05 |
| TheJulia | that looks like a solid typo | 17:05 |
| JayF | tell that to grep | 17:06 |
| JayF | if it's an err; I don't know where it comes from | 17:06 |
| TheJulia | https://review.opendev.org/c/openstack/ironic/+/864236/7/ironic/api/controllers/v1/node.py | 17:06 |
| TheJulia | missing typo second half | 17:06 |
| TheJulia | err | 17:06 |
| TheJulia | missing comma | 17:06 |
| JayF | I've only looked at that code 47 times without seeing that lolsob | 17:06 |
| JayF | might be a little optimistic; but with the no-mo-caching, I'm really going to try and get this hooked up completely today | 17:07 |
| JayF | so we need to get that spec landed so I can land this stuf lol | 17:07 |
| * TheJulia shrugs about need to have something before something else | 17:08 | |
| * TheJulia is laid back | 17:08 | |
| TheJulia | or at least... tries to be | 17:08 |
| JayF | There's "laid back" and there's "yolo landing API microversions without the in-review spec landing" | 17:10 |
| JayF | IDK exactly where my line is, and I think that's just over it | 17:10 |
| * TheJulia shrugs | 17:25 | |
| TheJulia | we've done worse before, trust me :) | 17:25 |
| TheJulia | speaking of things, has anyone heard of status on service role stuffs?\ | 17:28 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: DB & Object layer for node.shard https://review.opendev.org/c/openstack/ironic/+/864236 | 17:36 |
| dtantsur | folks, I'm on PTO until end of week, see you on Monday | 17:48 |
| kubajj | TheJulia: if I add something to the conf/inspector.py do I need to tell ironic somewhere to load this value or is this handled on its own? I am can't find a change similar to what I am trying to do | 17:59 |
| opendevreview | Jay Faulkner proposed openstack/ironic master: DB & Object layer for node.shard https://review.opendev.org/c/openstack/ironic/+/864236 | 18:03 |
| opendevreview | Jakub Jelinek proposed openstack/ironic master: WIP: Enable alternative storage for inventory https://review.opendev.org/c/openstack/ironic/+/866056 | 18:30 |
| kubajj | Is Swift used for anything in Ironic now? | 19:15 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Fix deprecated module ansible lint error https://review.opendev.org/c/openstack/bifrost/+/866040 | 19:16 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: Add a shard key https://review.opendev.org/c/openstack/ironic-specs/+/861803 | 19:20 |
| TheJulia | kubajj: temporary object storage | 19:21 |
| TheJulia | kubajj: if it is in the list, it is automatically available on the CONF object | 19:21 |
| kubajj | TheJulia: that is quite cool | 19:22 |
| TheJulia | there is a method call someplace towards the end of the file like register_opts | 19:22 |
| TheJulia | that does the magic :) | 19:22 |
| JayF | TheJulia: OK; so my code (and the comment on the spec) said I was going to use node.shard | 19:24 |
| JayF | TheJulia: do you feel strongly about node.shard_key? I can rework my patch for the [needs a second hand to count] time :P | 19:24 |
| TheJulia | oh | 19:26 |
| TheJulia | I thought you meant you used node.shard_key | 19:27 |
| TheJulia | uhhh.... shard is fine I guess | 19:27 |
| JayF | I neverr was good at the game of "telephone" :P | 19:27 |
| JayF | lol | 19:27 |
| * TheJulia tries to think of reasons against it | 19:27 | |
| JayF | shard_key is a bad name because it's a key by definition | 19:27 |
| JayF | we don't say "conductor_group_key" | 19:27 |
| TheJulia | yeah | 19:27 |
| TheJulia | yeah, makes sense | 19:28 |
| JayF | that's at least how I landed on `node.shard` | 19:28 |
| TheJulia | microwaving some lunch and I'll fix the spec | 19:30 |
| JayF | I'm excited, DB+Obj layer done, going to plug up the API then time to test | 19:30 |
| opendevreview | Julia Kreger proposed openstack/ironic-specs master: Add a shard key https://review.opendev.org/c/openstack/ironic-specs/+/861803 | 19:39 |
| stevebaker[m] | good morning | 19:57 |
| stevebaker[m] | TheJulia: I've reviewed those proliantutils changes | 19:57 |
| TheJulia | dtantsur: do you want me to just merge the inspector sqlalchemy changes at this point since you seem to want the first patch to have the fixes in the second patch... and in order to do it I have to cult over basically the entire content of the second patch.... soo.... | 20:10 |
| TheJulia | stevebaker[m]: thanks | 20:10 |
| opendevreview | Julia Kreger proposed openstack/ironic stable/zed: Change boot_interface order of iRMC driver https://review.opendev.org/c/openstack/ironic/+/865998 | 20:19 |
| JayF | TheJulia: I put a -1 on ^^^, I'm available for the next 2-3 hours if you wanna try and talk me out of it :) | 20:54 |
| stevebaker[m] | TheJulia: I think my esp.img is correct because grub boots, but it drops into the grub shell, like it can't find grub.cfg https://pasteboard.co/83Is3ylp2jqa.png | 21:59 |
| stevebaker[m] | should it be in /boot/grub2/grub.cfg rather than /boot/grub/grub.cfg? | 21:59 |
| JayF | yes | 21:59 |
| JayF | hmm, or not | 22:00 |
| JayF | on my machine; grub2; my files are at /boot/grub/grub.cfg | 22:00 |
| stevebaker[m] | oh, its an ironic.conf option! | 22:01 |
| JayF | oh, nice! | 22:02 |
| stevebaker[m] | this is for the boot iso that ironic builds for virtual-media-boot | 22:04 |
| stevebaker[m] | it worked \o/ | 22:19 |
| TheJulia | JayF: wrt that change, to me it kind of makes sense given that zed is still fairly new, and otherwise you can run into misconfigurations where you need entirely separate config or node specific overrides to try and navigate | 22:49 |
| TheJulia | JayF: in other words, not super friendly for operators who seek consistency | 22:50 |
| JayF | I figured that'd be the answer; but what about the operators who want it configured that way? We're going to break them with minimal notice (hope they read changelogs) | 22:51 |
| TheJulia | stevebaker[m]: \o/ | 22:51 |
| JayF | and if nobody in their right mind wants it configured that way; why is it supported in master? | 22:51 |
| JayF | AFAICT; we're just setting into stone a preference -- one that I hold as well -- but it's hard to justify that for already-released software | 22:52 |
| stevebaker[m] | metal3 and bifrost set grub_config_path to =EFI/{{ efi_distro }}/grub.cfg, but I just tried grub_config_path=EFI/BOOT/grub.cfg and it worked, so a distro agnostic UEFI default seems to exist | 22:53 |
| stevebaker[m] | JayF: I think its a throwback to Legacy BIOS defaults, we should change that now that UEFI is the default | 22:54 |
| JayF | I mean, I'm onboard with that for master; I just question the wisdom of breaking already-deployed Zed operators who may prefer the current default | 22:54 |
| JayF | I'm all for a loose stable policy; but causing people who could be currently working to have to reconfigure to maintain existing behavior seems really, really unfriendly to me | 22:54 |
| TheJulia | so, the conundrum is FJ is actually asking me for it | 22:55 |
| JayF | (in context of a stable branch) | 22:55 |
| stevebaker[m] | (soz I was talking about a different thing) | 22:55 |
| TheJulia | heh | 22:55 |
| JayF | stevebaker[m]: oooh | 22:55 |
| TheJulia | I think their visibility/testing of it has lagged behind release | 22:55 |
| JayF | stevebaker[m]: makes sense; although I prefer the EFI-is-boot-partition setup; but most distros don't :P | 22:55 |
| JayF | TheJulia: I mean, is it true that: 1) There is a use case for the existing, default configuration. 2) Users who want that configuration would need to reconfigure when updating to latest-stable if we pushed that change? | 22:56 |
| TheJulia | that being said the gray area for stable is operator friendly driver fixes..... soo... worth raising the question :) | 22:56 |
| JayF | TheJulia: I said I wasn't sure where my line was, I think ^^^^ that might be it for stable lol | 22:56 |
| TheJulia | the existing default config doesn't work if pxe is not enabled | 22:56 |
| TheJulia | that is the conundrum | 22:57 |
| stevebaker[m] | JayF: I believe the interface order is wrong to the point of it being broken by default | 22:57 |
| TheJulia | I think people have been routing around it without actually going "hey, this is broken" | 22:57 |
| stevebaker[m] | oh, if no pxe | 22:57 |
| JayF | Could we do something like, detect the broken configuration and flip behavior only in that case? That's probably not worth it from complexity | 22:57 |
| TheJulia | we would have no way to really know if it was intentional or not | 22:58 |
| TheJulia | and we would be enabling driver features if their config was intentional and one person didn't talk to the ironic admin | 22:58 |
| JayF | Is there a case/configuration, that makes sense for a user to have, that would break if we deployed this change? | 22:58 |
| JayF | that's the fulcrum of my concern | 22:58 |
| TheJulia | not really, since we're just changing the supported matrix | 23:00 |
| TheJulia | not the actual configured preference | 23:00 |
| TheJulia | and we're adding ipxe to that list, and the first option remains the same which is a vendor preferred method | 23:00 |
| TheJulia | if we removed an option from that list, that would be very breaking | 23:01 |
| JayF | what /is/ boot.IRMCPXEBoot and how is it different than usual pxe boot? /me goes to ask the code | 23:01 |
| JayF | it just backs up bios config for use in cleaning, interesting | 23:02 |
| TheJulia | yeah, and they don't want to adapt to the formal bios interface either :( | 23:02 |
| JayF | Why do we give them a choice? | 23:02 |
| JayF | We have a place for drivers that don't wanna work in our model. | 23:02 |
| TheJulia | that code was from *long* before we had the bios interface | 23:03 |
| JayF | so the only breaky situation I can think of with this | 23:03 |
| JayF | is if I intentionally want pxe, not ipxe | 23:03 |
| JayF | and I customized my pxe template, but not my ipxe template | 23:03 |
| JayF | yeah? | 23:03 |
| TheJulia | and you were *expecting the buggy behavior even though your enabled_boot_interfaces parameter was set to ipxe, pxe, magical-virtual-media-3 | 23:04 |
| TheJulia | that enabled list always take precedence as long as it is supported | 23:04 |
| JayF | so there's no way, outside of disabling pxe globally or merging 865998, to get that boot IF to use pxe? | 23:05 |
| TheJulia | to use ipxe | 23:05 |
| TheJulia | and correct | 23:05 |
| TheJulia | just... ipxe | 23:05 |
| JayF | okay, so if we merged this, I'm looking at the changelog and am concerned it's not ... noisy enough | 23:06 |
| TheJulia | heh | 23:06 |
| JayF | is there any way we can make that more directive+clear to operators of zed/other stable branches? | 23:06 |
| * JayF brb plumbers here | 23:07 | |
| TheJulia | funny you said that, because my draft review comments I somehow never posted were basically like "wow, this is a bit much" | 23:07 |
| TheJulia | JayF: likely... less words actually, the reno is very verbose | 23:07 |
| TheJulia | at least, that is my thought | 23:07 |
| JayF | So lets make this a clearer release note | 23:12 |
| JayF | and then *immediately* cut a **minor** (not patch) release of Ironic | 23:13 |
| JayF | to ensure it gets shipped as a minor release bump | 23:13 |
| TheJulia | sounds good to me | 23:13 |
| TheJulia | openstack release team might argue... depending on their mood and the phase of the moon | 23:14 |
| JayF | nah, elodilles suggested we do a minor release at one point b/c we added a dep | 23:15 |
| JayF | so I suspect there'll be some appreciation that we thought about it at all :P | 23:15 |
| JayF | TheJulia: https://review.opendev.org/c/openstack/ironic/+/865998/1#message-93140ce6955d15a0eacb26e5d697bfc7e1875d3b | 23:15 |
| TheJulia | ack, I'll rev it tomorrow | 23:19 |
| TheJulia | going to go paint now | 23:19 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
