| *** bfournie has joined #openstack-ironic | 00:00 | |
| *** mtreinish has quit IRC | 00:12 | |
| *** ijw has quit IRC | 00:14 | |
| *** Goneri has quit IRC | 00:23 | |
| *** yangyapeng has joined #openstack-ironic | 00:25 | |
| *** r-daneel has quit IRC | 00:29 | |
| *** yangyapeng has quit IRC | 00:30 | |
| *** mtreinish has joined #openstack-ironic | 00:37 | |
| *** ijw has joined #openstack-ironic | 00:38 | |
| *** rbudden has quit IRC | 00:39 | |
| *** rloo has quit IRC | 00:46 | |
| *** edleafe has quit IRC | 00:48 | |
| *** edleafe has joined #openstack-ironic | 00:51 | |
| *** edleafe has quit IRC | 00:55 | |
| *** ccollett has joined #openstack-ironic | 00:59 | |
| *** charcol-v2 has quit IRC | 01:02 | |
| *** edleafe has joined #openstack-ironic | 01:11 | |
| *** zshi has joined #openstack-ironic | 01:17 | |
| *** yangyapeng has joined #openstack-ironic | 01:22 | |
| openstackgerrit | Merged openstack/ironic-python-agent master: Replace use of functools.wraps() with six.wraps() https://review.openstack.org/540175 | 01:27 |
|---|---|---|
| *** mjturek has quit IRC | 01:38 | |
| *** etingof has quit IRC | 01:39 | |
| *** ijw has quit IRC | 01:42 | |
| TheJulia | regarding grenade, I've totally lost my mind digging into it tonight. it _looks_ like two different network issues with similar results. the commonality between the node and the qrouter is missing, FIP inbound no reply. one shows no dhcp logs, another two show that the nodes booted with dhcp before the failure... and observed on the subnode console/compute/conductor processes | 01:48 |
| TheJulia | dhcp being on the primary node, meaning traffic is crossing over the vlan between the two | 01:48 |
| *** zhangfei has joined #openstack-ironic | 01:51 | |
| *** etingof has joined #openstack-ironic | 01:52 | |
| *** zhangfei has quit IRC | 01:53 | |
| *** zhangfei has joined #openstack-ironic | 01:53 | |
| *** sapd__ has quit IRC | 02:04 | |
| *** zhangfei has quit IRC | 02:04 | |
| *** sapd has joined #openstack-ironic | 02:06 | |
| *** zhangfei has joined #openstack-ironic | 02:18 | |
| *** harlowja has quit IRC | 02:18 | |
| *** tuanla____ has joined #openstack-ironic | 02:53 | |
| *** d0ugal_ has joined #openstack-ironic | 03:16 | |
| *** d0ugal has quit IRC | 03:19 | |
| *** gcb has joined #openstack-ironic | 03:26 | |
| *** crushil has joined #openstack-ironic | 03:48 | |
| *** hshiina has joined #openstack-ironic | 03:56 | |
| *** dsneddon has quit IRC | 03:59 | |
| *** dsneddon has joined #openstack-ironic | 04:03 | |
| *** dsneddon has quit IRC | 04:11 | |
| *** devananda has quit IRC | 04:22 | |
| *** vsaienk0 has joined #openstack-ironic | 04:25 | |
| *** dsneddon has joined #openstack-ironic | 04:29 | |
| *** rbudden has joined #openstack-ironic | 04:30 | |
| openstackgerrit | Olivier Bourdon proposed openstack/bifrost master: Fix for proper deployment when heystone enabled and minimal OS https://review.openstack.org/540713 | 04:31 |
| openstackgerrit | Olivier Bourdon proposed openstack/bifrost master: Add creation of openrc file when keystone enabled https://review.openstack.org/540096 | 04:32 |
| *** rbudden has quit IRC | 04:34 | |
| *** vsaienk0 has quit IRC | 04:35 | |
| *** dsneddon has quit IRC | 04:37 | |
| *** dsneddon has joined #openstack-ironic | 04:38 | |
| *** crushil has quit IRC | 04:38 | |
| *** crushil has joined #openstack-ironic | 04:42 | |
| *** dsneddon has quit IRC | 04:43 | |
| *** harlowja has joined #openstack-ironic | 04:46 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic master: cleanup: Remove usage of some_dict.keys() https://review.openstack.org/541123 | 04:46 |
| *** zhangfei has quit IRC | 04:46 | |
| *** dsneddon has joined #openstack-ironic | 04:59 | |
| *** zhangfei has joined #openstack-ironic | 05:03 | |
| *** dsneddon has quit IRC | 05:08 | |
| *** dpawar has joined #openstack-ironic | 05:10 | |
| *** harlowja has quit IRC | 05:11 | |
| *** links has joined #openstack-ironic | 05:12 | |
| *** tuanla____ has quit IRC | 05:23 | |
| *** sdake_ is now known as sdake | 05:30 | |
| *** marios has joined #openstack-ironic | 05:45 | |
| *** dsneddon has joined #openstack-ironic | 05:53 | |
| *** zhangfei has quit IRC | 05:54 | |
| *** hw_wutianwei has quit IRC | 05:59 | |
| *** hw_wutianwei has joined #openstack-ironic | 05:59 | |
| *** gcb has quit IRC | 06:00 | |
| *** e0ne has joined #openstack-ironic | 06:04 | |
| *** zhangfei has joined #openstack-ironic | 06:07 | |
| *** charcol-v2 has joined #openstack-ironic | 06:08 | |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic-inspector master: Introducing an api_workers option to inspector https://review.openstack.org/540722 | 06:09 |
| *** ccollett has quit IRC | 06:10 | |
| *** d0ugal_ has quit IRC | 06:14 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic-inspector master: Imported Translations from Zanata https://review.openstack.org/541138 | 06:19 |
| *** d0ugal_ has joined #openstack-ironic | 06:23 | |
| *** ccollett has joined #openstack-ironic | 06:24 | |
| *** charcol-v2 has quit IRC | 06:26 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic-ui master: Imported Translations from Zanata https://review.openstack.org/541143 | 06:36 |
| *** crushil has quit IRC | 06:38 | |
| *** threestrands has quit IRC | 06:40 | |
| openstackgerrit | wudong proposed openstack/ironic master: Modify error quotation marks https://review.openstack.org/541146 | 06:41 |
| *** yolanda has joined #openstack-ironic | 06:43 | |
| *** ianychoi_ has quit IRC | 06:47 | |
| *** ianychoi_ has joined #openstack-ironic | 06:48 | |
| *** ccollett has quit IRC | 06:57 | |
| *** rwsu has joined #openstack-ironic | 07:01 | |
| *** olivierbourdon38 has quit IRC | 07:06 | |
| *** vsaienk0 has joined #openstack-ironic | 07:09 | |
| *** gcb has joined #openstack-ironic | 07:13 | |
| *** dsneddon has quit IRC | 07:17 | |
| *** crushil has joined #openstack-ironic | 07:19 | |
| *** dsneddon has joined #openstack-ironic | 07:20 | |
| *** mjura has joined #openstack-ironic | 07:22 | |
| *** e0ne has quit IRC | 07:22 | |
| openstackgerrit | wudong proposed openstack/ironic master: Modify error quotation marks https://review.openstack.org/541146 | 07:25 |
| *** AlexeyAbashkin has joined #openstack-ironic | 07:32 | |
| *** rcernin has quit IRC | 07:37 | |
| *** AlexeyAbashkin has quit IRC | 07:39 | |
| *** rbartal has joined #openstack-ironic | 07:41 | |
| *** olivierbourdon38 has joined #openstack-ironic | 07:42 | |
| *** AlexeyAbashkin has joined #openstack-ironic | 07:48 | |
| *** vsaienk0 has quit IRC | 07:48 | |
| *** vsaienk0 has joined #openstack-ironic | 07:49 | |
| openstackgerrit | Hironori Shiina proposed openstack/ironic master: Update iRMC document for classic driver deprecation https://review.openstack.org/541171 | 08:02 |
| *** v12aml has quit IRC | 08:04 | |
| *** v12aml has joined #openstack-ironic | 08:05 | |
| *** pcaruana has joined #openstack-ironic | 08:14 | |
| *** tesseract has joined #openstack-ironic | 08:22 | |
| *** dsneddon has quit IRC | 08:23 | |
| *** dsneddon has joined #openstack-ironic | 08:27 | |
| *** ianychoi_ has quit IRC | 08:27 | |
| *** ianychoi_ has joined #openstack-ironic | 08:28 | |
| *** amarao has joined #openstack-ironic | 08:34 | |
| *** dsneddon has quit IRC | 08:34 | |
| *** dsneddon has joined #openstack-ironic | 08:36 | |
| *** frti has joined #openstack-ironic | 08:48 | |
| *** crushil has quit IRC | 08:49 | |
| openstackgerrit | Harald Jensås proposed openstack/networking-baremetal master: Make the agent distributed using hashring and notifications https://review.openstack.org/524709 | 08:53 |
| *** mgoddard_ has joined #openstack-ironic | 08:54 | |
| *** priteau has joined #openstack-ironic | 08:56 | |
| *** hshiina is now known as hshiina|afk | 08:56 | |
| *** d0ugal_ has quit IRC | 08:57 | |
| *** d0ugal has joined #openstack-ironic | 08:57 | |
| *** d0ugal has quit IRC | 08:57 | |
| *** d0ugal has joined #openstack-ironic | 08:57 | |
| *** dsneddon has quit IRC | 08:59 | |
| Tengu | hello! would a systemd unit file and starting script be of any use for virtualBMC project? Just had to create both today. Would gladly contrib them. | 09:00 |
| *** hshiina|afk has quit IRC | 09:03 | |
| openstackgerrit | James E. Blair proposed openstack/ironic master: Zuul: Remove project name https://review.openstack.org/541190 | 09:04 |
| *** e0ne has joined #openstack-ironic | 09:05 | |
| *** threestrands has joined #openstack-ironic | 09:06 | |
| openstackgerrit | Madhuri Kumari proposed openstack/ironic-specs master: Add service management API for ironic https://review.openstack.org/471217 | 09:07 |
| *** dsneddon has joined #openstack-ironic | 09:09 | |
| sapd | Hi everyone. I have a supermicro service with Intel Raid support. I don't know how to launch an baremetal instance on this server with raid support. In the guide, I think I have to build agent ramdisk which support raid. But I can't find what driver? | 09:12 |
| openstackgerrit | ShangXiao proposed openstack/ironic master: [api-ref] Add fault names for the error response codes https://review.openstack.org/538717 | 09:13 |
| *** dsneddon has quit IRC | 09:14 | |
| *** dsneddon has joined #openstack-ironic | 09:15 | |
| *** dsneddon has quit IRC | 09:21 | |
| *** threestrands has quit IRC | 09:21 | |
| *** frti_ has joined #openstack-ironic | 09:21 | |
| *** frti_ has quit IRC | 09:25 | |
| *** frti has quit IRC | 09:25 | |
| *** dsneddon has joined #openstack-ironic | 09:27 | |
| *** hoonetorg has quit IRC | 09:28 | |
| *** dsneddon has quit IRC | 09:35 | |
| openstackgerrit | Merged openstack/ironic-inspector master: Imported Translations from Zanata https://review.openstack.org/541138 | 09:39 |
| *** hoonetorg has joined #openstack-ironic | 09:41 | |
| openstackgerrit | MaoyangLiu proposed openstack/ironic master: Replace chinese quotes to English quotes https://review.openstack.org/541206 | 09:43 |
| *** zhangfei has quit IRC | 09:44 | |
| *** derekh has joined #openstack-ironic | 09:45 | |
| *** zhangfei has joined #openstack-ironic | 09:56 | |
| openstackgerrit | Harald Jensås proposed openstack/networking-baremetal master: Add unit tests for member manager https://review.openstack.org/541216 | 09:57 |
| *** azaid_ has joined #openstack-ironic | 10:00 | |
| azaid_ | Hi , I have a problem in ironic which hypvisor have no cpu and no ram | 10:00 |
| azaid_ | so when I create instance it return no valid hosts | 10:01 |
| azaid_ | any idea ? | 10:01 |
| openstackgerrit | Harald Jensås proposed openstack/networking-baremetal master: Add unit tests for member manager https://review.openstack.org/541216 | 10:01 |
| *** jaianshu has joined #openstack-ironic | 10:14 | |
| *** zshi has quit IRC | 10:15 | |
| *** nmathew has joined #openstack-ironic | 10:19 | |
| *** dtantsur|afk is now known as dtantsur | 10:25 | |
| openstackgerrit | Olivier Bourdon proposed openstack/bifrost master: Add creation of openrc file when keystone enabled https://review.openstack.org/540096 | 10:25 |
| openstackgerrit | Olivier Bourdon proposed openstack/bifrost master: Add creation of openrc file when keystone enabled https://review.openstack.org/540096 | 10:27 |
| *** zshi has joined #openstack-ironic | 10:27 | |
| dtantsur | hi ironic | 10:28 |
| -openstackstatus- NOTICE: Our Zuul infrastructure is currently experiencing some problems and processing jobs very slowly, we're investigating. Please do not approve or recheck changes for now. | 10:30 | |
| openstackgerrit | Olivier Bourdon proposed openstack/bifrost master: Fix for proper deployment when keystone enabled and minimal OS https://review.openstack.org/540713 | 10:31 |
| *** azaid_ has quit IRC | 10:38 | |
| openstackgerrit | Mark Goddard proposed openstack/networking-generic-switch master: Add initial release note https://review.openstack.org/534201 | 10:43 |
| *** lucas-afk is now known as lucasagomes | 10:52 | |
| *** sambetts|afk is now known as sambetts | 10:54 | |
| sambetts | Morning all | 10:54 |
| sambetts | Tengu: I think that would be useful thanks! | 10:55 |
| sambetts | dtantsur: Morning o/ | 10:55 |
| dtantsur | o/ | 11:01 |
| dtantsur | Tengu: hi! sync with etingof, he was working on making vbmc a normal service | 11:01 |
| etingof | o/ | 11:02 |
| etingof | Tengu, here is my patch being reviewed and re-worked -- https://review.openstack.org/#/c/488874/ | 11:03 |
| patchbot | patch 488874 - virtualbmc - multiprocess server, socket-based management cli tool | 11:03 |
| *** zhangfei has quit IRC | 11:22 | |
| *** skazi has quit IRC | 11:29 | |
| sambetts | johnthetubaguy: Hey, are you around? I have some questions relating to nova to neutron interaction | 11:32 |
| *** nmathew has quit IRC | 11:38 | |
| *** zhangfei has joined #openstack-ironic | 11:38 | |
| *** links has quit IRC | 11:41 | |
| *** skazi has joined #openstack-ironic | 11:42 | |
| openstackgerrit | Sam Betts proposed openstack/ironic stable/ocata: Use zuul.override_checkout instead of custom branch_override var https://review.openstack.org/540830 | 11:45 |
| Tengu | sambetts: dtantsur etingof oh, if there's alreardy an existing patch, I'll check it and see if I can contrib on it, if you're ok? | 11:46 |
| Tengu | etingof: indeed your patch is (far) more advanced than the quick'n'dirty thing I've made :). | 11:47 |
| etingof | Tengu, I welcome your contributions indeed! ;) | 11:47 |
| Tengu | etingof: I'll check that out shortly (I hope) :). | 11:48 |
| etingof | Tengu, if you follow the review, a couple of significant changes were suggested ther | 11:48 |
| etingof | Tengu, so you may consider working on those if you have time and guts ;) | 11:48 |
| Tengu | yup, saw that in hurry. I'll check deeper later :). | 11:49 |
| etingof | Tengu, on the other hand, the patch I linked readily works with systemd so you may just use it as-is | 11:49 |
| Tengu | etingof: good to know - but if I can help to get it merged in some way, that will help a lot more of people than just me ;) | 11:50 |
| Tengu | will check what dtantsur requests in the latest review. Not sure I can help on that, my knowledge of openstack code is fresh. | 11:51 |
| etingof | Tengu, absolutely! for it to get merged we need to 1) address the suggestions in the comments and 2) lure core reviewers to actually review the patch ;-) | 11:52 |
| Tengu | ah, will need to learn a bit about cliff - and that will probably help me for certodia :). | 11:53 |
| Tengu | etingof: :] second part is the best I think :) | 11:53 |
| *** vsaienk0 has quit IRC | 11:53 | |
| Tengu | but right now : some rest, then eating, and with a full stomac code will be easier. | 11:53 |
| *** links has joined #openstack-ironic | 11:54 | |
| *** vsaienk0 has joined #openstack-ironic | 11:56 | |
| etingof | someone is risking a nap afterward ;) | 12:00 |
| openstackgerrit | liumengke proposed openstack/ironic master: Fixed some typos in test code. https://review.openstack.org/541251 | 12:03 |
| *** fyxim has quit IRC | 12:11 | |
| *** fyxim has joined #openstack-ironic | 12:11 | |
| *** olivierbourdon38 has quit IRC | 12:11 | |
| Tengu | etingof: :] | 12:14 |
| *** absubram has joined #openstack-ironic | 12:25 | |
| *** bfournie has quit IRC | 12:28 | |
| *** bfournie has joined #openstack-ironic | 12:28 | |
| mgoddard_ | sambetts: hey, I think johnthetubaguy is out today | 12:33 |
| *** bfournie has quit IRC | 12:33 | |
| sambetts | mgoddard_: ah no problem, need his in depth knowledge of nova's internals to answer a question I had | 12:35 |
| mgoddard_ | sambetts: what is the question? | 12:37 |
| sambetts | mgoddard_: I wanted to know if and how (if it does) nova resets the mac address on the neutron ports on instance delete if nova has overwriten then during deployment (in cases like SRIOV etc the hardware mac address gets set overriding the neutron generated one) | 12:39 |
| sambetts | because if it doesn't then you can end up with ports stuck with the hardware mac address until they are deleted | 12:41 |
| mgoddard_ | sambetts: good question. IIRC, being able to unset the MAC of a port is a relatively recent thing. Possibly predating SR-IOV support? So maybe they just don't | 12:45 |
| mgoddard_ | If it's done, it will be in nova/network/neutronv2/api.py. Nothing's leaping out at me. I can see where the MAC is set for SR-IOV | 12:46 |
| sambetts | mgoddard_: yeah... I'm trying to bring a fix to Ironic relating to the same thing, basically we overwrite the mac address on the port, but then we leave the port set that way which if the port is deleted breaks deployment the second time around due to "mac already in use" errors | 12:47 |
| sambetts | I was going to set the mac_addres to None when we unbind the port, and I expected neutron would regenerate the mac address, however it seems like you can't unset the mac address like that, so I'm left with having to store the old one | 12:48 |
| sambetts | and restore it | 12:48 |
| sambetts | but then that concerns me because what if neutron generates a mac matching the mac we've got stored for a different port | 12:48 |
| sambetts | then we'll end up in a real mess | 12:49 |
| mgoddard_ | maybe just generate another one? | 12:49 |
| sambetts | yeah thats the other option... but I was hoping to avoid it | 12:50 |
| *** jaianshu has quit IRC | 12:50 | |
| sambetts | which is why I wondered if nova had already solved this issue somehow | 12:50 |
| *** yangyapeng has quit IRC | 12:52 | |
| *** yangyapeng has joined #openstack-ironic | 12:53 | |
| *** yangyapeng has quit IRC | 12:53 | |
| *** yangyapeng has joined #openstack-ironic | 12:54 | |
| *** links has quit IRC | 12:55 | |
| peitolm | are the generated MAC addresses based on the underlying hardware's original address (ie assigned by the manufacturer? | 12:55 |
| *** jlabarre has joined #openstack-ironic | 12:55 | |
| *** takamatsu has joined #openstack-ironic | 12:56 | |
| sambetts | in neutron the macs are generated randomly but using a configurable prefix I believe | 12:58 |
| mgoddard_ | tricky one. I guess if we can't prevent neutron from reallocating the MAC, we're always going to need a generate code path. In that case, perhaps we should always generate? | 12:58 |
| sambetts | always generate a new one on port unbind? | 12:59 |
| mgoddard_ | could use a similar scheme, allowing the ironic & neutron prefixes to be different? | 12:59 |
| mgoddard_ | yes | 12:59 |
| *** AlexeyAbashkin has quit IRC | 12:59 | |
| sambetts | yeah... slightly concerned that that port doesn't have a neutron generated mac any more though :/ I think at the PTG I might asked the neutron folks if they think an ability to set mac to None to ask it to regenerate makes sense as a feature | 13:00 |
| peitolm | if the instances are going on bare-metal, why is the MAC being changed? | 13:00 |
| *** zhurong_ has joined #openstack-ironic | 13:01 | |
| *** AlexeyAbashkin has joined #openstack-ironic | 13:01 | |
| sambetts | peitolm: neutron generates a mac address for a port when it is created, when we deploy onto baremetal we can't change the mac address of the physical baremetal NIC, so we update neutron to match to NIC | 13:01 |
| sambetts | peitolm: when the instance is deleted, the port might not be deleted, so we need to undo the mac address update | 13:02 |
| *** yangyapeng has quit IRC | 13:02 | |
| sambetts | so that it can be used for another instance and won't conflict if someone else redeploys the same baremetal server | 13:02 |
| peitolm | ok. so if there is a prefix used as you suggest, (which would/should be different from any hardware vendor), then storing and returning it back to the stored one won't clash | 13:04 |
| peitolm | the stored one being the original hardware one i'm presuming | 13:04 |
| sambetts | I was suggesting storing the one that neutron generated, so that we can restore it, but if neutron has generated it, it can easily generate it again | 13:05 |
| peitolm | ah, | 13:05 |
| peitolm | so nova is going to override a generated one that neutron has set to override the hardware one? | 13:06 |
| sambetts | for nova VMs using SRIOV, nova overrides the generated one with the hardware one, and in Ironic we override the generated one with the baremetal server's one | 13:07 |
| *** links has joined #openstack-ironic | 13:08 | |
| peitolm | *nods* does the mac already in use error come from neutron? | 13:09 |
| sambetts | yup, when you try either create a port or update a port with a mac that is already in a network | 13:09 |
| *** jaganathan has quit IRC | 13:10 | |
| *** bfournie has joined #openstack-ironic | 13:10 | |
| peitolm | maybe neutron should be handling this centrally, ie, checking if the mac being set is one that's is already set on that host, (ie is just a null operation) | 13:11 |
| peitolm | rather than, is this mac set anywhere, | 13:11 |
| peitolm | by anywhere, the code suggests it's checking just the network, | 13:15 |
| peitolm | i'm probably missing something though | 13:16 |
| sambetts | all that logic is correct on the neutron side, as far as I know, its just that we have a set a port to have a mac address that can only ever match to single baremetal host, and when we're done with it we need to undo that so that the port not longer has any association to the baremetal it was attached too | 13:18 |
| *** mrunge_ has joined #openstack-ironic | 13:22 | |
| *** jaganathan has joined #openstack-ironic | 13:23 | |
| *** mrunge has quit IRC | 13:24 | |
| peitolm | you have a configuration where the port has a mac addres, that port is only ever matching a single host, which is valid, correct? | 13:25 |
| *** AlexeyAbashkin has quit IRC | 13:25 | |
| *** AlexeyAbashkin has joined #openstack-ironic | 13:27 | |
| openstackgerrit | Bob Fournier proposed openstack/ironic-inspector stable/ocata: Remove use of construct lib FieldError exception https://review.openstack.org/541271 | 13:29 |
| *** yangyapeng has joined #openstack-ironic | 13:30 | |
| openstackgerrit | Bob Fournier proposed openstack/ironic-inspector stable/ocata: Remove use of construct lib FieldError exception https://review.openstack.org/541271 | 13:33 |
| openstackgerrit | Ruby Loo proposed openstack/ironic master: Fix handling of 'timeout' parameter to power methods https://review.openstack.org/540150 | 13:33 |
| *** rloo has joined #openstack-ironic | 13:34 | |
| *** yangyapeng has quit IRC | 13:38 | |
| *** AlexeyAbashkin has quit IRC | 13:38 | |
| sambetts | peitolm: a user created a port in neutron without specifying a mac address, so neutron generates a mac address for them. then they go to nova and boot an baremetal instance passing in the uuid of the port they created. the nova scheduler selects a baremetal server to deploy their instance on to, and the information including that port uuid gets passed down through nova to ironic. ironic then | 13:38 |
| sambetts | updates that port with the hardware mac address from the server's NIC so that the information in neutron matches the real server's mac address, so that neutron's DHCP server etc is program with the right mac address. The user uses his instance to do some stuff. The user then deletes his instance via nova. Ironic then powers off the server, and cleans up the user's data. the instances is now | 13:38 |
| sambetts | removed from nova. However the port in neutron still exists, and the mac address is now set to the mac address of the baremetal server. So now another user does the same set of steps, and ends up on the same baremetal server, not knowing it was used before. when ironic tries to update the port (a different port owned by a different user but assigned to the same baremetal) we get a mac address | 13:38 |
| sambetts | already in use error, because of the port left behind by the first user | 13:38 |
| peitolm | I see | 13:39 |
| peitolm | thank you, sorry you had to type all of that | 13:39 |
| sambetts | no problem, its easier to explain with the whole picture | 13:39 |
| peitolm | i was missing the first bit, which is probably key | 13:40 |
| peitolm | i'm guessing the instance, may or may not run on ironic | 13:40 |
| peitolm | i wonder if neutron should do something like, "I've been informed that the mac address for this instance has been overriden by the provider, therefore i'll note that fact, along with the original address, so that when the instance goes, i can reset it" | 13:41 |
| peitolm | but I guess you're trying to avoid modifying neutron? | 13:42 |
| sambetts | yeah right now trying to fix that bug in ironic, without touching nova or neutron, there are two solutions I see if we make changes on the neutron side that could work, either like you suggest neutron keeps track and resets it, or neutron lets a user set the mac address to none which means generate me a new one please | 13:44 |
| sambetts | but unfortunatly I don't think a change like that will merge soon, so trying to work out how to resolve the bug without those | 13:44 |
| *** yangyapeng has joined #openstack-ironic | 13:44 | |
| sambetts | and so either we can track the old mac in ironic, and restore it, but because neutron then doesn't think the mac is assigned to a port, it might regenerate the same one and assign it to antoher port resulting the in the same conflict | 13:45 |
| sambetts | or we generate a random mac address and change it to that | 13:46 |
| peitolm | i thnk the first one is going to be a problem, depends on the login for generating a mac address | 13:46 |
| peitolm | is the generation of a mac address a function you can call from ironic? | 13:47 |
| sambetts | we'd have to generate it in ironic's code, and then call neutron.port_update to set it | 13:47 |
| peitolm | or you could create a temporary port, grab the mac from that, delete it anad assign that mac to the other port | 13:48 |
| peitolm | but that's still got a race condition | 13:48 |
| sambetts | yeah, another option I've thought of is deleting the original port, and then reassigning everything including the uuid, to a new one | 13:49 |
| peitolm | fraglie though | 13:49 |
| *** yangyapeng has quit IRC | 13:50 | |
| sambetts | yeah :/ and I'm not sure all the fields are writable | 13:50 |
| peitolm | have you asked teh neutron team about it? | 13:50 |
| sambetts | not yet, I was waiting for more of them to wake up | 13:50 |
| peitolm | storing it and restoreing it sounds like a workaround that'll be forgotten about | 13:51 |
| sambetts | yeah, and I think thats too fragile too | 13:51 |
| *** zhurong_ has quit IRC | 13:52 | |
| sambetts | I think the neatest solution would be to generate new ones until we get one that doesn't conflict with an existing one | 13:52 |
| sambetts | which should be only one or two | 13:52 |
| peitolm | well, the neatest solution from your side, would be to not have to deal with it :P | 13:52 |
| *** jcoufal has joined #openstack-ironic | 13:52 | |
| peitolm | but the neatest solution overwall is to get neutron to handle it | 13:52 |
| sambetts | yeah, I'm definatly going to talk to the neutron team and work out a better solution | 13:53 |
| *** rbudden has joined #openstack-ironic | 13:53 | |
| peitolm | I wonder if a more generic option may be to have a flag for static vs dynamic mac | 13:53 |
| sambetts | even if its just the ability to set mac_addres:"auto" in neutron | 13:53 |
| peitolm | ie, if the user sets the mac, then keep it, and if they move the instance to another host, that's their issu | 13:53 |
| peitolm | if it's dynamic, then at deploy time, it's re-generated | 13:53 |
| peitolm | auto is better than none | 13:54 |
| sambetts | yeah, thats similar to whats happening now with defered ip addresses in neutron | 13:54 |
| peitolm | what happens if you set it back to blank? is that possible? | 13:54 |
| *** Sukhdev_ has joined #openstack-ironic | 13:55 | |
| sambetts | that what I mean by setting it to None (null in json), but neutron rejects that request today | 13:55 |
| sambetts | I'd like it if setting it to blank/none/null resulted in a regen | 13:55 |
| peitolm | ah, i did wonder if you meant a null, or a literal none | 13:55 |
| peitolm | wonder what would happen if you set it to an invalid mac | 13:56 |
| sambetts | they have pretty good validation, and it rejects it with a invalid mac address error | 13:56 |
| *** VisBits has joined #openstack-ironic | 14:06 | |
| TheJulia | good morning everyone | 14:06 |
| dtantsur | morning TheJulia | 14:07 |
| peitolm | morning TheJulia | 14:07 |
| sambetts | o/ TheJulia | 14:10 |
| TheJulia | Anyone had any luck with grenade? | 14:10 |
| *** jrist has quit IRC | 14:11 | |
| *** yangyapeng has joined #openstack-ironic | 14:11 | |
| peitolm | sambetts: i found references to net.get_random_mac in the neutron code, but can' seem to pin down its source, looks like there was a discussion about it back in july 2016 | 14:12 |
| *** jrist has joined #openstack-ironic | 14:12 | |
| dtantsur | TheJulia: not me. got buried into one "funny" bug | 14:12 |
| TheJulia | is it actually funny? :) | 14:13 |
| peitolm | sambetts: ah, it's in neutron-lib | 14:13 |
| dtantsur | well.. https://launchpad.net/bugs/1643995 | 14:13 |
| openstack | Launchpad bug 1643995 in Ironic "ironic can't find nodes whose name ends in ".<mimetype>"" [High,Triaged] - Assigned to Dmitry Tantsur (divius) | 14:13 |
| TheJulia | I have a theory, I'm going to get some coffee, and issue a revert on neutron and test patch on our gate | 14:14 |
| TheJulia | dtantsur: ugh | 14:14 |
| peitolm | wow, that's a good definition of unintended consequences | 14:15 |
| *** Goneri has joined #openstack-ironic | 14:16 | |
| TheJulia | peitolm: which statement? | 14:17 |
| *** jaganathan has quit IRC | 14:17 | |
| peitolm | TheJulia: the lanchpad bug | 14:18 |
| dtantsur | yep | 14:18 |
| sambetts | what even is a troff-me content type... /me googles | 14:19 |
| dtantsur | .1 is the funniest | 14:20 |
| peitolm | dtantsur: as a random test, trying test.me fails, but test.me.json works | 14:20 |
| dtantsur | people really tried that | 14:20 |
| dtantsur | yeah | 14:20 |
| *** cdearborn has joined #openstack-ironic | 14:20 | |
| peitolm | dtantsur: who's the yeah at? me or sambetts | 14:21 |
| dtantsur | at both :) | 14:21 |
| peitolm | oh, did they report that in the redhat bug? | 14:22 |
| peitolm | i'm surprised it actually stripping off the supposed mime type though | 14:23 |
| * TheJulia tries to wake up | 14:23 | |
| dtantsur | peitolm: this is by design. a feature of the pecan library we use | 14:24 |
| peitolm | so "openstack baremetal node create --driver ipmi --name test.json" followed by "openstack baremetal node show test.json" fails, but if you do, "openstack baremetal node show test.json.json" it works | 14:24 |
| *** mjturek has joined #openstack-ironic | 14:26 | |
| *** cdearborn_ has joined #openstack-ironic | 14:26 | |
| dtantsur | right, because it only strips the prefix once | 14:26 |
| *** cdearborn_ has quit IRC | 14:26 | |
| peitolm | I wonder if you could sanitity check at node create/update time | 14:27 |
| peitolm | ie if the name matches a mimetype, issue a warning | 14:27 |
| dtantsur | well, I don't want to prevent users from creating such nodes, I'd rather make them work | 14:27 |
| dtantsur | users don't see warnings, and we don't have any kind of a warning API.. | 14:28 |
| peitolm | true | 14:29 |
| *** VisBits has quit IRC | 14:29 | |
| *** jaganathan has joined #openstack-ironic | 14:31 | |
| *** dpawar has quit IRC | 14:33 | |
| *** lucasagomes is now known as lucas-hungry | 14:34 | |
| *** AlexeyAbashkin has joined #openstack-ironic | 14:35 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: [WIP] Stop guessing mime types based on URLs https://review.openstack.org/541289 | 14:36 |
| dtantsur | peitolm, TheJulia, sambetts, what do you think about this approach? ^^^ | 14:36 |
| *** dtantsur is now known as dtantsur|bbl | 14:37 | |
| *** links has quit IRC | 14:38 | |
| *** mjturek has quit IRC | 14:39 | |
| *** zhangfei has quit IRC | 14:39 | |
| sambetts | dtantsur|bbl: seems sane enough to me | 14:39 |
| *** cdearborn has quit IRC | 14:41 | |
| TheJulia | sane, I'm not really a fan of the json fallback stuff, but I'm not entirely sure it si there... I'm also not fully caffinated yet | 14:42 |
| peitolm | the tests, test_get_one_with_json_in_name , that's with .json at the end right? rather than something.json.foo | 14:43 |
| peitolm | and test_get_one_with_double_json, shouldn't that be name=node.json.json | 14:43 |
| peitolm | ignore that last one, i see what you're testing | 14:44 |
| *** zhangfei has joined #openstack-ironic | 14:53 | |
| *** davidlenwell has quit IRC | 14:55 | |
| rloo | good morning everyone, peitolm, TheJulia, sambetts, dtantsur|bbl | 14:56 |
| *** davidlenwell has joined #openstack-ironic | 14:56 | |
| rloo | TheJulia: I haven't started to look at the grenade issue but it sounds like you've got a handle on it? if it is network related, I'm somewhat useless. | 14:56 |
| TheJulia | rloo: more eyes the better.... | 14:57 |
| rloo | TheJulia: ok, i'll look today. just going through priorities now to see if there is anything else that needs attention | 14:57 |
| rloo | sambetts: wrt ironic-tempest-plugin failure. is there more work to do, cuz this fails: https://review.openstack.org/#/c/540355/ | 14:59 |
| patchbot | patch 540355 - ironic-tempest-plugin - Switch to override-checkout for stable branch jobs | 14:59 |
| sambetts | rloo: its currently rechecking because of something I missed on the ocata branch backport | 14:59 |
| sambetts | rloo: the latest run (which is still running) should passed on all branches | 15:00 |
| rloo | sambetts: ah, ok. is the order wrt merging, master, pike, ocata starting with https://review.openstack.org/#/c/540432/ ? | 15:00 |
| patchbot | patch 540432 - ironic - Use zuul.override_checkout instead of custom bran... | 15:00 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: DNM TEST Lets try reverting the suspicious neutron change https://review.openstack.org/541298 | 15:00 |
| sambetts | rloo: yup, we need those three, and then the tempest one | 15:01 |
| rloo | sambetts: ok, i'll update the whiteboard with them so i know the order to review. | 15:01 |
| sambetts | rloo: thanks :) | 15:01 |
| rloo | sambetts: but... this fails, and it is the first one: https://review.openstack.org/#/c/540432/ | 15:02 |
| patchbot | patch 540432 - ironic - Use zuul.override_checkout instead of custom bran... | 15:02 |
| sambetts | rloo: from what I can tell the failure isn't related to the change I think its related to the thing TheJulia is investigating | 15:02 |
| rloo | sambetts: ugh. i thought it was only multinode grenade that was failing. | 15:03 |
| TheJulia | sambetts: it is, I spotted it on two separate jobs where somehow the connection to the VIP never mapped through... | 15:03 |
| TheJulia | well, the failure on grenade master to master was actually purely a timeout | 15:04 |
| TheJulia | the multinode one is the same multinode issue | 15:04 |
| *** zhangfei has quit IRC | 15:04 | |
| sambetts | ah so a recheck should fix the normal grenade one | 15:05 |
| sambetts | and multi is just sad | 15:05 |
| TheJulia | indeed | 15:05 |
| rloo | and we are sad too. thx TheJulia. | 15:05 |
| *** dsneddon has joined #openstack-ironic | 15:06 | |
| *** zshi has quit IRC | 15:07 | |
| TheJulia | I've not found any way to get additional debugging short of local recreation and interception or hacking on neutron code | 15:09 |
| *** dsneddon has quit IRC | 15:09 | |
| *** dsneddon has joined #openstack-ironic | 15:09 | |
| *** dtantsur|bbl is now known as dtantsur | 15:14 | |
| dtantsur | morning rloo | 15:14 |
| *** mgkwill has quit IRC | 15:14 | |
| rloo | dtantsur: :) | 15:14 |
| *** mgkwill has joined #openstack-ironic | 15:15 | |
| sambetts | TheJulia: hmmm, whats the theory around that patch breaking our grenade? | 15:15 |
| TheJulia | silent behavior in ovs firewall since we do actually have real ports that it can see/interact with | 15:16 |
| sambetts | oh right... because of our hacks to plug in the bms | 15:17 |
| TheJulia | bingo | 15:17 |
| *** rpioso|afk is now known as rpioso | 15:17 | |
| sambetts | le sigh.. I knew this stuff would bite us on the ass | 15:17 |
| rpioso | Good morning | 15:17 |
| *** dsneddon has quit IRC | 15:18 | |
| sambetts | TheJulia: hjensas has been working to restore my patches which make our BM connects to OVS less hacky if thats any help | 15:18 |
| dtantsur | morning rpioso | 15:18 |
| TheJulia | looks like the firewalling defaults to iptables_hybrid.... I'm wondering if we could just.. | 15:18 |
| TheJulia | oh shoot, the failure is on the pike code base, not master | 15:18 |
| TheJulia | Yeah, and that neutron change is only on master | 15:19 |
| rpioso | Hey dtantsur :) | 15:20 |
| *** zshi has joined #openstack-ironic | 15:20 | |
| TheJulia | so anyway, my revert attempt likely won't do anything | 15:21 |
| *** trown is now known as trown|outtypewww | 15:23 | |
| *** dsneddon has joined #openstack-ironic | 15:26 | |
| *** lucas-hungry is now known as lucasagomes | 15:30 | |
| dtantsur | folks, do I get it right that only nodes and port groups have user-settable names? | 15:34 |
| rloo | dtantsur: yeah, that sounds about right. | 15:38 |
| rloo | TheJulia: :-( do we need to fix on pike? | 15:38 |
| TheJulia | possibly, I'm hoping reproduce.sh yeilds something I can look at while broken | 15:38 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Stop guessing mime types based on URLs https://review.openstack.org/541289 | 15:46 |
| dtantsur | okay, please consider this bug fix for queens ^^^ | 15:46 |
| dtantsur | huh, I think I need to hack also DELETE operation.. will look later | 15:48 |
| *** dtantsur is now known as dtantsur|bbl | 15:49 | |
| NobodyCam | Good Morning Ironic'ers | 15:50 |
| *** pcaruana has quit IRC | 15:51 | |
| *** samueldmq has quit IRC | 15:52 | |
| rloo | dtantsur|bbl: please add link to whiteboard/priorities ^^ | 15:52 |
| rloo | morning NobodyCam! | 15:52 |
| *** samueldmq has joined #openstack-ironic | 15:52 | |
| NobodyCam | Good Morning rloo :) | 15:52 |
| openstackgerrit | Doug Hellmann proposed openstack/ironic-ui master: Update reno for stable/queens https://review.openstack.org/541319 | 16:00 |
| *** r-daneel has joined #openstack-ironic | 16:01 | |
| *** Sukhdev_ has quit IRC | 16:04 | |
| *** mjura has quit IRC | 16:08 | |
| *** rloo_ has joined #openstack-ironic | 16:10 | |
| *** rloo has quit IRC | 16:13 | |
| jlvillal | Good morning :) | 16:20 |
| jlvillal | That Zuul check queue is very very long... | 16:20 |
| *** retr0h has quit IRC | 16:20 | |
| *** retr0h has joined #openstack-ironic | 16:20 | |
| rpioso | Good morning NobodyCam and jlvillal | 16:21 |
| jlvillal | :) | 16:21 |
| *** e0ne has quit IRC | 16:22 | |
| NobodyCam | Good Morning rpioso :) | 16:25 |
| *** amarao has quit IRC | 16:27 | |
| *** jaganathan has quit IRC | 16:30 | |
| *** rbartal has quit IRC | 16:40 | |
| *** priteau has quit IRC | 16:42 | |
| *** hamzy has quit IRC | 16:43 | |
| jroll | good morning everyone | 16:55 |
| sambetts | hey jroll | 16:56 |
| *** mjturek has joined #openstack-ironic | 16:57 | |
| *** SerenaFeng has joined #openstack-ironic | 17:02 | |
| TheJulia | so.. I think I've reproduced it | 17:06 |
| TheJulia | sambetts: https://paste.openstack.org/show/663642 | 17:07 |
| sambetts | TheJulia: oh nice! | 17:08 |
| sambetts | TheJulia: hmmm ... I thought we fixed the issue with us removing the tap interfaces... | 17:08 |
| sambetts | there was a issue at one point with devstack where libvirt would delete the tap interfaces for the vms | 17:09 |
| sambetts | but I thought we fixed it, maybe we didn't back port it | 17:09 |
| TheJulia | so, we have the tap interfaces, only one vm is presently running, was 2 | 17:11 |
| TheJulia | node-3 has a tag... but I don't see that hooked anywhere else (of course, it is rare for me to have to dig into this stuff) | 17:11 |
| *** nicodemos has quit IRC | 17:11 | |
| sambetts | TheJulia: looks like we didn't backport the fix for the issue with the underlying interfaces getting removed which caused us a mojor headache before | 17:12 |
| sambetts | TheJulia: this --> https://review.openstack.org/#/c/510465/ | 17:12 |
| patchbot | patch 510465 - ironic - [Devstack] Replace tap with veth (MERGED) | 17:12 |
| *** ricardoas has quit IRC | 17:12 | |
| sambetts | it probably needs to be backported | 17:12 |
| sambetts | I think that patch was related to multinode failures (non-grenade) | 17:13 |
| sambetts | so add grenade in running the old version without the fix and *drum rollll* same error | 17:14 |
| * jroll can't connect to paste.o.o, wtf | 17:15 | |
| sambetts | jroll: I coukdn't either until I removed the https:// | 17:15 |
| sambetts | not sure why | 17:15 |
| jroll | ahhhhhhhhhhhhh | 17:15 |
| jroll | \o/ thanks | 17:15 |
| openstackgerrit | Sam Betts proposed openstack/ironic stable/pike: [Devstack] Replace tap with veth https://review.openstack.org/541378 | 17:16 |
| TheJulia | ahh, sam beat me to it | 17:16 |
| sambetts | :-P need a depends on for that ^ on your patch and we'll see if that resolves it | 17:16 |
| TheJulia | oh... I had to hand type the url because one of my computers is acting wonky | 17:16 |
| TheJulia | yup | 17:17 |
| sambetts | ahhh, no problem | 17:17 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: DNM TEST Lets the backport Depends-On: I04ff6a97bebf15636a794b51f748c62eda627d36 Change-Id: I831189540b9730bf4cf4e5de7e0a406f83283475 https://review.openstack.org/541298 | 17:18 |
| TheJulia | so now we wait | 17:19 |
| sambetts | \o/ waiting | 17:19 |
| *** rwsu has quit IRC | 17:19 | |
| sambetts | I've been waiting all day for the results of my recheck on https://review.openstack.org/#/c/540355/ | 17:20 |
| patchbot | patch 540355 - ironic-tempest-plugin - Switch to override-checkout for stable branch jobs | 17:20 |
| sambetts | :( | 17:20 |
| *** rwsu has joined #openstack-ironic | 17:20 | |
| TheJulia | I just noticed there were rechecks from YESTERDAY still running | 17:20 |
| sambetts | oh dear... | 17:20 |
| *** SerenaFeng has quit IRC | 17:20 | |
| TheJulia | yeah... | 17:20 |
| *** marios has quit IRC | 17:21 | |
| *** r-daneel has quit IRC | 17:22 | |
| *** r-daneel has joined #openstack-ironic | 17:22 | |
| *** dpawar has joined #openstack-ironic | 17:23 | |
| *** dpawar has quit IRC | 17:24 | |
| *** AlexeyAbashkin has quit IRC | 17:25 | |
| *** AlexeyAbashkin has joined #openstack-ironic | 17:25 | |
| *** AlexeyAbashkin has quit IRC | 17:29 | |
| *** vsaienk0 has quit IRC | 17:33 | |
| *** jlabarre has quit IRC | 17:34 | |
| TheJulia | I was going to ask in infra, but it looks like they are still fighting issues | 17:35 |
| TheJulia | sambetts: you may want to re-review https://review.openstack.org/#/c/534441/ | 17:36 |
| patchbot | patch 534441 - ironic - Don't try to lock for vif detach | 17:36 |
| openstackgerrit | Sam Betts proposed openstack/ironic-tempest-plugin master: Add [baremetal] available_nodes configuration option https://review.openstack.org/541383 | 17:37 |
| *** jlabarre has joined #openstack-ironic | 17:37 | |
| sambetts | TheJulia: I'll add it to the list :) | 17:37 |
| *** mtreinish has quit IRC | 17:43 | |
| peitolm | i'm trying to double check my install, and my "ovs-vsctl show" doesn't match step 6. of https://docs.openstack.org/ironic/pike/install/configure-networking.html, it's missing the options sections, http://paste.openstack.org/show/663703/ | 17:46 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: DNM TEST Lets the backport https://review.openstack.org/541298 | 17:46 |
| peitolm | any chance someone could check the output of ovs-vsctl on their system, and see if you're seeing the same, I'm not sure if it's a mis-configuration at my end, or if it's a change due to the ovs_version | 17:47 |
| sambetts | peitolm: if you have the neutron openvswitch agent running it'll create the phy-br- and int-br devices for you | 17:47 |
| *** dpawar has joined #openstack-ironic | 17:47 | |
| sambetts | if its configured correctly | 17:48 |
| peitolm | ok. i'll take a look at that then | 17:48 |
| peitolm | hmm, the br-int and br-eth2 are showing as down as well... | 17:49 |
| peitolm | hmm, " Agent out of sync with plugin!" | 17:50 |
| *** mtreinish has joined #openstack-ironic | 17:50 | |
| TheJulia | with the current status of the gate, I suspect we are not going to be able to land much in the way of bug fixes | 17:50 |
| *** rloo_ is now known as rloo | 17:56 | |
| *** crushil has joined #openstack-ironic | 17:58 | |
| *** crushil_ has joined #openstack-ironic | 17:58 | |
| * sambetts ---> home | 17:58 | |
| peitolm | hmmm | 18:00 |
| *** tesseract has quit IRC | 18:00 | |
| *** crushil has quit IRC | 18:00 | |
| sambetts | night all o/ | 18:01 |
| *** sambetts is now known as sambetts|afk | 18:01 | |
| *** derekh has quit IRC | 18:02 | |
| *** absubram has quit IRC | 18:03 | |
| TheJulia | goodnight sam | 18:06 |
| peitolm | 'night sam | 18:09 |
| *** hamzy has joined #openstack-ironic | 18:13 | |
| *** devananda has joined #openstack-ironic | 18:13 | |
| *** hamzy_ has joined #openstack-ironic | 18:15 | |
| snasir | Hi can a baremetal instance be suspended and or paused while in "active" state via Nova ? Just like we can suspend or pause VM instances via Nova ? Thanks | 18:15 |
| *** hamzy has quit IRC | 18:17 | |
| rloo | snasir: sorry, it can't be suspended or paused | 18:20 |
| rloo | bye sambetts|afk | 18:20 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Only set default network interface flat if enabled in config https://review.openstack.org/535830 | 18:21 |
| *** dtantsur|bbl is now known as dtantsur | 18:22 | |
| snasir | rloo: If that is the case, then how come we have those options available ? Does nova indeed try and suspend the instance, but then the operation is blocked by ironic-conductor ? | 18:24 |
| snasir | rloo: From horizon, I can see options such as "Migrate, suspend & Pause instances" even for the baremetal instance | 18:25 |
| TheJulia | snasir: nova is built around those concepts being supported, they simply don't have a matching in the ironic virt driver in nova-compute | 18:25 |
| rloo | snasir: oh, interesting. i suspect from a high level they don't know whether a vm is baremetal or not | 18:25 |
| TheJulia | snasir: also, horizon has 0 clue that it is baremetal under the hood | 18:25 |
| rloo | snasir: although i agree from the user's point of view, it isn't good. | 18:26 |
| snasir | TheJulia: Thanks for clarifying | 18:26 |
| rloo | snasir: the nova code raises an exception if you try it, not sure what trickles up to horizon | 18:27 |
| TheJulia | can anyone load the zuul status page? | 18:29 |
| jroll | fwiw, this is what the "capabilities API" we keep talking about is meant to solve. "what actions can be performed on this resource?" | 18:29 |
| * jroll tries | 18:30 | |
| TheJulia | :) | 18:30 |
| * jroll continues waiting | 18:30 | |
| jroll | gonna guess no | 18:30 |
| dtantsur | jroll: do you plan to update https://review.openstack.org/#/c/540402/ or should I? | 18:31 |
| patchbot | patch 540402 - ironic-inspector - Remove sample policy and config files | 18:31 |
| dtantsur | I'd like it before the release | 18:31 |
| hjensas | TheJulia: looks like the zuul status page took some time off ... | 18:31 |
| jroll | dtantsur: sorry, got distracted, I can do it quickly | 18:32 |
| dtantsur | thanks! | 18:32 |
| TheJulia | infra seems to be looking into the zuul issue | 18:33 |
| * TheJulia sighs | 18:33 | |
| *** harlowja has joined #openstack-ironic | 18:35 | |
| snasir | TheJulia: In addition, is it possible to create a snapshot for the instance as well, or thats not supported for baremetal either ? | 18:36 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-inspector master: Remove sample policy and config files https://review.openstack.org/540402 | 18:36 |
| *** olivierbourdon38 has joined #openstack-ironic | 18:36 | |
| jroll | dtantsur: there you are, sorry for the delay | 18:36 |
| TheJulia | snasir: not supported since the node is a physical server. If your doing something like boot from volume from a cinder volume, you can snap the volume directly | 18:36 |
| dtantsur | jroll: thanks! will wait for the CI now | 18:37 |
| TheJulia | snasir: we've talked about implementing the ability to take an offline snapshot/image the machine, except such things are a little complicated to do | 18:37 |
| jroll | "a little" | 18:39 |
| snasir | TheJulia: Thanks | 18:39 |
| TheJulia | s/a little/very very/ | 18:40 |
| jroll | TheJulia: this looks ready to land: https://review.openstack.org/#/c/524709/ | 18:54 |
| patchbot | patch 524709 - networking-baremetal - Make the agent distributed using hashring and noti... | 18:54 |
| jroll | ditto the unit tests on top | 18:57 |
| TheJulia | jroll: will look after I eat some lunch | 18:59 |
| *** AlexeyAbashkin has joined #openstack-ironic | 19:00 | |
| *** lucasagomes is now known as lucas-afk | 19:02 | |
| jroll | thanks, no rush | 19:04 |
| *** r-daneel_ has joined #openstack-ironic | 19:08 | |
| *** r-daneel has quit IRC | 19:09 | |
| *** r-daneel_ is now known as r-daneel | 19:09 | |
| peitolm | rloo: is it something that could be supported, if, forexample, you can talk to the BMC to trigger ACPI events? | 19:10 |
| *** vsaienk0 has joined #openstack-ironic | 19:11 | |
| rloo | peitolm: context? support what? | 19:12 |
| TheJulia | peitolm: you mean like soft power? | 19:14 |
| peitolm | yeah | 19:14 |
| peitolm | i think there's an ACPI event that could trigger suspend or hibernate | 19:15 |
| dtantsur | highly OS-dependent, I suspect | 19:17 |
| peitolm | mm, | 19:18 |
| peitolm | it's also optional in the IPMI spec | 19:18 |
| rpioso | I have a silly Gerrit review question. A change I'm reviewing has a Files table, just like every change. For one of its entries, the Comments column contains, 'drafts: 1 new: 1'. What does the 'new: 1' mean? For that line, I made a comment and the owner replied. I've drafted a reply to his reply. What's strange is that another file in the table has two lines with the same review pattern -- comment, reply, and | 19:19 |
| rpioso | draft. But it doesn't have that 'new' business. | 19:19 |
| *** Goneri has quit IRC | 19:19 | |
| TheJulia | dtantsur: fairly yes | 19:21 |
| TheJulia | peitolm: we do have soft on/off support, I'm not sure if that can be sent via nova though | 19:21 |
| *** vsaienk0 has quit IRC | 19:22 | |
| TheJulia | rpioso: unread I think | 19:22 |
| rpioso | TheJulia: I've viewed the file in every patch set, and clicked on each comment and reply. | 19:28 |
| TheJulia | cached data in the browser maybe? | 19:29 |
| rpioso | TheJulia: Different brower that I rarely use, same result. | 19:32 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Stop guessing mime types based on URLs https://review.openstack.org/541289 | 19:37 |
| dtantsur | this should be ready ^^^ now a good question is whether it needs a microversion.. | 19:37 |
| dtantsur | pros: the fix is discoverable, cons: the fix is not backportable | 19:37 |
| dtantsur | opinions? | 19:38 |
| dtantsur | rloo, TheJulia, afternoon bikeshedding time ^^^ :) | 19:38 |
| rloo | dtantsur: dam* you... | 19:39 |
| dtantsur | LOL :D | 19:39 |
| * TheJulia frowns at the expectation of bikeshedding time | 19:39 | |
| dtantsur | it's never bad time for arguing about microversions, is it? | 19:39 |
| TheJulia | lol | 19:40 |
| TheJulia | I need to wrap up an email, give me a couple minutes | 19:40 |
| dtantsur | sure | 19:40 |
| dtantsur | me as an API-SIG cores says "microversion it", me as someone who still maintains Kilo says "OH NO" | 19:40 |
| rloo | dtantsur: is it a bug? err, was it a bug? I mean, do people intend to specify /v1/nodes.json and expect it to be treated with requested content type Application/json? | 19:41 |
| *** baoli has joined #openstack-ironic | 19:41 | |
| rloo | dtantsur: any idea what rest of openstack community does? | 19:41 |
| dtantsur | rloo: /v1/nodes.json is fine. the problem is in /v1/nodes/mynode.1 | 19:41 |
| openstackgerrit | Merged openstack/bifrost master: Fix typo in playbook markdown readme https://review.openstack.org/540988 | 19:41 |
| openstackgerrit | Merged openstack/bifrost master: Don't copy ironic's policy.json sample https://review.openstack.org/540893 | 19:41 |
| dtantsur | rloo: API-SIG recommends microversioning bugs. I have no idea what the community actually does | 19:41 |
| * dtantsur summons edleafe | 19:41 | |
| *** baoli has quit IRC | 19:42 | |
| rloo | dtantsur: API-SIG recommends, it isn't a 'must' or a 'shall'. let me think/look at the patch... | 19:42 |
| dtantsur | right, we cannot say 'must' :) | 19:43 |
| dtantsur | huh, nearly 9pm, I guess I'll go | 19:47 |
| TheJulia | gah! complex reply | 19:47 |
| *** mgoddard_ has quit IRC | 19:47 | |
| dtantsur | mmm? | 19:47 |
| TheJulia | dtantsur: In my mind, it is a bug and we really shouldn't try to microversion it | 19:48 |
| dtantsur | I quite agree, even though most of API-SIG things we should version bugs | 19:48 |
| rloo | so is /v1/node/<uuid>.json valid? | 19:48 |
| dtantsur | rloo: it used to be, and still is with my patch | 19:48 |
| dtantsur | pecan magic (and ponies) | 19:49 |
| TheJulia | our client doesn't go for latest possible maximum... so... we can't logistically version and actually fix the issue at the same time | 19:49 |
| dtantsur | yep | 19:49 |
| dtantsur | and if we version this fix, it's likely to miss queens | 19:49 |
| * TheJulia wonders if ponies could somehow be added into a slide deck for the project update | 19:49 | |
| dtantsur | rloo: pecan by default allows /foo/bar.ext and tries to detect mime type by "ext" | 19:49 |
| dtantsur | which is a so-so idea to me, but who knows | 19:50 |
| rloo | dtantsur: well, that is pecan. but what does a user expect? | 19:50 |
| dtantsur | so when we have /v1/nodes/small.1, it treats it as a node called "small" returned as a man page :) | 19:50 |
| TheJulia | oh wow, bifrost changes merged | 19:50 |
| rloo | dtantsur: c'mon, that's really cool and is exactly what the user would want! ugh. | 19:50 |
| dtantsur | rloo: I try to not change this aspect. what I want to change is inability to get, patch or delete nodes by name if the name ends with a known extension | 19:51 |
| dtantsur | so, I have to do some trickery around .json, because /v1/nodes/small.json may mean 1. node called "small.json", 2. node called "small" returned as JSON | 19:51 |
| dtantsur | the former is what we document, the latter is what actually works before my patch | 19:51 |
| dtantsur | I try to keep both | 19:51 |
| TheJulia | yeah, more and more I think about it... unversioned | 19:52 |
| dtantsur | then please review :) | 19:52 |
| rloo | did we ever talk about using something else instead of pecan? | 19:53 |
| dtantsur | we did | 19:53 |
| dtantsur | the problem is to keep compatibility with all these quirks | 19:53 |
| dtantsur | and of course that's quite a lot of work :) | 19:54 |
| rloo | not sure how easy it would be to move over to something else, but i don't think we want to version this and then have to continue supporting both ways | 19:54 |
| jroll | as soon as I don't have to spend a cycle talking about using something else, I'm going to do it | 19:54 |
| TheJulia | dtantsur: just clicked send so I can now :) | 19:55 |
| jroll | dtantsur: I'm inclined to think we don't have users relying on this behavior (other than perhaps .json, which you handle), so I vote no microversion for it | 19:55 |
| dtantsur | I think we had some very early agreement to try Flask (which is something ironic-inspector uses) | 19:55 |
| dtantsur | 3:0 for not versioning, nice! | 19:55 |
| * jroll will review the patch later, it's meetings/lunch time | 19:56 | |
| dtantsur | take your time, I'm leaving for a day anyway | 19:56 |
| dtantsur | completely unrelated, but may be curious: a guy on devconf.cz showed me https://github.com/LAKostis/redirector | 19:56 |
| dtantsur | (NSFL warning: a lot of hacks in bash) | 19:56 |
| rloo | it seems to me that other openstack projects use pecan, and will have a similar problem. | 19:57 |
| rloo | we should deal with it the same way. | 19:57 |
| rloo | maybe we are the first to notice... | 19:57 |
| TheJulia | not safe for ?l? | 19:58 |
| jroll | life | 19:58 |
| jroll | :P | 19:58 |
| dtantsur | yes | 19:59 |
| dtantsur | rloo: we may be the only project with 1. unique names, 2. allowing names to be used in /v1/type/<NAME>, IIRC nova does /v1/servers/?name=<NAME> (which btw does a LIKE query) | 20:00 |
| dtantsur | though flavors may use the first form, not sure | 20:00 |
| dtantsur | I did not get to implementing flavors in rust-openstack yet :D | 20:00 |
| *** jtomasek has quit IRC | 20:00 | |
| * dtantsur learns things by implementing them on client side | 20:00 | |
| *** absubram has joined #openstack-ironic | 20:01 | |
| *** dtantsur is now known as dtantsur|afk | 20:03 | |
| dtantsur|afk | o/ | 20:03 |
| rloo | bye dtantsur|afk | 20:06 |
| rloo | so sad, that bug was reported in 2016 :-( | 20:06 |
| * edleafe catches up | 20:16 | |
| edleafe | dtantsur|afk: rloo: if a bug fix changes a 500 to something else, that doesn't need a microversion | 20:16 |
| *** r-daneel_ has joined #openstack-ironic | 20:16 | |
| edleafe | But if you're changing what you get back from a given URL, well, that's a prime candidate for a new version | 20:17 |
| *** r-daneel has quit IRC | 20:17 | |
| *** r-daneel_ is now known as r-daneel | 20:17 | |
| rloo | edleafe: so if you did GET /v1/nodes/foo.me and you got 404 (cuz of bug), and with this fix, you get 200 + body in response, it needs to be versioned? | 20:19 |
| edleafe | rloo: if the 404 was legitimate, yes. If it should have returned something else, than no | 20:21 |
| edleafe | rloo: A lot depends on how it was documented | 20:21 |
| rloo | edleafe: the 404 wasn't legitimate | 20:21 |
| rloo | edleafe: not documented. i mean, the document was that GET /v1/nodes/foo.me would return info about a node named 'foo.me'. which it didn't do. | 20:22 |
| edleafe | rloo: if it wasn't doing what it should have, then no new version is needed. If it was doing what it should have, but with an incomplete/inadequate response, then it should be a new version | 20:23 |
| *** mjturek has quit IRC | 20:24 | |
| *** Goneri has joined #openstack-ironic | 20:24 | |
| edleafe | IOW, would anyone have anything that depended on the old behavior? | 20:24 |
| rloo | edleafe: thx, that's what i wanted to hear. It didn't do what it should have, so no versioning. | 20:24 |
| edleafe | rloo: cool | 20:24 |
| *** e0ne has joined #openstack-ironic | 20:24 | |
| rloo | edleafe: nope, nothing 'normal' would depend on the old behaviour :) | 20:24 |
| rloo | edleafe: 'normal' cuz one never knows... | 20:24 |
| *** mgoddard_ has joined #openstack-ironic | 20:25 | |
| rloo | thank you edleafe for clarifying! | 20:25 |
| *** rloo has quit IRC | 20:29 | |
| *** hw_wutianwei1 has joined #openstack-ironic | 20:32 | |
| *** hw_wutianwei has quit IRC | 20:34 | |
| *** hw_wutianwei1 is now known as hw_wutianwei | 20:34 | |
| *** mgoddard_ has quit IRC | 20:38 | |
| *** rloo has joined #openstack-ironic | 20:41 | |
| *** mjturek has joined #openstack-ironic | 20:42 | |
| *** mjturek has quit IRC | 20:45 | |
| snasir | Hi, I'm trying to enable the serial console of the baremetal node, using the instructions here https://docs.openstack.org/ironic/latest/admin/console.html#node-serial-console | 20:48 |
| snasir | And keep seeing this error https://thepasteb.in/p/lOhOKAgq71KHB , when i try to do this add this propert" --console-interface ipmitool-socat". And I am using pxe_ipmitool | 20:49 |
| snasir | The console_interface is already set to "None", and i want to change it to "ipmitool-socat", any suggestions on what I may be doing wrong ? Thanks | 20:50 |
| *** olivierbourdon38 has quit IRC | 20:50 | |
| TheJulia | snasir: you need to change to a hardware type | 20:55 |
| snasir | TheJulia: how do i go about doing that ? I was just checking the enabled drivers at the moment ? Is that another parameter that needs to be added to the ironic node ? | 20:56 |
| snasir | TheJulia: enabled_hardware_types = ipmi , is set in my ironic.conf at the moment | 20:57 |
| *** mjturek has joined #openstack-ironic | 20:57 | |
| TheJulia | snasir: set driver to ipmi, set deploy_interface to iscsi, then set console_interface | 20:58 |
| *** takamatsu has quit IRC | 21:01 | |
| *** e0ne has quit IRC | 21:01 | |
| snasir | TheJulia: I have "enabled_drivers = pxe_ipmitool" , And "#enabled_deploy_interfaces = iscsi,direct" in my ironic.conf.... I think iscsi is set by default if you dont specify it explicitly right ? | 21:01 |
| snasir | And i also have this set "enabled_console_interfaces = ipmitool-socat,no-console" in my ironic.conf | 21:01 |
| TheJulia | snasir: so hardware_types are superceeding drivers | 21:01 |
| TheJulia | ipmi is the generic replacement for pxe_ipmitool and agent_ipmitool, the only difference is the deploy interface. iscsi and direct respectively. | 21:02 |
| TheJulia | I _think_ iscsi would be the default as it is first in the list | 21:02 |
| *** mjturek has quit IRC | 21:03 | |
| jroll | 99% sure you're correct there | 21:13 |
| *** dpawar_ has joined #openstack-ironic | 21:17 | |
| *** e0ne has joined #openstack-ironic | 21:17 | |
| *** dpawar has quit IRC | 21:19 | |
| *** dpawar_ has quit IRC | 21:21 | |
| *** AlexeyAbashkin has quit IRC | 21:21 | |
| TheJulia | rloo: https://review.openstack.org/#/c/531120/ | 21:27 |
| patchbot | patch 531120 - ironic - Rework exception handling on deploy failures in co... | 21:27 |
| rloo | needs a revie? | 21:27 |
| rloo | w | 21:27 |
| rloo | ok, in a few min | 21:27 |
| TheJulia | thanks | 21:29 |
| TheJulia | grenade multinode also passed on it *blinks* | 21:29 |
| *** threestrands has joined #openstack-ironic | 21:32 | |
| *** threestrands has quit IRC | 21:32 | |
| *** threestrands has joined #openstack-ironic | 21:32 | |
| * TheJulia muses with the thought of dropping all of the nv jobs util after release... likely too late though | 21:33 | |
| rloo | TheJulia: i'd prefer to keep them, just in case something breaks, we can go back to see where it might have happened | 21:35 |
| rloo | TheJulia: what about just dropping patches that we don't care about? | 21:35 |
| rloo | TheJulia: the abandon button will do that | 21:35 |
| rloo | TheJulia: is zuul overwhelmed? | 21:36 |
| TheJulia | rloo: I think there are 400 patches in the check queue | 21:36 |
| TheJulia | I'd rather not abandon/unabandon... Just have to wait it out I think | 21:37 |
| rloo | TheJulia: whoa! I suspect that there'll be an extension to the RC1 deadline. And really, we are probably safe if we get ironic in 1-2 days after the rest of the projects. i can't recall when things 'broke' last time, maybe a week after rc1. | 21:38 |
| *** mjturek has joined #openstack-ironic | 21:39 | |
| TheJulia | yeah | 21:41 |
| *** ijw has joined #openstack-ironic | 21:42 | |
| peitolm | where's the best place to start looking at port binding errors | 21:45 |
| rloo | TheJulia: do we care about string freeze? I'm guessing not? | 21:46 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: Fix for proper deployment when keystone enabled and minimal OS https://review.openstack.org/540713 | 21:47 |
| *** snasir has quit IRC | 21:49 | |
| *** snasir has joined #openstack-ironic | 21:49 | |
| peitolm | "Refusing to bind due to unsupported vnic_type: baremetal bind_port /usr/lib/python2.7/dist-packages/neutron/plugins/ml2/drivers/mech_agent.py:93" hmm... | 21:49 |
| *** mjturek has quit IRC | 21:51 | |
| *** etingof has quit IRC | 21:52 | |
| *** mjturek has joined #openstack-ironic | 21:53 | |
| openstackgerrit | Merged openstack/ironic master: Replace chinese quotes to English quotes https://review.openstack.org/541206 | 21:53 |
| TheJulia | rloo: given our release model, we pretty much have to be willing to take the patches as we go, so not really | 21:53 |
| rloo | TheJulia: that's what I thought :) | 21:54 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: Make Fedora26 CI job voting https://review.openstack.org/540524 | 21:54 |
| *** ijw has quit IRC | 21:58 | |
| *** etingof has joined #openstack-ironic | 22:06 | |
| *** ijw has joined #openstack-ironic | 22:07 | |
| * TheJulia is watching/waiting for http://zuul.openstack.org/stream.html?uuid=3f91077e2f9f4b7dac39ba03dfa2f37d&logfile=console.log | 22:08 | |
| TheJulia | rloo: if there are no disagreements and the backport to stable/pike for devstack changes cleans up the job issues, I'm thinking we should just immediately re-enable the multinode gate | 22:11 |
| rloo | TheJulia: ++ | 22:11 |
| *** mjturek has quit IRC | 22:13 | |
| TheJulia | oooh it is post-stacking | 22:13 |
| *** e0ne has quit IRC | 22:15 | |
| *** openstackgerrit has quit IRC | 22:16 | |
| TheJulia | networking-baremetal has 2 patches remaining | 22:17 |
| TheJulia | smoke test running | 22:26 |
| NobodyCam | oh is it 4:20 already :p | 22:26 |
| * NobodyCam _ducks_ | 22:26 | |
| *** Goneri has quit IRC | 22:27 | |
| *** mjturek has joined #openstack-ironic | 22:28 | |
| *** rcernin has joined #openstack-ironic | 22:28 | |
| TheJulia | heh | 22:31 |
| *** mjturek has quit IRC | 22:33 | |
| *** mjturek has joined #openstack-ironic | 22:34 | |
| *** openstackgerrit has joined #openstack-ironic | 22:39 | |
| openstackgerrit | Merged openstack/bifrost master: roles: bifrost-ironic-install: Download the iPXE loader file https://review.openstack.org/539298 | 22:39 |
| * peitolm wonders where to start looking at why neutron can't bind this port, http://paste.openstack.org/show/664104/ | 22:43 | |
| openstackgerrit | Merged openstack/bifrost master: Add creation of openrc file when keystone enabled https://review.openstack.org/540096 | 22:43 |
| TheJulia | peitolm: oh, sorry, completely got side tracked | 22:48 |
| peitolm | that's o.k., any help is appreciated, it just seems like a common,but v.difficult issue which boils down to "you've done something wrong" | 22:49 |
| TheJulia | so which network_interface are you using? | 22:50 |
| peitolm | um, sorry, i'm in a maze of twisty passages with that question, | 22:51 |
| TheJulia | oh my | 22:52 |
| peitolm | # grep network_inter /etc/ironic/ironic.conf | grep -v '^#' | 22:52 |
| peitolm | enabled_network_interfaces = flat,noop,neutron | 22:52 |
| peitolm | default_network_interface = neutron | 22:52 |
| TheJulia | I would guess flat, because there is no binding profile sent, which then makes me wonder if you have the required information on the node's port(s), and which network_interface your using | 22:52 |
| peitolm | is that what you mean | 22:53 |
| peitolm | ah | 22:53 |
| TheJulia | kind of | 22:53 |
| peitolm | provider:network_type | flat | 22:53 |
| TheJulia | ahh, okay | 22:53 |
| peitolm | from "# openstack network show 2ca2c8b2-f1e5-4eb2-8d4d-0b180d017f19 | 22:53 |
| peitolm | " | 22:53 |
| TheJulia | awesome | 22:53 |
| peitolm | (which is in the paste :P) | 22:53 |
| TheJulia | so, your default_network_interface should be likely "flat" in that case | 22:54 |
| peitolm | ah | 22:54 |
| TheJulia | I believe your port binding is likely to always report failed | 22:56 |
| TheJulia | in flat | 22:56 |
| *** ijw has quit IRC | 22:57 | |
| peitolm | ok. | 22:58 |
| peitolm | so that may be a red herring then | 22:58 |
| peitolm | i'm trying to get dhcp/bootp packets in | 22:59 |
| TheJulia | peitolm: this patch might be a little eye-opening. it landed on master about a month ago and describes the situation you with flat networks https://github.com/openstack/ironic/commit/f1e5b92e446a7bf15e052aaab2692c3d4f5710a1#diff-0f2bb064f00c135ecda1486ef58a0584R17 | 22:59 |
| TheJulia | in fron where? | 22:59 |
| TheJulia | and to where? | 22:59 |
| peitolm | from the baremetal host to the dhcp server on the cleaning network | 23:00 |
| TheJulia | okay, and the flat network you referred to. can you ping the interface on that network from the network baremetal nodes are attached to? | 23:00 |
| TheJulia | specifically by network interface in that last statement, I'm referring to the the router that allows connectivity to ironic-api | 23:01 |
| peitolm | http://paste.openstack.org/show/664130/ may help | 23:03 |
| * TheJulia looks | 23:04 | |
| peitolm | the node (pe1950) is trying to get a DHCP to do the initial deploy/cleaning/whatever | 23:05 |
| peitolm | i can see bootp on the raw interface "enx0050b622b509" | 23:05 |
| peitolm | as i understand it, i'd expect to see br-int and br-eth2 showing UP, and tap402407e1-53 to be bound onto the ovs bridge | 23:07 |
| TheJulia | okay | 23:09 |
| TheJulia | so neutron is basically looking for physnet1, except you have bridges in OVS and no actual physical interface attached to that | 23:10 |
| peitolm | # grep physnet1 /etc/neutron/plugins/ml2/ml2_conf.ini | 23:10 |
| peitolm | flat_networks = physnet1 | 23:11 |
| peitolm | bridge_mappings = physnet1:br-eth2 | 23:11 |
| TheJulia | okay | 23:11 |
| peitolm | unless that should be something else br-<interfacename> | 23:12 |
| TheJulia | but br-eth2 doesn't have anything in it physically | 23:12 |
| *** r-daneel has quit IRC | 23:12 | |
| *** r-daneel has joined #openstack-ironic | 23:13 | |
| peitolm | Port "enx0050b622b509" Interface "enx0050b622b509" | 23:13 |
| peitolm | that's the network interface onto the provisioning/cleaning/tenant network | 23:16 |
| TheJulia | doh, missed that | 23:16 |
| peitolm | no worries, it looks like line noise :) | 23:16 |
| peitolm | tbh, that's the device i'm tcpdumping | 23:16 |
| peitolm | as the bridge is down, i can't tcpdump that | 23:17 |
| TheJulia | oh, ifup the bridge | 23:17 |
| TheJulia | well | 23:17 |
| TheJulia | ip link blah blah blah set up | 23:18 |
| TheJulia | ifconfig br-eth2 up | 23:18 |
| * TheJulia thinks it is up | 23:18 | |
| peitolm | *nods* | 23:18 |
| peitolm | it is now | 23:18 |
| peitolm | wonder why openswitch isn't bringing it up | 23:18 |
| TheJulia | *shrugs* | 23:19 |
| peitolm | o.k. i can see packets on br-eth2, now to step over the link | 23:20 |
| * TheJulia crosses fingers | 23:20 | |
| peitolm | not seeing it on br-int (even after 'UP'ing it | 23:20 |
| TheJulia | anything on the tap? | 23:21 |
| peitolm | nope, even after adding it to the bridge | 23:22 |
| peitolm | # ovs-ofctl dump-flows br-eth2 | 23:23 |
| peitolm | cookie=0x6988756dbafd26d7, duration=7183.245s, table=0, n_packets=16, n_bytes=1327, priority=2,in_port="phy-br-eth2" actions=drop | 23:23 |
| peitolm | cookie=0x6988756dbafd26d7, duration=7183.250s, table=0, n_packets=423, n_bytes=240711, priority=0 actions=NORMAL | 23:23 |
| TheJulia | interesting | 23:23 |
| * TheJulia sighs due to grenade failure | 23:24 | |
| peitolm | i'll let you get back to that, i'm going to have to go to the land of nod | 23:24 |
| TheJulia | rest well! | 23:25 |
| peitolm | time to power off the noisy machines :) | 23:25 |
| TheJulia | :) | 23:25 |
| peitolm | thanks for your help Julia, hopefully i'll progress past complete newbie status soon | 23:27 |
| TheJulia | I'm sure you will | 23:29 |
| TheJulia | anyway, back to test failures :( | 23:29 |
| *** mjturek has quit IRC | 23:29 | |
| *** bfournie has quit IRC | 23:30 | |
| *** rbudden has quit IRC | 23:30 | |
| *** charcol-v2 has joined #openstack-ironic | 23:33 | |
| *** MattMan has quit IRC | 23:37 | |
| *** MattMan has joined #openstack-ironic | 23:38 | |
| *** absubram has quit IRC | 23:38 | |
| *** jcoufal has quit IRC | 23:42 | |
| *** absubram has joined #openstack-ironic | 23:51 | |
| *** absubram has quit IRC | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!