| *** gyee has quit IRC | 00:00 | |
| *** rpioso is now known as rpioso|afk | 00:13 | |
| *** rloo has quit IRC | 00:22 | |
| *** livelace has quit IRC | 00:43 | |
| *** livelace has joined #openstack-ironic | 00:43 | |
| *** MattMan_ has quit IRC | 00:45 | |
| *** MattMan_ has joined #openstack-ironic | 00:46 | |
| *** moshele has quit IRC | 00:53 | |
| *** phuongnh has joined #openstack-ironic | 01:02 | |
| *** Bhujay has joined #openstack-ironic | 02:00 | |
| *** zhangfei has joined #openstack-ironic | 02:00 | |
| *** gcb_ has joined #openstack-ironic | 02:07 | |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic-inspector master: Wrap rpc server into oslo.service https://review.openstack.org/584758 | 03:03 |
|---|---|---|
| *** Bhujay has quit IRC | 03:24 | |
| *** rh-jelabarre has quit IRC | 03:38 | |
| *** ijw has joined #openstack-ironic | 04:13 | |
| *** hoonetorg has quit IRC | 04:29 | |
| *** Bhujay has joined #openstack-ironic | 04:45 | |
| *** hoonetorg has joined #openstack-ironic | 04:47 | |
| *** Bhujay has quit IRC | 04:51 | |
| *** Bhujay has joined #openstack-ironic | 04:53 | |
| *** hoonetorg has quit IRC | 04:54 | |
| *** hoonetorg has joined #openstack-ironic | 05:06 | |
| *** skazi has joined #openstack-ironic | 05:08 | |
| *** ijw has quit IRC | 05:35 | |
| *** zhangfei has quit IRC | 05:59 | |
| *** moshele has joined #openstack-ironic | 06:01 | |
| *** rcernin has quit IRC | 06:38 | |
| *** rcernin has joined #openstack-ironic | 06:40 | |
| *** pcaruana has joined #openstack-ironic | 06:42 | |
| *** rcernin has quit IRC | 06:51 | |
| *** tssurya has joined #openstack-ironic | 06:52 | |
| openstackgerrit | wu.chunyang proposed openstack/python-ironicclient master: fix typo https://review.openstack.org/594046 | 07:02 |
| *** jtomasek has joined #openstack-ironic | 07:24 | |
| *** gkadam has joined #openstack-ironic | 07:25 | |
| *** jtomasek has quit IRC | 07:33 | |
| *** jtomasek has joined #openstack-ironic | 07:33 | |
| *** w-miller has joined #openstack-ironic | 07:52 | |
| openstackgerrit | zhaixiaojun proposed openstack/ironic-inspector master: add python 3.6 tox-functional unit test job https://review.openstack.org/594057 | 07:59 |
| yolanda | hi, question... how can i override the ipa-api-url endpoint? | 08:07 |
| yolanda | i tried to setup ipa-api-url setting inside [pxe] section of ironic.conf | 08:07 |
| yolanda | i also tried to add it into pxe_append_params | 08:07 |
| yolanda | but it still seems to be adding the ironic endpoint i have in keystone, not overridiing it | 08:08 |
| *** e0ne has joined #openstack-ironic | 08:10 | |
| *** zhangfei has joined #openstack-ironic | 08:12 | |
| *** pmannidi has quit IRC | 08:12 | |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic-inspector master: Wrap Flask into oslo.service https://review.openstack.org/561823 | 08:20 |
| rpittau | yolanda: my understanding is that the ipa-api-url value needs to be set under the [conductor] section as api_url | 08:28 |
| yolanda | ah, not under pxe? | 08:29 |
| yolanda | i was expecting it to be at same level as pxe_append_params | 08:29 |
| *** zhangfei has quit IRC | 08:29 | |
| rpittau | no, I'm actually reviewing different docs right now and I see that it's under [conductor] | 08:29 |
| rpittau | sounds weird to me too, although it's marked for removal and it will be replaced by [service_catalog]endpoint_override option | 08:30 |
| yolanda | rpittau, ok , let me try that | 08:33 |
| yolanda | i ended overriding the ipxe template and hardcoding, but that's not good | 08:33 |
| rpittau | yeah :) | 08:33 |
| yolanda | also, for using ironic standalone, is not possible to specify local images instead of http? i tried to pass a file:/// path | 08:35 |
| yolanda | and i get | 08:35 |
| yolanda | Error: Validation of image href file:///var/lib/ironic_images/cirros-0.4.0-x86_64-disk.img failed, reason: No connection adapters were found for 'file:/// | 08:35 |
| yolanda | seems to think it's an href | 08:35 |
| yolanda | i see people using file there | 08:37 |
| openstackgerrit | zhaixiaojun proposed openstack/ironic-inspector master: add python 3.6 unit test job https://review.openstack.org/594057 | 08:42 |
| rpittau | yes file href should be supported as well afaik | 08:42 |
| yolanda | mm | 08:43 |
| yolanda | Agent deploy supports only HTTP(S) URLs as instance_info['image_source'] or swift temporary URL | 08:43 |
| yolanda | i'd need to configure properly | 08:43 |
| *** zhangfei has joined #openstack-ironic | 08:45 | |
| openstackgerrit | zhaixiaojun proposed openstack/ironic-inspector master: add python 3.6 unit test job https://review.openstack.org/594057 | 08:46 |
| openstackgerrit | zhaixiaojun proposed openstack/ironic-inspector master: add python 3.6 unit test job https://review.openstack.org/594057 | 08:48 |
| openstackgerrit | Kaifeng Wang proposed openstack/ironic-inspector master: Wrap Flask into oslo.service https://review.openstack.org/561823 | 09:02 |
| *** ibmko has joined #openstack-ironic | 09:05 | |
| *** mgoddard has joined #openstack-ironic | 09:10 | |
| *** S4ren has joined #openstack-ironic | 09:29 | |
| openstackgerrit | Harald Jensås proposed openstack/ironic-inspector master: Add rootwrap filter for systemctl control of dnsmasq https://review.openstack.org/593114 | 09:44 |
| *** dtantsur|afk is now known as dtantsur | 09:59 | |
| dtantsur | morning ironic | 09:59 |
| openstackgerrit | Aija Jaunteva proposed openstack/sushy master: Add support for loading resources from archive file https://review.openstack.org/589147 | 10:01 |
| etingof | good morning dtantsur \o | 10:03 |
| openstackgerrit | Aija Jaunteva proposed openstack/sushy master: Add support for loading resources from archive file https://review.openstack.org/589147 | 10:08 |
| sambetts|afk | dtantsur: Do you know if anyone has starting writing a spec for standard firmware managment in ironic?/ | 10:17 |
| sambetts|afk | cdearborn: I saw you created an RFE related to this recently, any ideas? ^ | 10:18 |
| *** rpioso|afk is now known as rpioso | 10:19 | |
| rpioso | Good morning | 10:19 |
| *** Bhujay has quit IRC | 10:26 | |
| dtantsur | morning rpioso, isn't it super early for you? :) | 10:26 |
| dtantsur | sambetts|afk: I think cdearbord will | 10:26 |
| dtantsur | * cdearborn | 10:27 |
| sambetts|afk | dtantsur: awesome, got some renewed interest from downstream in that feature, so wasn't sure if anyone had picked up the work since Dublin | 10:28 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-tempest-plugin master: Stop relying on VCPU count in the introspection tests https://review.openstack.org/594120 | 10:33 |
| dtantsur | vdrok, sambetts|afk, jroll ^^^ | 10:33 |
| dtantsur | I wonder if it's right, but what we have now is broken :( | 10:33 |
| jroll | morning dtantsur, +W | 10:42 |
| dtantsur | thanks and good morning | 10:42 |
| jroll | sorry, I knew that landed but didn't realize we relied on it :( | 10:42 |
| sambetts|afk | I wondered if we needed it to take in a number to compare against incase the tempest plugin is run against an existing OpenStack that might already have other hypervisors enrolled | 10:43 |
| dtantsur | sambetts|afk: we don't know this number though | 10:43 |
| dtantsur | and this was not accounted for previously either, so we can make it a future improvement | 10:44 |
| * dtantsur puts on his backlog | 10:44 | |
| sambetts|afk | dtantsur: we have the number its len(self.node_ids) but yeah can be a follow up | 10:46 |
| sambetts|afk | to replicate the existing behaviour | 10:47 |
| sambetts|afk | that is, to make it better it should check the count, move them to inspected, then check the count goes back to the count it was before inspecting | 10:47 |
| jroll | sambetts|afk: eh, presumably if it's an existing install, len(self.node_ids) can change | 10:47 |
| jroll | because it's all available nodes at the time the tests start | 10:48 |
| sambetts|afk | yeah len(self.node_ids) only preserves the existing behaviour which is that all nodes are ironic nodes and all go into inspecting, the new logic in that patch will pass if only one BM succeeds inspecting right? | 10:49 |
| jroll | existing behavior is that at least one node for the flavor we use is available | 10:49 |
| jroll | expected_cpus = self.baremetal_flavor()['vcpus'] | 10:49 |
| jroll | if int(stats['hypervisor_statistics']['vcpus']) >= expected_cpus: | 10:50 |
| jroll | with the patch, it's at least one node is available at all | 10:50 |
| jroll | I think it's equally broken for existing installs | 10:50 |
| sambetts|afk | ah ok yeah, got it misunderstood I thought it waited for all Bms to finish, /me needs more coffee obviously | 10:51 |
| jroll | :) | 10:51 |
| sambetts|afk | has the logic merged in nova to stop removing the ironic hypevisors and make them reserved instead yet? I saw a patch for that a little while ago | 10:52 |
| sambetts|afk | because then the count won't change when the nodes move to inspecting | 10:52 |
| dtantsur | sambetts|afk: I think it has merged, yes | 10:53 |
| jroll | ooo | 10:53 |
| jroll | yes, it did | 10:53 |
| * dtantsur ponders replacing the function in question with sleep(60) :( | 10:53 | |
| dtantsur | well, the CI runs the inspection job for each branch. if it passes, we'll have some degree of confidence that this approach is an acceptable workaround for now :) | 10:54 |
| dtantsur | which reminds me: we should add rocky jobs for ironic-tempest-plugin. bring the total number of jobs to +insanity :) | 10:54 |
| sambetts|afk | you havn't seen insanity until you've seen the networking-cisco CI: https://review.openstack.org/#/c/592166/3 | 10:56 |
| patchbot | patch 592166 - networking-cisco - Remove MultiConfigParser from the device manger co... - 1h 55m 24s spent in CI | 10:56 |
| dtantsur | OMG | 10:56 |
| jroll | nice | 10:56 |
| jroll | I'm just impressed you can get things to run on mitaka | 10:57 |
| sambetts|afk | so am I :-P | 10:57 |
| *** S4ren has quit IRC | 11:03 | |
| rpioso | Morning, dtantsur | 11:07 |
| *** rpioso is now known as rpioso|afk | 11:09 | |
| ibmko | Hi all, may I have a question about multi-tenant networking in Ironic: what are the thoughts about having bare metal server connected to multiple networks? Is it somehow possible/planned to deliver multiple network via one physical port using VLAN tagging (meaning that operating system on the bare metal server would be tagging/untagging VLANs and having per vlan virtual interfaces) | 11:10 |
| sambetts|afk | ibmko: neutron supports "trunks" which if combined with an ML2 driver that supports baremetal and trunks can provide multiple networks over one physical port | 11:12 |
| sambetts|afk | the subinterfaces are/will need to be created either by the owner of the baremetal, or via cloud-init | 11:13 |
| *** Bhujay has joined #openstack-ironic | 11:13 | |
| ibmko | sambetts|afk, and these trunk ports are currently supported in Ironic Queens ? | 11:14 |
| ibmko | sambetts|afk, how can this be created by owned of the baremetal if he has no connectivity until he sets it up? | 11:15 |
| sambetts|afk | ibmko: yes, however the support is dependant more on the ml2 driver you are using than Ironic, neutron trunks appear to Ironic as normal neutron ports, except they have "sub-ports" | 11:15 |
| sambetts|afk | ibmko: every trunk port in neutron has a native vlan, which will always have connectivty | 11:16 |
| ibmko | sambetts|afk, that sounds good, when you are talking about ml2 drivers, are you talking about mechanism drivers? We use openvswitch - do you think trunks are supported here ? | 11:18 |
| sambetts|afk | ibmko: the openvswitch ml2 mechansim driver doesn't support baremetal that is only for VMs, you need an ML2 mechansim driver that supports the baremetal vnic type and will program the TOR above the baremetal, for example networking-baremetal provides an ml2 mech driver to generically programming most switches that have SSH communication, or networking-cisco provides an ML2 mech driver for | 11:22 |
| sambetts|afk | more advanced programming of Cisco Nexus 9000 series switches which includes baremeal and trunk support | 11:22 |
| sambetts|afk | sorry not netwokring-baremetal I meant https://github.com/openstack/networking-generic-switch | 11:23 |
| ibmko | sambetts|afk, ah, sorry I confused it as well - we actually use the netwokring-baremetal | 11:24 |
| ibmko | eeerh | 11:24 |
| ibmko | I mean the networking-generic-switch :)) | 11:24 |
| ibmko | that's what we use | 11:25 |
| ibmko | we use it to provision our juniper switch via ssh/netmiko | 11:26 |
| sambetts|afk | I don't see anything in the networking-gneric-switch repo that suggests support for trunks yet, but it is a feature that could be added I think | 11:26 |
| ibmko | sambetts|afk, so you think that this is currently only supported by networking-cisco ? | 11:28 |
| sambetts|afk | ibmko: as that is my teams driver I know its supported, not sure if other vendor's ML2 drivers have added support, I know arista had their own ML2 driver for baremetal too | 11:29 |
| ibmko | sambetts|afk, understood, and last question, do you think that whole provisioning could be automated if we use some input script for the baremetal to set up the inner virtual interfaces ? | 11:31 |
| sambetts|afk | ibmko: absolutely, I think there maybe enhancements to nova to support trunks for VMs by passing network config in via config-drive so cloud-init can program it, unsure if that works for Ironic nodes yet, but its completely doable to pass a user-data script into nova to configure the trunk ports how you would like, we did this for a downstream application | 11:34 |
| ibmko | sambetts|afk, great, thanks for all the information :) | 11:35 |
| sambetts|afk | no problem : | 11:35 |
| sambetts|afk | :) | 11:35 |
| *** e0ne has quit IRC | 11:36 | |
| *** zhangfei has quit IRC | 11:37 | |
| *** ibmko has quit IRC | 11:42 | |
| *** rh-jelabarre has joined #openstack-ironic | 11:49 | |
| *** trown|outtypewww is now known as trown | 11:50 | |
| *** phuongnh has quit IRC | 11:50 | |
| arne_wiebalck | TheJulia: Sorry, not yet: I was still struggling to boot an instance of a s/w RAID. Hence for now, I disabled the discovery by root hints and hardcoded what I’d like to have as a result. | 11:55 |
| openstackgerrit | Merged openstack/bifrost stable/queens: Refactor bifrost-integration-dhcp jobs https://review.openstack.org/593631 | 12:05 |
| *** ibmko has joined #openstack-ironic | 12:12 | |
| ibmko | sambetts|afk, sorry, maybe one more question :) | 12:12 |
| ibmko | sambetts|afk, regarding bonding - I see there is some support via the ironic port groups. I am wondering however if bonding configuration has to be prepared manually on the ToR switch or if the drivers have the capability to provision bonding configuration on the switch side too | 12:14 |
| TheJulia | arne_wiebalck: that is to be expected with raid I think | 12:16 |
| TheJulia | arne_wiebalck: but how your doing raid might impact that and all | 12:16 |
| *** adrianc has joined #openstack-ironic | 12:17 | |
| arne_wiebalck | TheJulia: You mean that root hints are not working? | 12:17 |
| arne_wiebalck | TheJulia: Or that it is difficult to boot off a s/w RAID? | 12:18 |
| TheJulia | arne_wiebalck: no they work, but we also have code wrapped up in that evaluation | 12:18 |
| TheJulia | arne_wiebalck: hardware raid is kind of hard to build and boot. I suspect mirrors would work, I expect we would (eventually) need to understand making mirror sets with partitions to hold the kernels/ramdisks | 12:19 |
| * TheJulia makes coffee | 12:20 | |
| dtantsur | morning TheJulia | 12:20 |
| arne_wiebalck | TheJulia: Basically what I have for now is instructing our h/w manager create the RAID and use it as the OS install drive. After that, the instance boots, but not from the RAID, only from one leg :-/ | 12:23 |
| arne_wiebalck | TheJulia: I changed the config on the node and it seems to boot from the s/w RAID now, but I need to figure out which of my changes did the trick … | 12:24 |
| * dtantsur hopes root_device setting | 12:24 | |
| TheJulia | I didn't see anyhting that would prevent /dev/md0 from being set as a root device | 12:25 |
| TheJulia | but I might have missed something when walking through ironic lib | 12:25 |
| TheJulia | in theory, /dev/md0 should ideally get picked up as a smaller device with my patch, but... | 12:25 |
| yolanda | hi, question.. i'm trying to debug some standalone problem, where it seems to complain about not finding the deployment image. I'm able to enter into the ipa, by ssh, but ironic is just powering down my node all the time. How can i avoid it? Setting into mainteannce mode will help? or it will have some other collateral effects? | 12:27 |
| arne_wiebalck | TheJulia: I’ll get back to the patch (and test the root device hints) once I understood how to make my RAID boot w/o manual help after deployment. | 12:27 |
| TheJulia | The one leg makes sense without setting root_device hints, at least for a mirror set | 12:27 |
| TheJulia | arne_wiebalck: awesome, thanks! | 12:27 |
| TheJulia | yolanda: I've typically short circuited continue_deployment's end to troubleshoot things like that | 12:28 |
| arne_wiebalck | TheJulia: Maybe I misunderstand what the hints are for … | 12:28 |
| TheJulia | arne_wiebalck: they are kind of confusing :( | 12:28 |
| TheJulia | yolanda: are the logs shipped up to the conductor not revealing enough information? | 12:29 |
| arne_wiebalck | TheJulia: I thought they are used as a help to fid the device onto which the OS shall be installed, no? | 12:29 |
| arne_wiebalck | s/fid/find/ | 12:29 |
| yolanda | TheJulia, no... i just get a Deploy failed for instance None. Error: [Errno 2] No such file or directory in the conductor | 12:29 |
| yolanda | the image is there | 12:29 |
| yolanda | and in the host i try to boot, i can see iscsi drive connected successfully | 12:29 |
| yolanda | so no idea | 12:29 |
| TheJulia | yolanda: sounds like your missing the folder on the conductor | 12:29 |
| TheJulia | for the conductor logs | 12:30 |
| yolanda | it's there | 12:30 |
| yolanda | for the logs? | 12:30 |
| TheJulia | yeah, where the conductor would write agent logs | 12:30 |
| *** jcoufal has joined #openstack-ironic | 12:30 | |
| yolanda | and if that doesn't exist, fails that that "No such file or directory" ? | 12:30 |
| TheJulia | The other possibility is permissions are blocking the conductor. I take that is from the conductor in debug logging mode? | 12:30 |
| TheJulia | yolanda: yeah, I suspect it can | 12:31 |
| yolanda | yep, conductor has debug=true | 12:31 |
| yolanda | it's on containers, i see all this info in docker logs in the container | 12:31 |
| yolanda | let me paste you more info | 12:31 |
| yolanda | mm, error actually comes from iscsi | 12:32 |
| yolanda | TheJulia, http://paste.openstack.org/show/728515/ | 12:32 |
| yolanda | in the host i provision, i see Created iscsi iqn... so that's ok | 12:34 |
| yolanda | last i see in the host, is 2018-08-21 12:24:54.688 1824 INFO root [-] Command log.collect_system_logs completed: Command name: collect_system_logs, params: {}, status: SUCCEEDED, result: {'system_logs': | 12:34 |
| yolanda | after that, it just powers down | 12:34 |
| sambetts|afk | ibmko: again that depends on the functionality of the ML2 driver, some are able to program it on the fly, others require it be preprogramed | 12:39 |
| ibmko | sambetts|afk, so how about the driver of yours ? :) | 12:40 |
| ibmko | the cisco one | 12:40 |
| sambetts|afk | ibmko: we create the bond on the fly so no need to preprogram them | 12:40 |
| ibmko | cool | 12:40 |
| ibmko | seems we need to ditch juniper :) | 12:41 |
| sambetts|afk | ibmko: looks like the juniper plugin for networking-generic-switch might support trunking after all: https://github.com/openstack/networking-generic-switch/blob/master/networking_generic_switch/devices/netmiko_devices/juniper.py#L57 but I don't see anything regarding bonds being dynamic | 12:44 |
| ibmko | sambetts|afk, cool thanks, I'll play around with it | 12:45 |
| TheJulia | yolanda: is this master branch? | 12:47 |
| TheJulia | yolanda: most likely the conductor can't find iscsiadm | 12:49 |
| TheJulia | it may be present but also may not be in the PATH that is being set for the process to execute with | 12:49 |
| *** rloo has joined #openstack-ironic | 12:55 | |
| yolanda | TheJulia, using pike | 12:55 |
| TheJulia | https://github.com/openstack/ironic/blob/stable/pike/ironic/drivers/modules/deploy_utils.py#L124 | 12:56 |
| TheJulia | We should likely put catches in to improve the logging "INSTALL XYZ" | 12:56 |
| yolanda | so it's a problem of communication between conductor and the host? | 12:59 |
| TheJulia | no | 12:59 |
| TheJulia | missing iscsiadm binary | 12:59 |
| yolanda | let me check | 12:59 |
| yolanda | mm, looks as a problem with the image when installing that | 13:04 |
| openstackgerrit | Merged openstack/bifrost stable/rocky: Refactor bifrost-integration-tinyipa jobs https://review.openstack.org/593626 | 13:04 |
| sambetts|afk | TheJulia, yolanda: could it be related too: https://review.openstack.org/#/c/557687/ I remember seeing a similarly weird log when I was writing a whole disk image as a partition image or vs versa (image_type was set wrong) and it tried to do the config drive but because the file system it had written was junk it couldn't open the location to write the config drive | 13:08 |
| patchbot | patch 557687 - ironic-python-agent - Attempt to read the partition table after writing ... - 6h 33m 35s spent in CI | 13:08 |
| yolanda | iscsiadm does not exist in my conductor | 13:09 |
| TheJulia | sambetts|afk: interesting | 13:12 |
| TheJulia | yolanda: that would do it | 13:13 |
| openstackgerrit | Merged openstack/ironic-tempest-plugin master: Stop relying on VCPU count in the introspection tests https://review.openstack.org/594120 | 13:15 |
| *** arne_wiebalck has quit IRC | 13:17 | |
| *** amoralej has joined #openstack-ironic | 13:18 | |
| *** arne_wiebalck has joined #openstack-ironic | 13:19 | |
| *** trown is now known as trown|brb | 13:27 | |
| *** trown|brb is now known as trown | 13:31 | |
| *** dougsz_ has joined #openstack-ironic | 13:34 | |
| *** e0ne has joined #openstack-ironic | 13:40 | |
| *** cdearborn has joined #openstack-ironic | 13:46 | |
| *** moshele has quit IRC | 13:53 | |
| *** jcoufal has quit IRC | 13:56 | |
| *** jcoufal has joined #openstack-ironic | 13:57 | |
| openstackgerrit | Ilya Etingof proposed openstack/sushy master: Add a virtual media resource https://review.openstack.org/570810 | 14:12 |
| dtantsur | TheJulia: I would like to release queens and pike for sushy, any objections? | 14:26 |
| dtantsur | oh, I see you have pike. then queens, I guess. | 14:27 |
| NobodyCam | Good Morning Ironic'ers | 14:29 |
| dtantsur | morning NobodyCam | 14:29 |
| NobodyCam | Hey Hey dtantsur happy Tuesday :) | 14:29 |
| dtantsur | good news: the inspector CI seems back up | 14:32 |
| *** cdearborn has quit IRC | 14:32 | |
| NobodyCam | \o/ | 14:35 |
| anupn | Hi Folks, good morning. I have a quick one, I see "InstanceDeployFailure: Failed to deploy instance: Failed to start the iSCSI target to deploy the node" in ir-cond logs, can someone point me where should I look to fix this issue. I have correct hypervisor stats. | 14:37 |
| openstackgerrit | Dmitry Tantsur proposed openstack/sushy-tools master: Bump hacking version to 1.0 (capped by 1.2) https://review.openstack.org/594248 | 14:38 |
| *** cdearborn has joined #openstack-ironic | 14:44 | |
| anupn | I see "DeviceNotFound: Error finding disk or partition devices" inside the deploy logs, not sure why is this, because I am using whole-disk image for deploying | 14:45 |
| TheJulia | dtantsur: ++ for queens | 14:46 |
| TheJulia | dtantsur: \o/ | 14:46 |
| TheJulia | Good morning NobodyCam | 14:46 |
| NobodyCam | Howdy TheJulia :) how goes New Mexico | 14:47 |
| TheJulia | anupn: try switching to the direct interface | 14:47 |
| TheJulia | NobodyCam: dry | 14:47 |
| NobodyCam | :) thats a good thing | 14:47 |
| *** r-daneel has joined #openstack-ironic | 14:47 | |
| NobodyCam | unless its a Dry state | 14:47 |
| NobodyCam | that could be bad | 14:47 |
| *** rpioso|afk is now known as rpioso | 14:48 | |
| anupn | TheJulia: Okay cool | 14:50 |
| anupn | TheJulia: ALso what I noticed is my virsh VM that I am deploying on is in power off state. I will just power that on, to make sure that's not a problem | 14:52 |
| TheJulia | anupn: I suspect something is failing connectivity wise that might be causing iscsi to work, if your doing virutal testing, it could be hypervisor firewalling. | 14:52 |
| TheJulia | anupn: It can take a moment to start up, but it sounds like there may be a deeper problem there | 14:53 |
| anupn | TheJulia: What do you mean firewalling? Ok, I am now trying with direct deploy | 14:53 |
| TheJulia | hypervisors can have firewall rules that impact the VMs, so if your hitting this in development/testing, it seems like for some reason an iscsi connection is not able to start up... that or maybe even iscsi tools are missing from the IPA ramdisk your using. | 14:54 |
| anupn | TheJulia: Hmm I see, I am using the master copy of IPA, the one devstack use. | 14:56 |
| anupn | but sounds then if iscsi is not starting up, direct interface might work there | 14:57 |
| openstackgerrit | Paul Belanger proposed openstack/bifrost stable/pike: Refactor bifrost-integration-dhcp jobs https://review.openstack.org/594267 | 14:57 |
| openstackgerrit | Paul Belanger proposed openstack/bifrost stable/pike: Refactor bifrost-integration-dibipa-debian jobs https://review.openstack.org/594268 | 14:57 |
| openstackgerrit | Paul Belanger proposed openstack/bifrost stable/pike: Refactor bifrost-integration-tinyipa jobs https://review.openstack.org/594269 | 14:57 |
| rpioso | dtantsur, jroll, sambetts|afk, TheJulia: Yesterday, I asked TheJulia if an RFE is required, preferred, or not needed for adding a supported interface to a h/w type. The default interface would not be changed. It's about adding the Redfish management and power interfaces to the idrac h/w type. | 14:57 |
| rpioso | I mentioned that there may be a need to enhance tripleo's support of the idrac h/w type, but that's off this channel's topic. The interfaces expect their creds to be on the ironic node object. | 14:58 |
| rpioso | Thoughts? | 14:58 |
| anupn | TheJulia: Same message "Error finding the disk or partition to deploy image onto. No suitable device was found for deployment" | 15:00 |
| sambetts|afk | rpioso: I think ideally there would be a very short RFE just to make sure the work is tracked somewhere, but no need for a spec or anything, the RFE is almost purly for tracking purposes, I don't think the tripleO part affects it at all, we already have h/w types with mixed interfaces that require different driver_info for different interfaces | 15:02 |
| TheJulia | rpioso: I think it would be good to track as an RFE from our point or view, that also gives you more credence if tripleo folks push back at the change you'd need to make. | 15:02 |
| dtantsur | rpioso: just an RFE, no spec, I think. | 15:02 |
| dtantsur | tripleo is a different topic, but we can fix it as well. I can point you where. | 15:02 |
| TheJulia | I have absolutely no objection to the additional interfaces though, it was what we intended with hardware types and interface model. | 15:03 |
| rpioso | dtantsur, sambetts|afk, TheJulia: Thank you | 15:03 |
| rpioso | dtantsur: A pointer would be appreciated :) | 15:03 |
| dtantsur | rpioso: let's move to #tripleo | 15:04 |
| jroll | simple RFE is good with me as well | 15:04 |
| rpioso | dtantsur: ack | 15:05 |
| rpioso | jroll: Thanks :) | 15:05 |
| jroll | :) | 15:05 |
| *** pcaruana has quit IRC | 15:09 | |
| *** priteau has joined #openstack-ironic | 15:13 | |
| anupn | TheJulia: I see in IPA logs, that it waits for root device to come up, but then it throws No Mellanox device found in the logs. Here http://paste.openstack.org/show/728532/ | 15:14 |
| anupn | does that mean it is trying to find the Mellanox device? | 15:15 |
| sambetts|afk | anupn: that is an expected debug log becuase the mellanox device manger will only process devices if it find a mellonox device, its shouldn't affect anything | 15:16 |
| *** sambetts|afk is now known as sambetts | 15:16 | |
| anupn | sambetts: Ah alright | 15:16 |
| * anupn wonders then why it is unable to find a device | 15:17 | |
| TheJulia | anupn: all found block devices are smaller than 4294967296 | 15:17 |
| TheJulia | anupn: at least 4GB of attached storage is required. | 15:17 |
| *** moshele has joined #openstack-ironic | 15:22 | |
| anupn | TheJulia: I have local_gb set to 8gb. I will double check if VM does not have that much | 15:23 |
| *** amoralej is now known as amoralej|off | 15:26 | |
| *** Bhujay has quit IRC | 15:27 | |
| *** luzC has quit IRC | 15:42 | |
| *** e0ne has quit IRC | 15:44 | |
| *** cdearborn has quit IRC | 15:45 | |
| *** marlinc has quit IRC | 15:45 | |
| rpioso | dtantsur, jroll, sambetts|afk, TheJulia: I created a new RFE -- https://storyboard.openstack.org/#!/story/2003514 | 15:45 |
| rpioso | Does that need a spec? | 15:45 |
| *** strigazi has quit IRC | 15:46 | |
| openstackgerrit | Merged openstack/bifrost stable/queens: Refactor bifrost-integration-dibipa-debian jobs https://review.openstack.org/593632 | 15:46 |
| *** strigazi has joined #openstack-ironic | 15:46 | |
| openstackgerrit | Merged openstack/bifrost stable/queens: Refactor bifrost-integration-tinyipa jobs https://review.openstack.org/593633 | 15:46 |
| rpioso | ajya[m], etingof: ^^^ | 15:47 |
| rpioso | bdodd: ^^^ | 15:47 |
| jroll | rpioso: probably not unless there's something not-so-normal about the implementation | 15:48 |
| *** marlinc has joined #openstack-ironic | 15:53 | |
| TheJulia | rpioso: Yeah, if it is feature parity and nothing requires anything magical or new user facing, then it does not need a spec | 15:54 |
| rpioso | jroll: I don't expect anything unusual. There may be an opportunity to increase code reuse by locating common RAID code in a module that other drivers could use. However, other drivers' use of it would not be part of this RFE. | 15:55 |
| rpioso | Famous last words ;-) | 15:56 |
| jroll | heh | 15:56 |
| jroll | yeah, just for adding existing capabilities to drivers, shouldn't need a spec | 15:56 |
| *** gyee has joined #openstack-ironic | 15:56 | |
| rpioso | TheJulia: Would there be any reason to discuss this at the PTG? I'm thinking that it should come up during the cycle priorities section. Beyond that, I'm not sure. | 15:59 |
| TheJulia | rpioso: cycle priorities is the only reason I can think | 16:02 |
| rpioso | TheJulia: Should I list it on the Stein PTG etherpad? | 16:03 |
| jroll | I don't think it even needs to be in the priority list | 16:03 |
| jroll | it'll likely be straightforward, easy to review code | 16:03 |
| jroll | rubber stamp city | 16:03 |
| rpioso | jroll: I appreciate your confidence :) | 16:04 |
| TheJulia | rpioso: please | 16:05 |
| rpioso | jroll, TheJulia: It may require changes to sushy | 16:05 |
| TheJulia | my first introduction to rubber stamps was the library scene in indiana jones... | 16:05 |
| TheJulia | it shall forever be burned into my brain | 16:05 |
| jroll | oh right, we own the client library | 16:05 |
| rpioso | TheJulia: ack | 16:05 |
| jroll | I was hoping to just review `def apply_raid(...): sushy.do_the_raid_thing() | 16:06 |
| TheJulia | rpioso: yeah, additive in nature, but still likely not spec requiring because the interface is standardized | 16:06 |
| TheJulia | and I think some of the sushy stuff is coming about on its own | 16:06 |
| rpioso | TheJulia: I've seen encouraging activity there. | 16:07 |
| rpioso | TheJulia: Added, ~lines 150-152 | 16:10 |
| rpioso | jroll, TheJulia: Thanks, again! | 16:10 |
| TheJulia | rpioso: thanks! | 16:11 |
| rpioso | yw | 16:11 |
| * rpioso goes to refuel | 16:11 | |
| *** rpioso is now known as rpioso|afk | 16:12 | |
| *** adrianc has quit IRC | 16:16 | |
| *** jtomasek_ has joined #openstack-ironic | 16:25 | |
| *** jtomasek has quit IRC | 16:26 | |
| yolanda | TheJulia, that was it. Some issue with selinux being enabled on the host, and not be able to install the package on the docker image due to that | 16:27 |
| TheJulia | fun :( | 16:27 |
| TheJulia | stepping away to get a marriage license (yay!) | 16:27 |
| yolanda | congrats! | 16:27 |
| *** dtantsur is now known as dtantsur|afk | 16:38 | |
| dtantsur|afk | \o | 16:38 |
| jroll | TheJulia: \o/ | 16:41 |
| *** S4ren has joined #openstack-ironic | 16:43 | |
| *** trown is now known as trown|lunch | 16:46 | |
| *** sambetts is now known as sambetts|afk | 16:52 | |
| sambetts|afk | night all o/ | 16:52 |
| *** w-miller has quit IRC | 17:00 | |
| *** mgoddard has quit IRC | 17:01 | |
| openstackgerrit | Mark Goddard proposed openstack/networking-generic-switch master: Convert error detection to use a decorator https://review.openstack.org/592009 | 17:23 |
| *** e0ne has joined #openstack-ironic | 17:42 | |
| *** tssurya has quit IRC | 17:50 | |
| *** trown|lunch is now known as trown | 17:53 | |
| *** portdirect has quit IRC | 17:54 | |
| *** mjturek has joined #openstack-ironic | 18:10 | |
| *** NobodyCam has quit IRC | 18:37 | |
| *** leifmadsen has quit IRC | 18:37 | |
| *** vkmc has quit IRC | 18:37 | |
| *** kencjohnston has quit IRC | 18:37 | |
| *** zer0c00l has quit IRC | 18:37 | |
| *** fungi has quit IRC | 18:37 | |
| *** kencjohnston_ has joined #openstack-ironic | 18:38 | |
| *** leifmadsen has joined #openstack-ironic | 18:40 | |
| *** vkmc has joined #openstack-ironic | 18:40 | |
| *** fbouliane has quit IRC | 18:42 | |
| *** TheJulia has quit IRC | 18:42 | |
| *** fyx has quit IRC | 18:42 | |
| *** mnaser has quit IRC | 18:42 | |
| *** nicholas has quit IRC | 18:42 | |
| *** TheJulia has joined #openstack-ironic | 18:44 | |
| *** jrist has quit IRC | 18:46 | |
| *** fbouliane has joined #openstack-ironic | 18:46 | |
| *** cdearborn has joined #openstack-ironic | 18:47 | |
| *** fungi has joined #openstack-ironic | 18:48 | |
| *** e0ne has quit IRC | 18:57 | |
| *** priteau has quit IRC | 19:00 | |
| TheJulia | Reminder: vote for bug day time https://doodle.com/poll/ef4m9zmacm2ey7ce | 19:14 |
| TheJulia | since that is next week. | 19:14 |
| *** rpioso|afk is now known as rpioso | 19:17 | |
| rpioso | TheJulia: Thanks for your assistance on the backport of the idrac UEFI bug fix to stable/ocata. I suppose that means our friends released a new version of sushy. | 19:18 |
| *** bnemec has quit IRC | 19:20 | |
| *** bnemec has joined #openstack-ironic | 19:20 | |
| *** r-daneel has quit IRC | 19:24 | |
| *** r-daneel has joined #openstack-ironic | 19:24 | |
| *** moshele has quit IRC | 19:25 | |
| *** melwitt has quit IRC | 19:33 | |
| *** sdake has quit IRC | 19:33 | |
| *** BadCub has quit IRC | 19:34 | |
| *** chihhsin has quit IRC | 19:34 | |
| *** melwitt has joined #openstack-ironic | 19:34 | |
| *** sdake has joined #openstack-ironic | 19:34 | |
| *** moshele has joined #openstack-ironic | 19:36 | |
| openstackgerrit | Merged openstack/ironic stable/ocata: Fix iDRAC hardware type does not work with UEFI https://review.openstack.org/588949 | 19:37 |
| *** etingof has quit IRC | 19:39 | |
| *** jistr|off has quit IRC | 19:39 | |
| *** mraineri has quit IRC | 19:39 | |
| *** Bico_Fino_ has quit IRC | 19:39 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost master: Begin change over to use python3 https://review.openstack.org/588325 | 19:40 |
| *** sambetts|afk has quit IRC | 19:42 | |
| openstackgerrit | Julia Kreger proposed openstack/bifrost stable/queens: Pin get-pip.py to 3.2 https://review.openstack.org/594529 | 19:44 |
| openstackgerrit | Julia Kreger proposed openstack/bifrost stable/pike: Pin get-pip.py to 3.2 https://review.openstack.org/594530 | 19:45 |
| *** sambetts_ has joined #openstack-ironic | 19:45 | |
| *** moshele has quit IRC | 19:45 | |
| *** dirk has quit IRC | 19:46 | |
| *** samueldmq has quit IRC | 19:46 | |
| *** cdearborn has quit IRC | 19:47 | |
| *** jistr has joined #openstack-ironic | 19:47 | |
| *** ibmko has quit IRC | 19:48 | |
| *** samueldmq has joined #openstack-ironic | 19:49 | |
| *** chihhsin has joined #openstack-ironic | 19:49 | |
| *** r-daneel has quit IRC | 20:03 | |
| *** r-daneel has joined #openstack-ironic | 20:03 | |
| *** MattMan has joined #openstack-ironic | 20:11 | |
| *** gmann_ has joined #openstack-ironic | 20:13 | |
| *** ddurst- has joined #openstack-ironic | 20:18 | |
| *** MattMan_ has quit IRC | 20:19 | |
| *** ddurst has quit IRC | 20:19 | |
| *** gmann has quit IRC | 20:19 | |
| *** gmann_ is now known as gmann | 20:19 | |
| *** gkadam has quit IRC | 20:27 | |
| anupn | TheJulia: Yes less memory was the issue, somehow I was thinking it has larger disk size. Thanks | 20:32 |
| *** etingof has joined #openstack-ironic | 20:36 | |
| *** gkadam has joined #openstack-ironic | 20:43 | |
| *** harlowja has joined #openstack-ironic | 21:02 | |
| *** trown is now known as trown|outtypewww | 21:07 | |
| *** mjturek has quit IRC | 21:08 | |
| *** rh-jelabarre has quit IRC | 21:22 | |
| *** gkadam has quit IRC | 21:26 | |
| *** jcoufal has quit IRC | 21:55 | |
| *** dougsz_ has quit IRC | 22:03 | |
| *** rcernin has joined #openstack-ironic | 22:10 | |
| *** cxc99 has joined #openstack-ironic | 22:31 | |
| *** moshele has joined #openstack-ironic | 22:46 | |
| *** jtomasek_ has quit IRC | 22:47 | |
| *** r-daneel has quit IRC | 23:09 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!