*** rbudden has joined #openstack-ironic | 00:07 | |
*** baoli has quit IRC | 00:08 | |
*** baoli has joined #openstack-ironic | 00:09 | |
*** baoli has quit IRC | 00:14 | |
*** liuzz has quit IRC | 00:17 | |
*** liuzz has joined #openstack-ironic | 00:19 | |
*** rbudden has quit IRC | 00:19 | |
*** cjloader has joined #openstack-ironic | 00:27 | |
*** r-daneel has quit IRC | 00:29 | |
*** cjloader has quit IRC | 00:31 | |
*** jcoufal has joined #openstack-ironic | 00:47 | |
*** trungnv has quit IRC | 00:53 | |
*** trungnv has joined #openstack-ironic | 00:54 | |
*** rbudden has joined #openstack-ironic | 00:58 | |
*** rbudden has quit IRC | 00:59 | |
*** phuongnh has joined #openstack-ironic | 01:02 | |
*** baoli has joined #openstack-ironic | 01:05 | |
*** tiendc has joined #openstack-ironic | 01:06 | |
*** rbudden has joined #openstack-ironic | 01:10 | |
*** zshi has joined #openstack-ironic | 01:14 | |
*** fragatina has quit IRC | 01:16 | |
*** fragatina has joined #openstack-ironic | 01:18 | |
*** fragatina has quit IRC | 01:19 | |
*** hemna_ has quit IRC | 01:32 | |
*** namnh has joined #openstack-ironic | 01:40 | |
*** fragatin_ has joined #openstack-ironic | 01:52 | |
*** rwsu has joined #openstack-ironic | 02:03 | |
openstackgerrit | zenghui.shi proposed openstack/ironic master: Change exception msg of BIOS caching https://review.openstack.org/568193 | 02:07 |
---|---|---|
*** zhangfei has joined #openstack-ironic | 02:09 | |
*** cjloader has joined #openstack-ironic | 02:19 | |
openstackgerrit | Tuan Luong-Anh proposed openstack/ironic master: [WIP] Implement iRMC BIOS configuration https://review.openstack.org/534595 | 02:23 |
*** cjloader has quit IRC | 02:24 | |
openstackgerrit | Tuan Luong-Anh proposed openstack/ironic master: [WIP] Implement iRMC BIOS configuration https://review.openstack.org/534595 | 02:27 |
openstackgerrit | Merged openstack/ironic master: Add reno for new config [disk_utils]partprobe_attempts https://review.openstack.org/559779 | 02:35 |
*** jcoufal has quit IRC | 02:35 | |
*** lbragstad has quit IRC | 02:39 | |
*** masber has quit IRC | 02:49 | |
*** baoli has quit IRC | 02:51 | |
*** baoli has joined #openstack-ironic | 02:51 | |
*** masber has joined #openstack-ironic | 02:52 | |
openstackgerrit | Tuan Luong-Anh proposed openstack/ironic master: [WIP] Implement iRMC BIOS configuration https://review.openstack.org/534595 | 02:56 |
*** baoli has quit IRC | 02:56 | |
openstackgerrit | zenghui.shi proposed openstack/ironic master: BIOS Settings: add admin doc https://review.openstack.org/568504 | 02:56 |
openstackgerrit | Tuan Luong-Anh proposed openstack/ironic master: [WIP] Implement iRMC BIOS configuration https://review.openstack.org/534595 | 03:03 |
*** akhilaki has quit IRC | 03:07 | |
*** jcoufal has joined #openstack-ironic | 03:13 | |
*** cjloader has joined #openstack-ironic | 03:19 | |
*** baoli has joined #openstack-ironic | 03:22 | |
*** cjloader has quit IRC | 03:24 | |
*** jcoufal has quit IRC | 03:30 | |
*** baoli has quit IRC | 03:30 | |
*** cjloader has joined #openstack-ironic | 03:46 | |
*** masuberu has joined #openstack-ironic | 03:56 | |
*** gyankum has joined #openstack-ironic | 03:59 | |
*** masber has quit IRC | 04:00 | |
*** zhangfei has quit IRC | 04:01 | |
*** zhangfei has joined #openstack-ironic | 04:02 | |
*** masber has joined #openstack-ironic | 04:02 | |
*** masuberu has quit IRC | 04:05 | |
*** links has joined #openstack-ironic | 04:24 | |
*** gyee has quit IRC | 04:26 | |
openstackgerrit | Merged openstack/ironic master: Remove excessive usage of mock_the_extension_manager in unit tests - part 1 https://review.openstack.org/568585 | 04:28 |
*** jesusaur has quit IRC | 04:39 | |
*** jesusaur has joined #openstack-ironic | 04:42 | |
*** cjloader has quit IRC | 04:47 | |
*** cjloader has joined #openstack-ironic | 04:48 | |
*** cjloader has quit IRC | 04:53 | |
*** marios has joined #openstack-ironic | 05:05 | |
*** jtomasek has quit IRC | 05:36 | |
*** liuzz has quit IRC | 05:47 | |
*** liuzz has joined #openstack-ironic | 05:48 | |
*** mjura has joined #openstack-ironic | 05:48 | |
*** penick has quit IRC | 05:53 | |
*** penick has joined #openstack-ironic | 05:58 | |
*** jaganathan has joined #openstack-ironic | 06:15 | |
*** liuzz_ has joined #openstack-ironic | 06:15 | |
*** marios has quit IRC | 06:16 | |
*** liuzz has quit IRC | 06:16 | |
*** rbartal has joined #openstack-ironic | 06:17 | |
*** marios has joined #openstack-ironic | 06:17 | |
*** liuzz_ has quit IRC | 06:20 | |
*** liuzz has joined #openstack-ironic | 06:21 | |
*** liuzz has quit IRC | 06:23 | |
*** liuzz has joined #openstack-ironic | 06:25 | |
*** jaganathan has quit IRC | 06:28 | |
*** jaganathan has joined #openstack-ironic | 06:31 | |
*** liuzz_ has joined #openstack-ironic | 06:32 | |
*** liuzz has quit IRC | 06:33 | |
*** Arkady has joined #openstack-ironic | 06:43 | |
*** rbudden has quit IRC | 06:46 | |
olivierb- | good morning everyone | 06:57 |
*** tesseract has joined #openstack-ironic | 07:01 | |
*** cjloader has joined #openstack-ironic | 07:18 | |
*** ajya has joined #openstack-ironic | 07:20 | |
*** cjloader has quit IRC | 07:23 | |
*** mjura has quit IRC | 07:38 | |
*** mjura has joined #openstack-ironic | 07:38 | |
*** AlexeyAbashkin has joined #openstack-ironic | 07:47 | |
*** racedo has joined #openstack-ironic | 07:51 | |
*** racedo has quit IRC | 07:51 | |
*** milan has joined #openstack-ironic | 08:04 | |
*** mgoddard has joined #openstack-ironic | 08:05 | |
*** rcernin has quit IRC | 08:12 | |
*** dougsz has joined #openstack-ironic | 08:14 | |
*** jtomasek has joined #openstack-ironic | 08:16 | |
*** cjloader has joined #openstack-ironic | 08:19 | |
*** cjloader has quit IRC | 08:23 | |
*** dtantsur|afk is now known as dtantsur | 08:26 | |
dtantsur | morning ironic | 08:26 |
*** zhangfei has quit IRC | 08:28 | |
etingof|afk | good [almost Friday] morning everyone o/ | 08:29 |
*** etingof|afk is now known as etingof | 08:30 | |
*** links has quit IRC | 08:32 | |
*** gyankum has quit IRC | 08:33 | |
*** jtomasek has quit IRC | 08:37 | |
*** e0ne has joined #openstack-ironic | 08:39 | |
*** jtomasek has joined #openstack-ironic | 08:39 | |
*** zhangfei has joined #openstack-ironic | 08:41 | |
*** MattMan has quit IRC | 08:42 | |
*** MattMan has joined #openstack-ironic | 08:42 | |
olivierb- | any of you having more clue today about the CI failures in ironic-lib ??? | 08:45 |
olivierb- | TheJulia made a tentative (partial?) patch with https://review.openstack.org/568828 but seems like it still fails | 08:45 |
patchbot | patch 568828 - ironic-lib - Slightly increase job timeouts | 08:45 |
*** jcoufal has joined #openstack-ironic | 08:45 | |
*** links has joined #openstack-ironic | 08:49 | |
*** jcoufal has quit IRC | 08:50 | |
*** d0ugal__ has quit IRC | 08:51 | |
*** d0ugal has joined #openstack-ironic | 08:51 | |
*** d0ugal has quit IRC | 08:51 | |
*** d0ugal has joined #openstack-ironic | 08:51 | |
*** derekh has joined #openstack-ironic | 08:51 | |
*** gyankum has joined #openstack-ironic | 09:04 | |
*** gyankum has quit IRC | 09:08 | |
openstackgerrit | zenghui.shi proposed openstack/ironic master: BIOS Settings: add admin doc https://review.openstack.org/568504 | 09:12 |
*** dims has quit IRC | 09:20 | |
*** rwsu has quit IRC | 09:28 | |
openstackgerrit | zenghui.shi proposed openstack/ironic master: Change exception msg of BIOS caching https://review.openstack.org/568193 | 09:43 |
*** gyankum has joined #openstack-ironic | 09:46 | |
*** gyankum has quit IRC | 09:50 | |
*** AlexeyAbashkin has quit IRC | 09:53 | |
*** gyankum has joined #openstack-ironic | 09:56 | |
openstackgerrit | yolanda.robla proposed openstack/ironic master: Add Node BIOS support - REST API https://review.openstack.org/512579 | 09:57 |
*** namnh has quit IRC | 10:00 | |
*** stephenbm has joined #openstack-ironic | 10:05 | |
*** stephenbm has quit IRC | 10:07 | |
*** dims has joined #openstack-ironic | 10:07 | |
*** dims has quit IRC | 10:17 | |
openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic stable/queens: Mark xclarity password as secret https://review.openstack.org/567639 | 10:22 |
*** hoonetorg has quit IRC | 10:43 | |
*** jcoufal has joined #openstack-ironic | 10:46 | |
*** jcoufal has quit IRC | 10:51 | |
*** rwsu has joined #openstack-ironic | 10:51 | |
*** AlexeyAbashkin has joined #openstack-ironic | 10:54 | |
openstackgerrit | Olivier Bourdon proposed openstack/ironic-python-agent master: Fix for fatal error when GPT was used and only MBR was cleaned https://review.openstack.org/569086 | 10:58 |
*** hoonetorg has joined #openstack-ironic | 11:00 | |
*** gyan_ has joined #openstack-ironic | 11:14 | |
*** jcoufal has joined #openstack-ironic | 11:15 | |
*** gyankum has quit IRC | 11:16 | |
*** links has quit IRC | 11:16 | |
*** cjloader has joined #openstack-ironic | 11:19 | |
openstackgerrit | Sam Betts proposed openstack/ironic-tempest-plugin master: Stop double json encoding the error responses https://review.openstack.org/569089 | 11:19 |
openstackgerrit | Sam Betts proposed openstack/ironic master: Stop double json encoding the error message responses https://review.openstack.org/567837 | 11:19 |
*** jcoufal has quit IRC | 11:20 | |
*** cjloader has quit IRC | 11:23 | |
*** rh-jelabarre has joined #openstack-ironic | 11:24 | |
*** phuongnh has quit IRC | 11:28 | |
*** links has joined #openstack-ironic | 11:29 | |
*** zhangfei has quit IRC | 11:32 | |
*** lbragstad has joined #openstack-ironic | 11:33 | |
*** UForgotten has quit IRC | 11:35 | |
*** UForgotten has joined #openstack-ironic | 11:35 | |
*** bfournie has quit IRC | 11:38 | |
*** tiendc has quit IRC | 11:44 | |
openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: OpenStack cloud driver added https://review.openstack.org/523879 | 11:49 |
*** dprince has joined #openstack-ironic | 11:56 | |
*** milan_ has joined #openstack-ironic | 11:57 | |
*** milan has quit IRC | 11:57 | |
*** liuzz_ has quit IRC | 11:57 | |
*** milan_ is now known as milan | 11:59 | |
*** trown|outtypewww is now known as trown | 12:03 | |
*** jcoufal has joined #openstack-ironic | 12:07 | |
*** bfournie has joined #openstack-ironic | 12:21 | |
olivierb- | Do you think I need to adjust some timeouts to have https://review.openstack.org/#/c/569086/ pass CI. because in the log trace I can see: | 12:31 |
patchbot | patch 569086 - ironic-python-agent - Fix for fatal error when GPT was used and only MBR... | 12:31 |
olivierb- | this takes 10 - 15 minutes (logs in logs/devstacklog.txt.gz) | 12:31 |
olivierb- | however next trace TS is +29mn after so I am suspecting a timeout furthermore as in the other logs I see broken pipe. | 12:31 |
olivierb- | however as this is the 1st time I am dealing with these tests I might very well be wrong | 12:32 |
*** sapd_ has quit IRC | 12:34 | |
*** sapd_ has joined #openstack-ironic | 12:34 | |
*** sapd_ has quit IRC | 12:35 | |
*** sapd_ has joined #openstack-ironic | 12:35 | |
*** sapd_ has quit IRC | 12:36 | |
*** sapd_ has joined #openstack-ironic | 12:36 | |
*** rbudden has joined #openstack-ironic | 12:43 | |
*** dims has joined #openstack-ironic | 12:43 | |
*** Goneri has joined #openstack-ironic | 12:46 | |
*** gyan_ has quit IRC | 12:54 | |
*** rloo has joined #openstack-ironic | 12:55 | |
vdrok | olivierb-: this info is just out of date, it is normal for devstack setup take around 30 min I think | 12:55 |
vdrok | also morning | 12:55 |
olivierb- | morning vdrok | 12:56 |
vdrok | olivierb-: the actual error seems to be http://logs.openstack.org/86/569086/1/check/ipa-tempest-dsvm-wholedisk-bios-pxe_ipmitool-tinyipa-src/54cfe1f/logs/devstacklog.txt.gz#_2018-05-17_11_33_14_543 | 12:57 |
vdrok | seems like some kind of problem with scsi package in tinycore | 12:57 |
vdrok | ah, it's just download time out | 12:57 |
jroll | morning everyone | 12:58 |
vdrok | morning jroll | 12:58 |
olivierb- | and thx for the info, I now see what you mean | 12:58 |
olivierb- | I'll launch rechech then | 12:58 |
vdrok | yup | 12:58 |
olivierb- | many thx again | 12:59 |
vdrok | no problem, you're welcome | 12:59 |
openstackgerrit | Kaifeng Wang proposed openstack/ironic master: Power fault recovery: db and rpc implementation https://review.openstack.org/555708 | 13:00 |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: Remove excessive usage of mock_the_extension_manager in unit tests - part 2 https://review.openstack.org/569105 | 13:08 |
*** links has quit IRC | 13:09 | |
TheJulia | good morning | 13:18 |
openstackgerrit | Jim Rollenhagen proposed openstack/ironic master: Remove endpoint_type from configuration https://review.openstack.org/567628 | 13:21 |
jroll | dtantsur: fixed those nits for you, thanks :) | 13:21 |
dtantsur | thnx! | 13:21 |
jroll | morning TheJulia | 13:21 |
rloo | good morning everyone, jroll, dtantsur, TheJulia, vdrok, olivierb- | 13:22 |
dtantsur | morning rloo, jroll, TheJulia | 13:22 |
vdrok | Morning dtantsur TheJulia and rloo ! | 13:24 |
dtantsur | hey vdrok | 13:24 |
olivierb- | hello rloo | 13:25 |
* rloo such a wonderful group of people to work with :) | 13:25 | |
olivierb- | yep, very helpful and so kind with newbies | 13:26 |
*** baoli has joined #openstack-ironic | 13:28 | |
dtantsur | ugh, is ironic removing neutron ports on undeploy? Oo | 13:29 |
*** sambetts|afk is now known as sambetts | 13:29 | |
sambetts | dtantsur: it definatly should not be | 13:30 |
sambetts | dtantsur: in an old version it used too | 13:30 |
dtantsur | well, it's master... | 13:30 |
dtantsur | and I see a port disappearing, I cannot get where | 13:30 |
sambetts | dtantsur: nova will delete the port if you have deployed only giving it a network | 13:31 |
dtantsur | sambetts: there's no nova.. | 13:31 |
jroll | sambetts: I think you have it backward, we used to not do it, now we do | 13:31 |
sambetts | jroll: not the neutron port, we remove the vif from ironic | 13:31 |
dtantsur | jroll: VIFs attachment or port themselves? | 13:31 |
jroll | oh yes, only vifs | 13:31 |
*** baoli has quit IRC | 13:32 | |
jroll | it's deleting the neutron port? O_o | 13:32 |
sambetts | only time we should delete a neutron port is during cleaning or provisioning tear down | 13:32 |
*** baoli has joined #openstack-ironic | 13:32 | |
sambetts | and we only delete it in provisioning in the neutron driver | 13:32 |
dtantsur | welllllll | 13:33 |
jroll | https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/network/neutron.py#L104 | 13:33 |
jroll | we do delete them | 13:33 |
sambetts | jroll: only for provisioning like I said | 13:33 |
sambetts | jroll: we never delete a tenant port | 13:33 |
jroll | oh, oops | 13:34 |
dtantsur | well, here we go http://logs.openstack.org/70/569070/4/check/metalsmith-integration-glance/bf54a52/controller/logs/screen-ir-cond.txt.gz#_May_17_13_19_14_679344 | 13:34 |
jroll | unless provisioning net and tenant net are the same :) | 13:34 |
dtantsur | this is flat networking (I hope) | 13:34 |
sambetts | jroll: thats why this does nothing: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/network/flat.py#L91 | 13:34 |
dtantsur | but we delete the port | 13:34 |
* dtantsur wonders who is crazy: ironic, himself or both? | 13:35 | |
jroll | sambetts: right, I was implying neutron network driver but using the same networks (which is a mistake, but a valid mistake) | 13:35 |
sambetts | jroll: that will never work because we try to create a port in neutron in that network but there is already one so you get a mac address conflict | 13:35 |
jroll | fair | 13:36 |
sambetts | so tenant can't == provisioning | 13:36 |
dtantsur | folks, we're deleting ports in *flat* networking, which I don't think is correct.. | 13:36 |
jroll | dtantsur: yes, understand, looking | 13:37 |
sambetts | dtantsur: we only remove the neutron port we create in neutron | 13:37 |
jroll | I was explaining my comment before you said it was flat | 13:37 |
sambetts | we create one for cleaning, and then delete it | 13:37 |
dtantsur | sambetts: this is not what I see | 13:37 |
sambetts | that log you sent the node has just finished cleaning | 13:37 |
dtantsur | sambetts: nope, it's starting cleaning | 13:38 |
dtantsur | (and I know this port's UUID) | 13:38 |
sambetts | oooooo I know what happening! | 13:38 |
TheJulia | vif teardown | 13:38 |
sambetts | nope | 13:38 |
sambetts | I think add_cleaning_network removes all ports from the network first to prevent mac address conflcits | 13:39 |
sambetts | when we create a port we own | 13:39 |
sambetts | because there shouldn't be a tenant port at cleaning time | 13:39 |
sambetts | in the cleaning network | 13:39 |
TheJulia | Yeah, there is that | 13:39 |
*** sdake_ is now known as sdake | 13:39 | |
dtantsur | it means, we delete tenant ports before cleaning? | 13:39 |
sambetts | its this line: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/network/flat.py#L115-L116 | 13:40 |
sambetts | it ensures we can proceed through cleaning | 13:41 |
sambetts | but results in the tenant port getting deleted if its left in the cleaning network | 13:41 |
sambetts | without that we'll fail with a mac address conflict when attempting to add a cleaning port | 13:42 |
* dtantsur also screwed up something around creating the port apparently: http://logs.openstack.org/70/569070/4/check/metalsmith-integration-glance/bf54a52/controller/logs/screen-neutron-api.txt.gz#_May_17_13_15_49_956907 | 13:42 | |
jroll | sambetts: yep, that's it | 13:43 |
jroll | possible solution is to check if there's a port, update it if needed | 13:43 |
sambetts | dtantsur: that error is expected if your using devstack and hacking baremetal into a vxlan network | 13:43 |
jroll | could probably do that for all of flat networking tbh | 13:43 |
TheJulia | dtantsur: vlan instead of vxlan maybe? | 13:43 |
sambetts | dtantsur: and not running a real flat network | 13:44 |
dtantsur | sambetts: I guess I should modify some devstack setting? or just set --provider stuff when creating a port? | 13:44 |
dtantsur | (this is re vxlan, dunno what to do about deletion) | 13:44 |
sambetts | dtantsur: its the network type not the port type | 13:44 |
sambetts | dtantsur: the network type needs to be "flat" but in ironic devstack stuff we cxreate a vxlan network and then hack our way into it | 13:44 |
sambetts | dtantsur: resulting in weird errors becuase the network type is wrong | 13:45 |
dtantsur | sambetts: oh, so it's normal for the default devstack? | 13:45 |
sambetts | dtantsur: yup | 13:45 |
dtantsur | ok, thanks! one problem less :) | 13:45 |
dtantsur | now, the port deletion... I guess we don't have a CI job using a pre-created port? | 13:45 |
sambetts | I had many patches to resolve that and change the network to a real network, but they became inpossible to maintain | 13:45 |
sambetts | dtantsur: we don't have on that reuses a port no | 13:46 |
dtantsur | nice, so I'm the first to test it in the CI :D | 13:46 |
sambetts | dtantsur: I was working on a fix for this though | 13:46 |
sambetts | the deletion thing | 13:46 |
dtantsur | https://review.openstack.org/#/c/569070/4/playbooks/integration/exercise.yaml | 13:46 |
patchbot | patch 569070 - metalsmith - Allow attaching existing ports | 13:46 |
dtantsur | (the last task fails with port not found) | 13:46 |
sambetts | dtantsur: https://review.openstack.org/#/c/539981/ | 13:46 |
patchbot | patch 539981 - ironic - Ensure we unbind flat network ports and clear BM m... | 13:46 |
sambetts | dtantsur: basically instead of deleting the ports we just need to reset their mac addresses from the baremetal mac address so they aren't associated any more | 13:47 |
sambetts | dtantsur: because otherwise if you use pre-created ports, you run into mac address conflicts the next time you deploy the same node with a different port | 13:48 |
dtantsur | sambetts: need help with that patch? I can throw a depend-on | 13:48 |
TheJulia | sambetts: I had been thinkign about that with recent issues with heat/nova that I 've been chasing down, but I'm worried that we may confuse SDN integration by doing updates that change the mac and that port could have been wired elsewhere... my worry really is about long term port re-use | 13:48 |
sambetts | dtantsur: Yeah, any help would be good to get this resolved | 13:49 |
sambetts | TheJulia: thats why we need to reset the mac when we unbind, because the port is no longer associated with the baremetal | 13:49 |
dtantsur | TheJulia: removing the pre-created port is also confusing.. | 13:49 |
* TheJulia is on a call at the moment | 13:50 | |
sambetts | but currently we leave the mac address behind which still ties the port to the baremetal node, even though its not bound for that node any more | 13:50 |
dtantsur | sambetts: do you have a bug number for all this? | 13:50 |
sambetts | I thought I created one but I don't see it on the gerrit patch :( | 13:51 |
* sambetts fails at searching storyboard for it | 13:52 | |
dtantsur | sambetts: I guess we can remove https://review.openstack.org/#/c/539981/4/ironic/drivers/modules/network/flat.py@138 now, right? | 13:52 |
patchbot | patch 539981 - ironic - Ensure we unbind flat network ports and clear BM m... | 13:52 |
dtantsur | hmm, and https://review.openstack.org/#/c/539981/4/ironic/drivers/modules/network/flat.py@159 will probably try the same... | 13:53 |
patchbot | patch 539981 - ironic - Ensure we unbind flat network ports and clear BM m... | 13:53 |
sambetts | dtantsur: with this change unconfigure tenant networks should reset the mac address of the tenant port, but I think the rollback is there for the case of retrying cleaning too | 13:54 |
TheJulia | sambetts: We would need community consensus w/r/t ml2 driver owners to ensure that it is sane to them because they might be using the mac address to track the ports existence | 13:54 |
sambetts | its just an accident that the tenant network is getting rolled into it | 13:54 |
dtantsur | sambetts: then we need to record which ports we created | 13:54 |
sambetts | TheJulia: at this point the port is unbound, the ml2 drivers have no idea it even exists | 13:54 |
sambetts | dtantsur: we have that already, cleaning_vif_id, so yeah rollback should only deal with that | 13:55 |
dtantsur | TheJulia: I think the problem with user experience is worse, to be honest. at least assuming that nova does not delete ports on server deletion | 13:55 |
dtantsur | right | 13:55 |
TheJulia | I'm not 100% sure about that | 13:55 |
TheJulia | dtantsur: nova delete the port if you ask for a network attachment because it creates the port | 13:55 |
TheJulia | otherwise it does not delete the port | 13:55 |
dtantsur | TheJulia: right, we're talking about the latter | 13:55 |
TheJulia | but the port is functionally unusable | 13:56 |
TheJulia | which goes to sambetts's point that we shoudl be resetting the mac back | 13:56 |
dtantsur | yep | 13:56 |
TheJulia | which might work just fine, I just have concerns | 13:56 |
*** zhangfei has joined #openstack-ironic | 13:57 | |
sambetts | we clear the binding:host_id and binding:profile, that defines a port as unbound, and allows us to change the mac address, neutron disallows changing the mac address of a bound port specifially because of your concern | 13:58 |
TheJulia | okay, whew | 13:59 |
sambetts | because if its bound then it might be being used for something, once its unbound its just a blob of data | 13:59 |
*** ianychoi_ has quit IRC | 13:59 | |
TheJulia | in theory, I feel like at least one of the ml2 drivers I've looked at does some wonky things still tracking the blobs of data | 14:00 |
sambetts | TheJulia: the only thing that could go wrong with my code in this case is that we unluckly manage to generate a random mac address that already exists in the network, but I think thats a very small chance | 14:01 |
sambetts | and we could wrap it in a retry | 14:01 |
sambetts | the code I use to generate the mac is taken straight from neutron too, so it uses the exact same scheme | 14:02 |
TheJulia | I wonder if we could store the original mac | 14:02 |
sambetts | TheJulia: neutron might regenerate that mac though | 14:02 |
TheJulia | true | 14:02 |
sambetts | TheJulia: so its actually safer to generate a new one | 14:02 |
TheJulia | yeah | 14:02 |
TheJulia | is it just unsetting the mac fhat forces it to regenerate? | 14:03 |
sambetts | neutron doesn't let you unset the mac, so we have to generate one on our end | 14:03 |
sambetts | https://review.openstack.org/#/c/539981/4/ironic/common/neutron.py @ L93 | 14:03 |
patchbot | patch 539981 - ironic - Ensure we unbind flat network ports and clear BM m... | 14:03 |
sambetts | I was hoping you could just set it to null/none and it would generate a new one neutron side | 14:04 |
sambetts | but alas nope | 14:04 |
TheJulia | ugh, fun | 14:04 |
sambetts | yeah | 14:04 |
sambetts | this is one of those patches I was working on and kinda slipped into the background because of other things :( but this bug caused me and massive headache in a system a few months ago that were trying to use pre-existing neutron ports | 14:06 |
dtantsur | ditto :) | 14:07 |
TheJulia | Yeah, I just got off a call where they were hitting this because of heat orphaned ports :( | 14:08 |
sambetts | well at least the code is already written, even if it got forgotten for a bit | 14:09 |
dtantsur | I can try resurrecting it | 14:09 |
dtantsur | probably tomorrow with a fresher head | 14:09 |
TheJulia | doesn't neutron have a range it generates macs in? | 14:10 |
sambetts | TheJulia: yup, I basically stole the code striaght from neutron, so we generate the macs in exactly the same way | 14:10 |
*** skazi has quit IRC | 14:10 | |
TheJulia | ok | 14:10 |
sambetts | TheJulia: and its configurable in the same way | 14:10 |
TheJulia | okay, yeah, see the base mac now | 14:11 |
sambetts | I believe all thats missing from that patch is UTs | 14:11 |
jroll | so we used to generate macs in exactly the same way until the copy/paste evolved on both ends* | 14:11 |
jroll | :( | 14:11 |
sambetts | jroll: yeah thats the only thing thats a pain ... | 14:12 |
* TheJulia goes and gets coffee | 14:12 | |
jroll | sambetts: it's not a pain, it's bad practice, why not pull that out into neutron-lib or something? | 14:12 |
jroll | probably a separate library actually | 14:12 |
jroll | idk | 14:13 |
jroll | any reason we can't just null the mac out? | 14:13 |
sambetts | actually its already in neutron-lib https://github.com/openstack/neutron-lib/blob/master/neutron_lib/utils/net.py#L28 | 14:13 |
sambetts | jroll: it won't let us :( that was my original goal | 14:14 |
jroll | okay, why not import neutron-lib? :) | 14:14 |
sambetts | jroll: new dep for ironic | 14:14 |
jroll | yes, and? | 14:14 |
jroll | are dependencies bad again? | 14:14 |
jroll | better question, are dependencies worse than copy/paste? | 14:14 |
openstackgerrit | Ilya Etingof proposed openstack/sushy-tools master: OpenStack cloud driver added https://review.openstack.org/523879 | 14:14 |
jroll | it doesn't require much that ironic doesn't have already: https://github.com/openstack/neutron-lib/blob/master/requirements.txt | 14:15 |
sambetts | tbh I wasn't even sure this patch would fly, so yeah totally can use neutron-lib to avoid the copy/paste | 14:15 |
TheJulia | +1 to using neutron-lib and not copy/paste | 14:16 |
sambetts | cool left a comment to that affect on the patch | 14:17 |
jroll | thanks | 14:18 |
*** ianychoi has joined #openstack-ironic | 14:18 | |
*** baha has joined #openstack-ironic | 14:20 | |
*** hemna_ has joined #openstack-ironic | 14:24 | |
*** hamzy_ has joined #openstack-ironic | 14:28 | |
*** hamzy has quit IRC | 14:29 | |
openstackgerrit | Sam Betts proposed openstack/ironic-tempest-plugin master: Stop double json encoding the error responses https://review.openstack.org/569089 | 14:32 |
*** jaypipes has quit IRC | 14:34 | |
openstackgerrit | Mark Hamzy proposed openstack/bifrost master: pip version 10.0.0b1 will not uninstall system packages https://review.openstack.org/568719 | 14:43 |
*** hamzy_ is now known as hamzy | 14:43 | |
olivierb- | TheJulia once back from coffee could you please have a brief look at the very short fix I made concerning what we talked about yesterday and is resumed in commit message of https://review.openstack.org/569086, many thanks. I have successfully tested this several times today on my test environments. Comments from others also very welcomed | 14:44 |
patchbot | patch 569086 - ironic-python-agent - Fix for fatal error when GPT was used and only MBR... | 14:44 |
TheJulia | olivierb-: in the middle of helping someone with a deployment issue in their system, when I circle back from that | 14:54 |
*** cjloader has joined #openstack-ironic | 14:54 | |
olivierb- | no pb at all, take your time TheJulia I am in absolute no hurry on my side as I already patched my env | 14:55 |
*** mjura has quit IRC | 15:01 | |
*** rbartal has quit IRC | 15:07 | |
*** fragatin_ has quit IRC | 15:08 | |
*** fragatina has joined #openstack-ironic | 15:10 | |
*** AlexeyAbashkin has quit IRC | 15:12 | |
*** AlexeyAbashkin has joined #openstack-ironic | 15:15 | |
dtantsur | one more big boring but necessary patch: https://review.openstack.org/569105 | 15:19 |
patchbot | patch 569105 - ironic - Remove excessive usage of mock_the_extension_manag... | 15:19 |
*** dprince has quit IRC | 15:24 | |
*** cz2 has quit IRC | 15:24 | |
*** zhangfei has quit IRC | 15:25 | |
prometheanfire | https://paste.pound-python.org/show/roqd3nvt1QhalGZUeKpa/ helps, but still getting what looks like a unicode error | 15:35 |
prometheanfire | https://gist.githubusercontent.com/prometheanfire/d4bc95be41d401471b040e72ab82d7e7/raw/d6aff3e1f00c80b5642bfc379d998162d7d59709/gistfile1.txt | 15:35 |
*** gyee has joined #openstack-ironic | 15:41 | |
*** dprince has joined #openstack-ironic | 15:43 | |
*** marios has quit IRC | 15:45 | |
*** lucasagomes is now known as lucas-afk | 15:48 | |
openstackgerrit | Kaifeng Wang proposed openstack/ironic master: Power fault recovery: apply fault https://review.openstack.org/556758 | 15:49 |
*** e0ne has quit IRC | 15:51 | |
*** cz2 has joined #openstack-ironic | 15:52 | |
*** baha has quit IRC | 15:53 | |
rloo | TheJulia: if you have a few min to hopefully put your mark so we can proceed: https://review.openstack.org/#/c/568627/ | 15:56 |
patchbot | patch 568627 - ironic-specs - Use node.fault field for power fault recovery work | 15:56 |
*** milan has quit IRC | 15:59 | |
TheJulia | as long as my mark is not signing away my soul to sail on a ship for the rest of time | 16:01 |
*** trown is now known as trown|lunch | 16:02 | |
TheJulia | rloo: did you see Kaifeng's most recent comment? | 16:04 |
rloo | TheJulia: no, i didn't. I don't think we should change it to a json field; we haven't spent time thinking about what it means to support more than one fault at a time. | 16:05 |
TheJulia | That too is reasonable :) | 16:06 |
rloo | TheJulia: we could add new field .faults for that. and/or look at the orig spec that proposes a more robust solution. | 16:06 |
rloo | TheJulia: i will respond; thx for pointing it out! | 16:06 |
TheJulia | yeah, and that actually may be a better path, that way we have a clean migration there if we ever get there | 16:06 |
TheJulia | rloo: no problem, let me know when you want me to get out the +A stamp | 16:07 |
rloo | TheJulia: oh, you can do the +A now :) | 16:07 |
rloo | TheJulia: thx! | 16:07 |
openstackgerrit | Matthew Thode proposed openstack/ironic-lib master: Fall back to PARTUUID if UUID returns nothing. https://review.openstack.org/569158 | 16:09 |
TheJulia | Stamped | 16:09 |
openstackgerrit | Matthew Thode proposed openstack/ironic-python-agent master: write byte objects when using os.write https://review.openstack.org/569159 | 16:12 |
TheJulia | olivierb-: your ipa change lgtm, super simple, although I did have to go double check the manual for sgdisk | 16:15 |
prometheanfire | mind reviewing https://review.openstack.org/569160 as well? | 16:16 |
patchbot | patch 569160 - diskimage-builder - IPA requires iptables | 16:16 |
openstackgerrit | Merged openstack/ironic-specs master: Use node.fault field for power fault recovery work https://review.openstack.org/568627 | 16:19 |
TheJulia | prometheanfire: sure in a little bit | 16:20 |
prometheanfire | yes | 16:20 |
*** baha has joined #openstack-ironic | 16:20 | |
TheJulia | oh jeeze | 16:21 |
TheJulia | done btw :) | 16:21 |
prometheanfire | ? | 16:21 |
TheJulia | iptables missing | 16:21 |
prometheanfire | :D | 16:22 |
*** jaganathan has quit IRC | 16:23 | |
*** jaganathan has joined #openstack-ironic | 16:23 | |
*** jrist has quit IRC | 16:24 | |
TheJulia | stepping away for realz for a little while | 16:25 |
*** Arkady has quit IRC | 16:27 | |
* sambetts --> home | 16:33 | |
sambetts | night all | 16:33 |
*** sambetts is now known as sambetts|afk | 16:33 | |
*** mjturek has joined #openstack-ironic | 16:38 | |
*** tesseract has quit IRC | 16:38 | |
openstackgerrit | Ruby Loo proposed openstack/ironic master: Disable periodic tasks if interval set to 0 https://review.openstack.org/569166 | 16:46 |
openstackgerrit | Dmitry Tantsur proposed openstack/ironic master: [WIP] Convert test_manager to hardware types https://review.openstack.org/569167 | 16:47 |
*** trown|lunch is now known as trown|mtg | 17:02 | |
openstackgerrit | Kaifeng Wang proposed openstack/ironic master: Power fault recovery: API implementation https://review.openstack.org/556015 | 17:03 |
dtantsur | g'night | 17:04 |
*** AlexeyAbashkin has quit IRC | 17:04 | |
*** dtantsur is now known as dtantsur|afk | 17:04 | |
*** dougsz has quit IRC | 17:04 | |
*** derekh has quit IRC | 17:05 | |
*** mgoddard has quit IRC | 17:06 | |
*** jtomasek has quit IRC | 17:07 | |
anupn | Hi TheJulia: I was able to provision the machine correctly. It has to have separate provisioning network specified and need OVS_BRIDGE_MAPPINGS to be specified. That way br-int and the ovs bridge connected to the physical network interface can talk to each other | 17:36 |
*** akhilaki has joined #openstack-ironic | 17:38 | |
TheJulia | \o/ | 17:42 |
*** baha has quit IRC | 17:44 | |
*** retr0h has quit IRC | 17:44 | |
*** retr0h has joined #openstack-ironic | 17:44 | |
*** baha has joined #openstack-ironic | 17:46 | |
*** mgoddard has joined #openstack-ironic | 17:46 | |
*** dprince has quit IRC | 17:56 | |
*** links has joined #openstack-ironic | 17:59 | |
*** cjloader_ has joined #openstack-ironic | 18:07 | |
*** cjloader has quit IRC | 18:07 | |
*** cjloader_ has quit IRC | 18:07 | |
*** cjloader has joined #openstack-ironic | 18:07 | |
*** trown|mtg is now known as trown | 18:22 | |
openstackgerrit | Matthew Thode proposed openstack/ironic-lib master: Fall back to PARTUUID if UUID returns nothing. https://review.openstack.org/569158 | 18:32 |
*** ElCoyote_ has joined #openstack-ironic | 18:33 | |
*** cjloader has quit IRC | 18:34 | |
*** cjloader has joined #openstack-ironic | 18:35 | |
*** cjloader has quit IRC | 18:35 | |
*** links has quit IRC | 18:35 | |
*** cjloader has joined #openstack-ironic | 18:36 | |
TheJulia | prometheanfire: I think it should... | 19:13 |
prometheanfire | think what should? | 19:14 |
*** mjturek has quit IRC | 19:25 | |
*** dtantsur|afk has quit IRC | 19:29 | |
TheJulia | the patch, that it should work. Unit tests likely won't like it much :) | 19:30 |
*** dtantsur has joined #openstack-ironic | 19:36 | |
*** jrist has joined #openstack-ironic | 19:36 | |
*** ajya has quit IRC | 19:37 | |
*** ajya has joined #openstack-ironic | 19:37 | |
openstackgerrit | Merged openstack/ironic-python-agent master: rework ATA secure erase https://review.openstack.org/557755 | 19:39 |
*** dtantsur has quit IRC | 19:40 | |
*** jrist has quit IRC | 19:48 | |
*** dtantsur has joined #openstack-ironic | 19:51 | |
*** jrist has joined #openstack-ironic | 19:51 | |
*** zshi has quit IRC | 19:54 | |
*** zshi has joined #openstack-ironic | 19:54 | |
prometheanfire | TheJulia: ah, question, should this image work with ironic-python-agent? | 19:55 |
prometheanfire | disk-image-create -o baremetal-$DISTRO_NAME-$DIB_RELEASE ubuntu baremetal block-device-gpt bootloader dhcp-all-interfaces local-config slow-network devuser | 19:56 |
TheJulia | prometheanfire: I believe so yes | 19:56 |
prometheanfire | ok, I'm getting this, it's with my patches https://paste.pound-python.org/show/MwRhLPfXUrChtIiKUp8t/ | 19:58 |
TheJulia | that is... bizzar | 20:08 |
prometheanfire | yep | 20:12 |
prometheanfire | trying with bionic instead of fedora next | 20:12 |
*** ajya has quit IRC | 20:13 | |
prometheanfire | or not, dib doesn't support it yet | 20:15 |
prometheanfire | what do we do with stuff like this? do we want UUID or PARTUUID? | 20:19 |
prometheanfire | KNAME="nvme0n1p1" UUID="F0CC-7859" PARTUUID="69497f55-a6e8-46ff-8c30-eb01a2ab9854" TYPE="part" | 20:19 |
TheJulia | I think we want part uuid, but it looks like it is not reading the underlying uuid correctly | 20:20 |
prometheanfire | ya, IPA needs a change for PARTUUID as well | 20:20 |
prometheanfire | the log showed it mapping the UUID to /dev/sda2, wasn't the right? | 20:20 |
TheJulia | depends on the image | 20:21 |
prometheanfire | so... I'll check for PARTUUID and use that, if that doesn't exist try UUID | 20:21 |
TheJulia | what it thought was correct matched what it wrote I beleive | 20:21 |
prometheanfire | rather than the other way around | 20:21 |
TheJulia | okay | 20:24 |
openstackgerrit | Matthew Thode proposed openstack/ironic-python-agent master: fall back to PARTUUID if UUID not found. https://review.openstack.org/569250 | 20:25 |
prometheanfire | though for ipa the order doesn't mater, just the matching | 20:25 |
*** fragatina has quit IRC | 20:32 | |
*** dprince has joined #openstack-ironic | 20:39 | |
prometheanfire | I'm surprised no one else ran into this | 20:40 |
prometheanfire | makes me think I'm doing something wrong | 20:40 |
*** jcoufal has quit IRC | 20:50 | |
TheJulia | could be your hardware or kernel... | 20:51 |
*** ElCoyote_ has quit IRC | 20:51 | |
*** fragatina has joined #openstack-ironic | 20:59 | |
prometheanfire | just dell and hp servers | 21:01 |
*** fragatina has quit IRC | 21:01 | |
*** fragatina has joined #openstack-ironic | 21:02 | |
*** dtantsur has quit IRC | 21:02 | |
*** milan has joined #openstack-ironic | 21:02 | |
*** dtantsur has joined #openstack-ironic | 21:03 | |
*** baha has quit IRC | 21:03 | |
*** fragatina has quit IRC | 21:03 | |
*** fragatina has joined #openstack-ironic | 21:04 | |
*** ElCoyote_ has joined #openstack-ironic | 21:08 | |
*** Goneri has quit IRC | 21:08 | |
*** trown is now known as trown|outtypewww | 21:09 | |
*** rh-jelabarre has quit IRC | 21:15 | |
*** rh-jelabarre has joined #openstack-ironic | 21:15 | |
NobodyCam | Good Afternoon Ironic'ers | 21:27 |
*** cjloader_ has joined #openstack-ironic | 21:28 | |
*** cjloader has quit IRC | 21:28 | |
*** dprince has quit IRC | 21:28 | |
*** bfournie has quit IRC | 21:30 | |
*** bfournie has joined #openstack-ironic | 21:31 | |
*** bfournie has quit IRC | 21:31 | |
*** cjloader_ has quit IRC | 21:33 | |
*** hemna_ has quit IRC | 21:34 | |
*** fragatina has quit IRC | 21:39 | |
*** mgoddard has quit IRC | 21:40 | |
*** fragatina has joined #openstack-ironic | 21:46 | |
*** rcernin has joined #openstack-ironic | 22:04 | |
*** hemna_ has joined #openstack-ironic | 22:05 | |
*** baoli has quit IRC | 22:09 | |
*** baoli has joined #openstack-ironic | 22:10 | |
*** baoli has quit IRC | 22:14 | |
*** milan has quit IRC | 22:15 | |
*** hemna_ has quit IRC | 22:24 | |
*** hemna_ has joined #openstack-ironic | 22:24 | |
*** cjloader has joined #openstack-ironic | 22:26 | |
*** cjloader has quit IRC | 22:28 | |
*** cjloader has joined #openstack-ironic | 22:28 | |
*** ElCoyote_ has quit IRC | 22:33 | |
*** threestrands has joined #openstack-ironic | 22:37 | |
*** rpioso|afk is now known as rpioso | 22:41 | |
rpioso | TheJulia: I have a favor to ask. Do you have the link to Jimmy McArthur’s announcement about the revised early PTG registration deadline handy? | 22:45 |
rpioso | TheJulia: I don’t recall which list he sent it to. | 22:46 |
TheJulia | Not readily, did you check the openstack-dev archive? | 22:46 |
rpioso | TheJulua: y | 22:46 |
TheJulia | Hmmmm | 22:47 |
rpioso | TheJulia: Oops | 22:47 |
rpioso | TheJulia: He sent it in April. | 22:47 |
*** ElCoyote_ has joined #openstack-ironic | 22:51 | |
prometheanfire | is ironic tested with diskimage built images? | 22:54 |
* prometheanfire is really confused how anyone is using ironic with the current IPA issues, maybe they are using old DIB? | 22:57 | |
*** ElCoyote_ has quit IRC | 23:03 | |
*** rbudden has quit IRC | 23:03 | |
rpioso | TheJulia: Never mind. I found it. | 23:11 |
TheJulia | prometheanfire: bifrost has a job that does it. I can check it when i get back in front of a computer | 23:12 |
prometheanfire | TheJulia: thanks, atm I just want to find something that works, just to compare the current deploy to it | 23:14 |
TheJulia | prometheanfire: http://logs.openstack.org/19/568719/4/check/bifrost-integration-dibipa-debian/d1a49a5/ | 23:16 |
prometheanfire | ok, I'll work with that, thanks | 23:24 |
*** baoli has joined #openstack-ironic | 23:25 | |
*** baoli has quit IRC | 23:25 | |
*** cjloader has quit IRC | 23:29 | |
*** bfournie has joined #openstack-ironic | 23:35 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!