| *** zhenguo has joined #openstack-ironic | 00:01 | |
| *** achanda has quit IRC | 00:05 | |
| *** jamielennox|away is now known as jamielennox | 00:06 | |
| *** davideagnello has joined #openstack-ironic | 00:06 | |
| *** davideagnello has quit IRC | 00:11 | |
| *** gabriel-bezerra has joined #openstack-ironic | 00:14 | |
| Haomeng | mrda: :) | 00:20 |
|---|---|---|
| openstackgerrit | Haomeng,Wang proposed openstack/python-ironicclient: (WIP)validate input uuid from client side https://review.openstack.org/172618 | 00:20 |
| *** naohirot has joined #openstack-ironic | 00:41 | |
| *** kkoski has joined #openstack-ironic | 01:02 | |
| *** igordcard_ has quit IRC | 01:04 | |
| *** kkoski has quit IRC | 01:06 | |
| *** achanda has joined #openstack-ironic | 01:21 | |
| *** Marga_ has quit IRC | 01:50 | |
| *** chenglch has joined #openstack-ironic | 02:22 | |
| *** chenglch has quit IRC | 02:36 | |
| *** chenglch has joined #openstack-ironic | 02:37 | |
| *** ricolin has quit IRC | 02:39 | |
| *** ramineni has joined #openstack-ironic | 02:40 | |
| *** Haomeng|2 has joined #openstack-ironic | 02:43 | |
| *** Haomeng has quit IRC | 02:43 | |
| *** achanda has quit IRC | 02:44 | |
| *** achanda has joined #openstack-ironic | 02:48 | |
| *** oomichi has quit IRC | 02:55 | |
| *** achanda has quit IRC | 02:59 | |
| *** lintan_ has joined #openstack-ironic | 03:05 | |
| *** Marga_ has joined #openstack-ironic | 03:11 | |
| *** Marga_ has quit IRC | 03:12 | |
| *** subscope_ has joined #openstack-ironic | 03:12 | |
| *** Marga_ has joined #openstack-ironic | 03:12 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: Improve AMT driver doc https://review.openstack.org/169606 | 03:20 |
| *** naohirot has quit IRC | 03:28 | |
| Haomeng|2 | mrda: are you around? | 03:40 |
| Haomeng|2 | mrda: do you know if we can change the exception string now, I understand it is in freezon status, right? | 03:41 |
| *** achanda has joined #openstack-ironic | 03:59 | |
| *** mdbooth has quit IRC | 04:01 | |
| *** naohirot has joined #openstack-ironic | 04:01 | |
| *** achanda has quit IRC | 04:05 | |
| *** lintan_ has quit IRC | 04:07 | |
| *** mdbooth has joined #openstack-ironic | 04:08 | |
| *** saripurigopi has joined #openstack-ironic | 04:17 | |
| mrda | Haomeng|2: Sorry for delay in getting back to you. I don't think we want to make any string changes now. | 04:17 |
| mrda | only bug fixes, and even then, no string changes. I think otherwise our i18n folk are too overworked. | 04:18 |
| *** lintan_ has joined #openstack-ironic | 04:23 | |
| Haomeng|2 | mrda: ok, thk | 04:27 |
| *** subscope_ has quit IRC | 04:43 | |
| *** chenglch has quit IRC | 05:07 | |
| openstackgerrit | Anusha Ramineni proposed stackforge/proliantutils: UPDATE_PERSISTENT_BOOT to 'NETWORK' fails on Gen9 https://review.openstack.org/172785 | 05:09 |
| saripurigopi | Is there a way to enroll ironic nodes from horizon ? | 05:10 |
| openstackgerrit | Anusha Ramineni proposed stackforge/proliantutils: UPDATE_PERSISTENT_BOOT to 'NETWORK' fails on Gen9 https://review.openstack.org/172785 | 05:15 |
| openstackgerrit | Anusha Ramineni proposed stackforge/proliantutils: UPDATE_PERSISTENT_BOOT to 'NETWORK' fails on Gen9 https://review.openstack.org/172785 | 05:22 |
| Haomeng|2 | saripurigopi: tripleo UI have such function, can you try with tuskar-ui? | 05:29 |
| saripurigopi | Haomengl2, sure will try with tuskar-ui. | 05:29 |
| *** pradipta has joined #openstack-ironic | 05:35 | |
| *** chenglch has joined #openstack-ironic | 05:35 | |
| saripurigopi | Looking at the ironic specs, kilo-archive is created for postponed BPs. Is it open to submit the review changes? | 05:49 |
| *** achanda has joined #openstack-ironic | 05:53 | |
| Haomeng|2 | saripurigopi: yes, we can create bp for next release now, welcome | 05:54 |
| *** yuanying has quit IRC | 06:04 | |
| *** yuanying has joined #openstack-ironic | 06:08 | |
| *** jcoufal has joined #openstack-ironic | 06:26 | |
| *** rwsu has joined #openstack-ironic | 06:34 | |
| saripurigopi | Haomengl2, do I need to submit the BP again or will it be automatically moved to Liberty? | 06:35 |
| Haomeng|2 | saripurigopi: no need to submit again, will not move to *L* release, so dont care | 06:41 |
| Haomeng|2 | saripurigopi: which is your bp patch? can you share the link? | 06:41 |
| saripurigopi | Haomengl2, okay. pxe_ucs driver - https://review.openstack.org/#/c/159734 | 06:43 |
| *** ukalifon1 has joined #openstack-ironic | 06:46 | |
| Haomeng|2 | saripurigopi: I understand this is the code patch, do you have blueprint patch/ | 06:49 |
| Haomeng|2 | saripurigopi: that shoud be blueprint's spec | 06:49 |
| saripurigopi | Haomengl2 , sorry , here is BP review - https://review.openstack.org/#/c/139517/ | 06:50 |
| Haomeng|2 | saripurigopi: oh, your bp is approved, so dont worry | 06:51 |
| Haomeng|2 | saripurigopi: once main branch is *L*, we can go ahead to do these patches | 06:52 |
| saripurigopi | Haomengl2: okay. | 06:52 |
| Haomeng|2 | saripurigopi: so no need to submit bp again:) | 06:52 |
| Haomeng|2 | saripurigopi: thanks for your bp:) | 06:52 |
| saripurigopi | Haomengl2: can I submit the code changes for review, or should I wait till main branch is "L" | 06:53 |
| *** yuanying has quit IRC | 06:54 | |
| Haomeng|2 | saripurigopi: I see deva's comments, so it is better we start to new patch during the second or third week of April:) | 07:07 |
| saripurigopi | Haomengl2: okay. :-) | 07:08 |
| Haomeng|2 | saripurigopi: and we can fix some Jenkins issue now, and run ut with latest code in our local dev env I think:) | 07:08 |
| Haomeng|2 | saripurigopi: good luck:) | 07:08 |
| Haomeng|2 | saripurigopi: :) | 07:08 |
| saripurigopi | Haomengl2: sure, thank you.(y) | 07:08 |
| *** jamielennox is now known as jamielennox|away | 07:11 | |
| *** yuanying has joined #openstack-ironic | 07:21 | |
| *** achanda has quit IRC | 07:21 | |
| *** achanda has joined #openstack-ironic | 07:22 | |
| *** jistr has joined #openstack-ironic | 07:24 | |
| *** chlong has quit IRC | 07:25 | |
| *** yuanying has quit IRC | 07:28 | |
| *** yuanying has joined #openstack-ironic | 07:31 | |
| *** mrda is now known as mrda-away | 07:32 | |
| *** Marga_ has quit IRC | 07:32 | |
| *** yuanying has quit IRC | 07:33 | |
| *** ifarkas has joined #openstack-ironic | 07:41 | |
| *** rsalevsky has joined #openstack-ironic | 07:43 | |
| *** yuanying has joined #openstack-ironic | 07:48 | |
| *** yuanying has quit IRC | 07:51 | |
| *** yuanying has joined #openstack-ironic | 07:52 | |
| *** yuanying has quit IRC | 07:54 | |
| *** yuanying has joined #openstack-ironic | 07:59 | |
| *** yuanying has quit IRC | 08:02 | |
| *** chenglch has quit IRC | 08:03 | |
| *** ijw has quit IRC | 08:03 | |
| *** yuanying has joined #openstack-ironic | 08:07 | |
| *** lucasagomes has joined #openstack-ironic | 08:09 | |
| *** andreykurilin has joined #openstack-ironic | 08:11 | |
| *** a1exhughe5 has joined #openstack-ironic | 08:12 | |
| *** dtantsur|afk is now known as dtantsur | 08:14 | |
| dtantsur | Morning everyone :) | 08:14 |
| pshige | dtantsur: morning :) | 08:14 |
| *** edcragg has joined #openstack-ironic | 08:15 | |
| *** stendulker has joined #openstack-ironic | 08:16 | |
| *** lsmola_ has joined #openstack-ironic | 08:16 | |
| *** MattMan has joined #openstack-ironic | 08:21 | |
| *** achanda has quit IRC | 08:22 | |
| lucasagomes | dtantsur, morning | 08:22 |
| dtantsur | o/ | 08:22 |
| pshige | lucasagomes: morning :) | 08:23 |
| lucasagomes | pshige, morning there | 08:23 |
| *** romcheg has joined #openstack-ironic | 08:23 | |
| *** ndipanov has joined #openstack-ironic | 08:28 | |
| *** tiagogomes_ has joined #openstack-ironic | 08:29 | |
| *** derekh has joined #openstack-ironic | 08:32 | |
| openstackgerrit | SHIGEMATSU Mitsuhiro proposed openstack/ironic: Add Ceilometer to Ironic's Conceptual Architecture https://review.openstack.org/172384 | 08:46 |
| pshige | dtantsur: thank you for your comment. I fixed it. | 08:48 |
| dtantsur | ok | 08:48 |
| dtantsur | ifarkas, hey-hey, how was your PTO? ready for some reviews? ;) | 08:50 |
| ifarkas | dtantsur, hi. thanks, it was excellent! | 08:51 |
| ifarkas | dtantsur, I will check the open patches for discoverd | 08:51 |
| dtantsur | ifarkas, yep, there should be 3 of my patches there. And you were lucky enough to avoid a huge documentation patch :D | 08:52 |
| dtantsur | thanks | 08:52 |
| ifarkas | hehe | 08:52 |
| *** pelix has joined #openstack-ironic | 08:53 | |
| *** athomas has joined #openstack-ironic | 09:01 | |
| *** jamielennox|away is now known as jamielennox | 09:02 | |
| dtantsur | lucasagomes, ifarkas, btw have you seen our folks proposing an openstackclient plugin for ironicclient? | 09:06 |
| dtantsur | it's starting with https://review.openstack.org/171672 | 09:06 |
| ifarkas | dtantsur, I haven't yet, but I will take a look | 09:07 |
| lucasagomes | dtantsur, yeah i think you posted it here on friday | 09:10 |
| lucasagomes | lemme look at the patch see if there's comments/opnions about it | 09:10 |
| dtantsur | deva wants to discuss it a bit later, otherwise seems like no comments | 09:11 |
| lucasagomes | yeah | 09:12 |
| lucasagomes | but looks good :-) | 09:12 |
| naohirot | lucasagomes: good morning, I have a question about Local boot with partition images | 09:14 |
| lucasagomes | naohirot, hi there, sure | 09:14 |
| naohirot | lucasagomes: https://github.com/openstack/ironic/commit/9a1544b76a0bb75bcf5c28647ede9079a10e097e#diff-fe0ee0bce65182e98f6966120e008656R908 | 09:14 |
| naohirot | lucasagomes: Hi, here "deployed with Ironic **must** contain ``grub2`` installed within it." | 09:14 |
| pshige | naohirot: good evening :) | 09:15 |
| lucasagomes | naohirot, yeah, just the packaged installed into it | 09:15 |
| lucasagomes | because the ramdisk uses the grub2 version from the image | 09:15 |
| lucasagomes | and isntall it on the disk | 09:15 |
| naohirot | lucasagomes: I'm not sure how to create ubuntu image with grub2. | 09:15 |
| naohirot | pshige: Hi good evening :) | 09:15 |
| lucasagomes | naohirot, with dib? yeah I had a patch up but apparently it didn't work for some people (ramesh87 said it worked for him tho) | 09:16 |
| lucasagomes | lemme try to find it | 09:16 |
| lucasagomes | naohirot, you can install it manually if needed too | 09:16 |
| lucasagomes | https://review.openstack.org/#/c/167929/ | 09:16 |
| naohirot | lucasagomes: I see, you are working on it. I couldn't find how to create the image using dib. | 09:17 |
| naohirot | lucasagomes: without dib, is there any way to create the image? | 09:18 |
| naohirot | lucasagomes: will the image be a whole disk image? | 09:19 |
| lucasagomes | naohirot, theoretically you could use other image buildings projects. Or even just download the images directly from fedora or ubuntu | 09:19 |
| lucasagomes | DIB just customize them | 09:19 |
| lucasagomes | naohirot, oh no, if u install grub2 from that package it will only install the grub2 package and that's it | 09:20 |
| lucasagomes | it won't create a full disk image (I just replied greghaynes comment there) | 09:20 |
| naohirot | lucasagomes: Okay, I think you are saying I can do it manually by refereeing the https://review.openstack.org/#/c/167929/ | 09:20 |
| naohirot | lucasagomes: right? | 09:20 |
| lucasagomes | yeah please try that patch | 09:21 |
| lucasagomes | apparently it didn't work for ubuntu with lintan_ | 09:21 |
| lucasagomes | :-/ I need to find a way to fix it | 09:21 |
| naohirot | lucasagomes: It seems readhat element has some code which invokes grub2-install, here https://github.com/openstack/diskimage-builder/blob/master/elements/redhat-common/pre-install.d/15-remove-grub | 09:23 |
| lucasagomes | yeah | 09:23 |
| lucasagomes | this element will install it anyway | 09:23 |
| lucasagomes | I believe they do remove grub by default to save space | 09:23 |
| lucasagomes | (since grub can add ~30mb in the image that won't be used if not local boot) | 09:24 |
| naohirot | naohirot: If I try redhat, can I create the image using current version of dib? | 09:25 |
| lucasagomes | naohirot, yeah you can, but for local boot you'll still need #167929 to install the bootloader | 09:26 |
| lucasagomes | or you can install it after the image generation by hand (by mouting the image and installing it in the chroot) | 09:27 |
| lucasagomes | dtantsur, can we close this https://bugs.launchpad.net/ironic/+bug/1397988 ? | 09:27 |
| openstack | Launchpad bug 1397988 in Ironic "Ironic should allow choosing the root device" [Medium,In progress] - Assigned to Lucas Alvares Gomes (lucasagomes) | 09:27 |
| naohirot | lucasagomes: Aha, 15-remove-grub | 09:27 |
| naohirot | removes grub2. So apparently I need to change it so that it doesn't remove it. | 09:27 |
| lucasagomes | naohirot, yeah but not all images come with grub2 | 09:27 |
| lucasagomes | afaict fedora cloud images doesn't | 09:27 |
| openstackgerrit | Merged openstack/ironic: Use mock patch decorator for eventlet.greenthread.sleep https://review.openstack.org/172047 | 09:28 |
| lucasagomes | the best way I would say it to use #167929 which will make sure grub2 is installed as part of the image generation | 09:28 |
| dtantsur | lucasagomes, I guess we can | 09:29 |
| lucasagomes | cool yeah because the spec code is already merged and all | 09:29 |
| naohirot | lucasagomes: I'd like to make sure additionally one thing, do we need grub2 both deploy and boot images? or just deploy image? | 09:33 |
| lucasagomes | naohirot, just for the image being deployed | 09:33 |
| lucasagomes | the deploy ramdisk doesn't need any change (it needs to be updated, since this feature is a K feature) | 09:33 |
| naohirot | lucasagomes: Okay grub2 element needs to be specified in the disk-image-create command, but not ramdisk-image-create | 09:35 |
| lucasagomes | naohirot, yes | 09:35 |
| naohirot | lucasagomes: dtantsur: okay I'll try disk-image-create with the grub2 element, thanks :) | 09:36 |
| lucasagomes | naohirot, yw :-) lemme know if it works for u | 09:36 |
| naohirot | lucasagomes: sure! | 09:36 |
| ramineni | dtantsur: hi | 09:38 |
| *** lintan_ has quit IRC | 09:39 | |
| ramineni | dtantsur: this patch https://review.openstack.org/#/c/169162/ , not getting merged since Friday, not sure what is the issue | 09:40 |
| openstackgerrit | Merged openstack/ironic: Improve AMT driver doc https://review.openstack.org/169606 | 09:43 |
| dtantsur | there was some problem with Zuul, let's recheck once more | 09:44 |
| dtantsur | done | 09:44 |
| dtantsur | ramineni, ^^ | 09:44 |
| ramineni | dtantsur: thanks | 09:45 |
| *** lintan_ has joined #openstack-ironic | 09:52 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic-specs: iRMC Virtual Media Deploy Driver for Ironic https://review.openstack.org/134865 | 09:56 |
| *** naohirot has quit IRC | 09:57 | |
| *** igordcard_ has joined #openstack-ironic | 09:58 | |
| *** romcheg has quit IRC | 10:03 | |
| *** romcheg has joined #openstack-ironic | 10:07 | |
| openstackgerrit | Merged openstack/ironic: Add Ceilometer to Ironic's Conceptual Architecture https://review.openstack.org/172384 | 10:10 |
| *** igordcard_ has quit IRC | 10:10 | |
| *** igordcard_ has joined #openstack-ironic | 10:11 | |
| *** athomas has quit IRC | 10:11 | |
| *** amitpp has joined #openstack-ironic | 10:14 | |
| *** athomas has joined #openstack-ironic | 10:18 | |
| *** amitpp has quit IRC | 10:22 | |
| kejiahu | good morning all, I am trying to use devstack deploying openstack with Ironic on HP moonshot, does anybody have any clue on the configs? thanks. | 10:25 |
| tiagogomes_ | the ssh_pxe driver passing by a ssh password doesn't work for me, I had to pass the private key | 10:27 |
| *** romcheg has quit IRC | 10:28 | |
| *** gillllliard is now known as gilliard | 10:29 | |
| *** romcheg has joined #openstack-ironic | 10:30 | |
| dtantsur | lucasagomes, what about writing a spec for https://bugs.launchpad.net/ironic/+bug/1443327 ? | 10:31 |
| openstack | Launchpad bug 1443327 in Ironic "Custom PXE parameters per node" [Wishlist,New] - Assigned to Lucas Alvares Gomes (lucasagomes) | 10:31 |
| dtantsur | lucasagomes, I suspect some people will want a pet-vs-cattle discussion again :) | 10:32 |
| lucasagomes | hmm | 10:32 |
| lucasagomes | yeah well I can do that | 10:32 |
| lucasagomes | I thought it could be very simple and not need a spec | 10:32 |
| lucasagomes | it won't change any data model or api | 10:32 |
| *** igordcard_ has quit IRC | 10:32 | |
| lucasagomes | just an extra parameter at driver_info | 10:32 |
| *** igordcard_ has joined #openstack-ironic | 10:33 | |
| kejiahu | tiagogomes_, cheers, I haven't got to the step, but I will make a node of that issue, thanks | 10:33 |
| kejiahu | s/node/note/ | 10:33 |
| dtantsur | lucasagomes, personally I don't mind it, not sure what folks will say. though as we already have a per-node deploy ramdisk, this makes sense to me | 10:33 |
| dtantsur | ok, let's go with just a bug for now | 10:34 |
| lucasagomes | ack | 10:34 |
| lucasagomes | yeah we can discuss it on the patch | 10:34 |
| lucasagomes | if it turns out we need a spec I will put one up | 10:34 |
| dtantsur | cool | 10:34 |
| tiagogomes_ | kejiahu, that was not particularly addressed to you :) | 10:35 |
| kejiahu | :) still worth to make an attention. | 10:36 |
| *** takadayuiko has joined #openstack-ironic | 10:37 | |
| *** chenglch has joined #openstack-ironic | 10:37 | |
| *** kozhukalov has joined #openstack-ironic | 10:39 | |
| *** andreykurilin has quit IRC | 10:43 | |
| *** stendulker has quit IRC | 10:50 | |
| *** ramineni has quit IRC | 10:53 | |
| tiagogomes_ | the dnsmasq process for providing an IP to baremetal seems not being able to reolve my controller hostname, as it runs with --no-resolv and --no-hosts | 10:56 |
| tiagogomes_ | any way of fixing that? | 10:56 |
| *** yuanying has quit IRC | 10:58 | |
| lucasagomes | tiagogomes_, neutron is running the dnsmasq process for u? | 11:00 |
| tiagogomes_ | lucasagomes, yes | 11:00 |
| lucasagomes | if so, I think it's possible to solve that by having a custom dnsmasq.conf | 11:00 |
| lucasagomes | you can set it in the dhcp_agent.ini | 11:00 |
| lucasagomes | /etc/neutron/dhcp_agent.ini | 11:01 |
| lucasagomes | under [DEFAULT] there's an option called "dnsmasq_config_file" | 11:01 |
| lucasagomes | I believe you can overwrite the options by setting it | 11:01 |
| * lucasagomes is not a neutron expert tho | 11:01 | |
| tiagogomes_ | lucasagomes, ok, I'll try | 11:02 |
| tiagogomes_ | thanks | 11:02 |
| lucasagomes | np | 11:02 |
| *** Haomeng has joined #openstack-ironic | 11:08 | |
| saripurigopi | any of the drivers support out-of-band h/w introspection now? | 11:10 |
| dtantsur | ilo | 11:11 |
| *** Haomeng|2 has quit IRC | 11:11 | |
| saripurigopi | dtantsur: okay. will have a look at it. | 11:17 |
| *** romcheg has quit IRC | 11:19 | |
| openstackgerrit | Yuiko Takada proposed stackforge/ironic-discoverd: Use keystonemiddleware auth credentials https://review.openstack.org/170782 | 11:21 |
| *** chenglch has quit IRC | 11:23 | |
| *** ukalifon3 has joined #openstack-ironic | 11:34 | |
| *** ukalifon3 has quit IRC | 11:34 | |
| *** ukalifon2 has joined #openstack-ironic | 11:35 | |
| *** ukalifon5 has joined #openstack-ironic | 11:37 | |
| *** ukalifon1 has quit IRC | 11:38 | |
| *** ukalifon2 has quit IRC | 11:40 | |
| *** mgoddard has joined #openstack-ironic | 11:42 | |
| *** romcheg has joined #openstack-ironic | 11:42 | |
| *** lucasagomes is now known as lucas-hungry | 11:47 | |
| *** andreykurilin has joined #openstack-ironic | 11:58 | |
| *** andreykurilin has quit IRC | 11:59 | |
| *** andreykurilin has joined #openstack-ironic | 12:00 | |
| *** romcheg has quit IRC | 12:00 | |
| *** Haomeng|2 has joined #openstack-ironic | 12:00 | |
| *** Haomeng has quit IRC | 12:03 | |
| *** ukalifon5 has quit IRC | 12:03 | |
| *** ukalifon1 has joined #openstack-ironic | 12:09 | |
| *** jamielennox is now known as jamielennox|away | 12:12 | |
| *** pradipta has quit IRC | 12:13 | |
| *** maurosr is now known as maurosr_ | 12:16 | |
| *** takadayuiko has quit IRC | 12:18 | |
| *** EmilienM|afk is now known as EmilienM | 12:19 | |
| *** zz_jgrimm has quit IRC | 12:22 | |
| *** maurosr_ has quit IRC | 12:22 | |
| *** zz_jgrimm has joined #openstack-ironic | 12:24 | |
| *** dtantsur is now known as dtantsur|brb | 12:25 | |
| *** maurosr_ has joined #openstack-ironic | 12:25 | |
| *** maurosr_ is now known as maurosr | 12:26 | |
| *** dprince has joined #openstack-ironic | 12:26 | |
| *** saripurigopi has quit IRC | 12:33 | |
| *** andreykurilin has quit IRC | 12:39 | |
| *** lucas-hungry is now known as lucasagomes | 12:41 | |
| *** jroll has quit IRC | 12:50 | |
| *** jroll has joined #openstack-ironic | 12:50 | |
| jroll | morning | 12:58 |
| lintan_ | morning jroll | 12:58 |
| jroll | hiya lintan_ :) | 12:59 |
| *** romcheg has joined #openstack-ironic | 13:00 | |
| *** openstackgerrit has quit IRC | 13:00 | |
| pshige | jroll: morning :) | 13:03 |
| lucasagomes | jroll, lintan_ morning | 13:03 |
| *** openstackgerrit has joined #openstack-ironic | 13:03 | |
| jroll | hey pshige, lucasagomes :) | 13:03 |
| pshige | lintan_: morning :) | 13:03 |
| *** krtaylor has quit IRC | 13:03 | |
| *** lintan_ has quit IRC | 13:04 | |
| *** trown|outttypeww is now known as trown | 13:04 | |
| *** Marga_ has joined #openstack-ironic | 13:05 | |
| *** rloo has joined #openstack-ironic | 13:08 | |
| *** kkoski has joined #openstack-ironic | 13:13 | |
| openstackgerrit | Brad P. Crochet proposed openstack/python-ironicclient: Implementation of baremetal reboot command https://review.openstack.org/172517 | 13:30 |
| openstackgerrit | Brad P. Crochet proposed openstack/python-ironicclient: Implementation of baremetal port create https://review.openstack.org/172461 | 13:30 |
| *** andreykurilin has joined #openstack-ironic | 13:30 | |
| NobodyCam | good morning Ironic | 13:40 |
| TheJulia | Good morning | 13:43 |
| rloo | morning TheJulia, NobodyCam | 13:44 |
| *** absubram has joined #openstack-ironic | 13:45 | |
| NobodyCam | morning TheJulia and rloo :) | 13:45 |
| NobodyCam | how was the weekend | 13:45 |
| rloo | NobodyCam: always too short :-). How was yours? | 13:46 |
| *** mjturek1 has joined #openstack-ironic | 13:46 | |
| NobodyCam | oh the same. and the "kids" made sleep ruff lastnight :-p | 13:46 |
| NobodyCam | but other wise I'm doing good | 13:47 |
| NobodyCam | :) | 13:47 |
| TheJulia | Busy, next weekend is house cleaning and then listing | 13:47 |
| rloo | oh, TheJulia. You're moving?! | 13:47 |
| NobodyCam | oh! | 13:48 |
| TheJulia | rloo: I've mostly been living with my partner for the last 6 months, so it was time to dispose of my house :) | 13:49 |
| TheJulia | s/was/is/ | 13:49 |
| rloo | TheJulia: gotcha. Makes sense! :) | 13:50 |
| NobodyCam | lol | 13:51 |
| *** mjturek1 has quit IRC | 13:51 | |
| *** mjturek1 has joined #openstack-ironic | 13:51 | |
| *** dtantsur|brb is now known as dtantsur | 13:56 | |
| dtantsur | morning jroll, NobodyCam, TheJulia, rloo! | 13:57 |
| NobodyCam | mornign dtantsur :) | 13:57 |
| NobodyCam | brb | 13:57 |
| *** a1exhughe5 has quit IRC | 14:01 | |
| *** tiagogomes_ has quit IRC | 14:01 | |
| *** edcragg has quit IRC | 14:01 | |
| *** tiagogomes_ has joined #openstack-ironic | 14:01 | |
| *** a1exhughe5 has joined #openstack-ironic | 14:01 | |
| *** edcragg has joined #openstack-ironic | 14:02 | |
| rloo | hi dtantsur, jrist | 14:02 |
| rloo | oops, jrist. hi to you too, but I meant jroll :) | 14:02 |
| *** stendulker has joined #openstack-ironic | 14:04 | |
| lucasagomes | rloo, NobodyCam TheJulia morning | 14:04 |
| lucasagomes | jroll, you there | 14:04 |
| lucasagomes | ? | 14:04 |
| NobodyCam | morning lucasagomes | 14:05 |
| NobodyCam | rloo: dtantsur: got a free second to look at https://review.openstack.org/#/c/172582 ? | 14:06 |
| * dtantsur is looking | 14:07 | |
| NobodyCam | :) awesome TY | 14:07 |
| *** jerryz has joined #openstack-ironic | 14:10 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Agent: Do not power off node if deployment fail https://review.openstack.org/172932 | 14:11 |
| NobodyCam | is jlvillal about? | 14:12 |
| pshige | NobodyCam, TheJulia, rloo: morning :) | 14:12 |
| NobodyCam | morning pshige :) | 14:13 |
| rloo | evening? pshige, aft lucasagomes | 14:14 |
| lucasagomes | :-) | 14:15 |
| pshige | rloo: actually just before midnight in Japan :) | 14:18 |
| rloo | pshige: good NIGHT then! | 14:18 |
| NobodyCam | pshige: omg thats late | 14:20 |
| *** krtaylor has joined #openstack-ironic | 14:20 | |
| *** BadCub_Hiding is now known as BadCub | 14:20 | |
| BadCub | mornin folks | 14:21 |
| NobodyCam | Morning BadCub | 14:21 |
| openstackgerrit | Zhenguo Niu proposed openstack/ironic: Fix typos in vendor-passthru.rst https://review.openstack.org/172934 | 14:21 |
| pshige | BadCub: morning | 14:21 |
| BadCub | mhiya pshige, NobodyCam :) | 14:22 |
| BadCub | hiya even | 14:22 |
| * BadCub obviously needs more coffee | 14:22 | |
| BadCub | Anyone available to look at https://review.openstack.org/#/c/171334/ ? | 14:27 |
| NobodyCam | lol: http://techcrunch.com/2015/04/09/this-custom-apple-ii-watch-is-my-new-favorite-thing/?ncid=rss&cps=gravity_1462_652061677749533598#.sw5hky:Mbwo | 14:27 |
| dtantsur | BadCub, I will | 14:28 |
| BadCub | dtantsur: awesome :) | 14:28 |
| dtantsur | BadCub, I guess devananda will want to, if he's around today | 14:29 |
| *** jistr is now known as jistr|mtg | 14:29 | |
| BadCub | dtantsur: He will most likely take a look, but we do need to land it anyway too :) | 14:30 |
| *** saripurigopi has joined #openstack-ironic | 14:30 | |
| *** achanda has joined #openstack-ironic | 14:31 | |
| NobodyCam | devananda: should be back some time today. I expect he's recovering from his PTO | 14:31 |
| jlvillal | BadCub, I'm awake now :) | 14:32 |
| NobodyCam | morning jlvillal | 14:32 |
| BadCub | jlvillal: I am working on being awake! lol | 14:32 |
| jlvillal | Sorry I meant NobodyCam :) | 14:32 |
| NobodyCam | thank you for all the work on 171334 | 14:33 |
| BadCub | LOL | 14:33 |
| NobodyCam | lol | 14:33 |
| jlvillal | NobodyCam, You're welcome. | 14:33 |
| jlvillal | NobodyCam, Hopefully some other people can try it out and see if there are any mistakes | 14:33 |
| NobodyCam | ++ | 14:33 |
| BadCub | brb | 14:33 |
| openstackgerrit | Merged stackforge/ironic-discoverd: Check hooks availability on start-up https://review.openstack.org/172394 | 14:34 |
| openstackgerrit | Merged stackforge/ironic-discoverd: Delay failure from pre-processing hooks https://review.openstack.org/172051 | 14:39 |
| *** achanda has quit IRC | 14:41 | |
| devananda | morning, all | 14:47 |
| BadCub | devananda: morning | 14:47 |
| NobodyCam | good morning devananda | 14:48 |
| devananda | what's the status of things? | 14:48 |
| * devananda is just making first cup of coffee | 14:48 | |
| BadCub | devananda: hope you had a good long weekend! :-) | 14:48 |
| NobodyCam | devananda: take a look at 171334 | 14:48 |
| devananda | BadCub: not long enough :p | 14:48 |
| openstackgerrit | Merged stackforge/ironic-discoverd: Doublecheck node provision state on receiving data from ramdisk https://review.openstack.org/172094 | 14:48 |
| BadCub | devananda: we still have two major patches to land before we can cut. Also two patches outside Ironic have not been landed. | 14:49 |
| devananda | BadCub: I see a new bug: cleaning results in deadlocks | 14:50 |
| BadCub | devananda: yes, ^ was discovered late on Friday | 14:50 |
| *** MattMan has quit IRC | 14:51 | |
| openstackgerrit | Brad P. Crochet proposed openstack/python-ironicclient: Implementation of baremetal power on and off https://review.openstack.org/172946 | 14:52 |
| pshige | devananda: morning :) | 14:54 |
| *** rwsu has quit IRC | 14:54 | |
| devananda | BadCub: everything on https://etherpad.openstack.org/p/IronicReviewDay I presume? | 14:54 |
| *** rwsu has joined #openstack-ironic | 14:54 | |
| dtantsur | devananda, morning | 14:54 |
| lucasagomes | devananda, morning | 14:57 |
| lucasagomes | devananda, back to US already? | 14:57 |
| BadCub | devananda: correct ^ is up_to_date | 14:57 |
| devananda | BadCub: cheers | 15:01 |
| devananda | lucasagomes: yup! already getting pinged by ttx about our RC status too | 15:01 |
| lucasagomes | devananda, awesome! | 15:01 |
| BadCub | devananda: :) | 15:01 |
| devananda | anyone working on the client negotiation patch now? it looks like it needs a tiny bit more work this morning | 15:02 |
| NobodyCam | BadCub: https://review.openstack.org/#/c/168844 +a'd | 15:04 |
| NobodyCam | devananda: jlvillal was working on it. | 15:04 |
| NobodyCam | jlvillal: are you still working on it? | 15:04 |
| BadCub | NobodyCam: awesomeness! TY :) | 15:04 |
| BadCub | brb | 15:06 |
| jlvillal | NobodyCam, devananda: I can stop if someone else wants to work on it. | 15:06 |
| jlvillal | I'm getting ready for work at the moment | 15:06 |
| devananda | jlvillal: nope. just checking to see that someone is :) | 15:06 |
| devananda | jlvillal: I just posted two more comments | 15:07 |
| *** jistr|mtg is now known as jistr | 15:07 | |
| jlvillal | devananda, Thanks. I will work on it. Hopefully have an updated patch in a bit. | 15:07 |
| rloo | jlvillal: I, too, have just posted some comments. Morning jlvillal :-) | 15:07 |
| jlvillal | rloo, Good morning and thanks. | 15:07 |
| *** jerryz has quit IRC | 15:07 | |
| rloo | hi lucasagomes, if you have a minute, question about your comment: https://review.openstack.org/#/c/171334/16//COMMIT_MSG | 15:09 |
| lucasagomes | rloo, sure | 15:09 |
| rloo | lucasagomes: what did you mean? by "fall forward"? | 15:09 |
| jlvillal | devananda, I think I can only do self._parse_version_headers() if I rename the function. Because the functions have the same name, unless I'm missing something. | 15:09 |
| rloo | lucasagomes: I think you mean there's a bug in ironic. nothing wrong with the patch, right? | 15:09 |
| jlvillal | devananda, But dtantsur suggested I rename the function :) | 15:09 |
| devananda | jlvillal: exactly | 15:10 |
| lucasagomes | rloo, yeah, that's something we need to fix in Ironic to not return 406 | 15:10 |
| lucasagomes | and be backward compat | 15:10 |
| devananda | jlvillal: also, you can use super(Class, self).function | 15:10 |
| rloo | lucasagomes: something we should fix before rc1 i think? | 15:10 |
| dtantsur | jlvillal, yeah, because these are 2 different functions, not one overloaded | 15:10 |
| jlvillal | devananda, Thanks | 15:10 |
| devananda | i can't type - that's the wrong method def. but the right idea | 15:10 |
| lucasagomes | rloo, that was what I was trying to point out when we were discussing that logical name patches remember? | 15:10 |
| * devananda sips is coffee | 15:10 | |
| rloo | lucasagomes: yeah, I remember that. And don't know why it wasn't fixed. Didn't we agree it should be fixed? | 15:10 |
| lucasagomes | rloo, hmm for the logical name itself I think we cna fix it in rc1, that should be simple I suppose | 15:11 |
| lucasagomes | lemme look | 15:11 |
| lucasagomes | rloo, yeah, not sure what was the agreement. But yeah it def should be fixed | 15:11 |
| *** stendulker has quit IRC | 15:12 | |
| devananda | lucasagomes: i dont understand your comment there | 15:12 |
| rloo | lucasagomes: can't find the etherpad from mrda, looking... | 15:12 |
| lucasagomes | devananda, it's because 406 is not only returned when the version requested is > than what is supported | 15:13 |
| lucasagomes | it could be the other way around too | 15:13 |
| devananda | I think I'm just not parsing "fall forward" yet | 15:13 |
| lucasagomes | so falling back to the minimum version will not help | 15:13 |
| lucasagomes | devananda, so we maybe should not fall foward | 15:13 |
| devananda | oh. you mean that an older client talking to a newer server may see new error codes? | 15:13 |
| lucasagomes | devananda, yes, that's what I was pointing in the logical name patch before | 15:14 |
| devananda | eg, if I dont upgrade my client, whereas I used to get a 400, I may start getting a 406 error? | 15:14 |
| lucasagomes | basically before the microversion if you tried to update a node with a atrribute that didnt exist it would return 400 | 15:14 |
| lucasagomes | now with microversioning and name | 15:14 |
| devananda | i think that's just fine -- it informs the client of the change | 15:14 |
| lucasagomes | if you try to add name to an old version it returns 406 | 15:14 |
| lucasagomes | imo it should return 400, as before | 15:14 |
| lucasagomes | to be backward compat | 15:14 |
| devananda | i disagree :) | 15:14 |
| lucasagomes | right | 15:15 |
| lucasagomes | the idea of microversion is not make the api looks and feel like an old api> | 15:15 |
| lucasagomes | ?* | 15:15 |
| *** zz_jgrimm is now known as jgrimm | 15:15 | |
| devananda | it is to allow the API to change while maintaining backwards compatibility with older clients, by allowing both client and server to negotiate the semantics of their exchange | 15:16 |
| devananda | whereas a client previously got a 400 error for a particular request (which it then should not attempt to repeat) | 15:16 |
| rloo | lucasagomes, devananda: found the etherpad: https://etherpad.openstack.org/p/ironic-microversion-handling | 15:17 |
| lucasagomes | right, so it's in our plan to implement a "roll forward" | 15:17 |
| devananda | if that client should continue making the request with no version header, it will start working when the server is upgraded | 15:17 |
| rloo | lucasagomes, devananda: see 1B. I think this is what you're discussing. | 15:17 |
| devananda | if that client should contoinue making hte request with a version header, it will get a different error message, indicating the server is capable of a different exchange | 15:17 |
| devananda | in either case, we haven't broken the old client | 15:17 |
| devananda | yes, i've changed my mind on this as a result of spending more time implementing the client side of this and using microversions | 15:20 |
| devananda | that said, I'd like to suggest we focus today on finalizing the RC and addressing those things which actually negatively affect a lot of users | 15:21 |
| rloo | devananda: ok. so to be clear (maybe I've lost track). with an old client + new server, no header is sent to the server. for lucas' example, the server defaults to 1.1, and will 406 cuz logical names aren't supported. whereas it would have 400'd with an old server. | 15:22 |
| devananda | rloo: right | 15:22 |
| rloo | devananda: ok. I"m going to update that etherpad and add a comment to lucas' so it is 'recorded'. then i'll focus on the rc1 stuff :-) | 15:23 |
| lucasagomes | ack | 15:23 |
| lucasagomes | thanks rloo | 15:23 |
| *** MattMan has joined #openstack-ironic | 15:23 | |
| devananda | I don't think this will affect our users nearly as much as, say, cleaning causing deadlocks, or the new client refusing to connect to any pre-1.6 server | 15:23 |
| devananda | but I agree that it could, potentially, affect someone who was relying on the server failing with an HTTp 400 error | 15:24 |
| lucasagomes | yea | 15:24 |
| lucasagomes | agree with that, is not serious, but is a breakage | 15:24 |
| devananda | when they tried, repeatedly, to update a property which did not exist :) | 15:24 |
| dtantsur | well, we used microversion to not break people who expect 400 from node-set-provision-state inspect | 15:25 |
| devananda | lucasagomes: but who does it affect? only someone who expected to attempt to update a specific non-existent property | 15:25 |
| dtantsur | while it's weird to me, we could continue gating literally everything | 15:25 |
| lucasagomes | devananda, yeah, that's true... but that's one example... dtantsur is has another | 15:26 |
| devananda | dtantsur: I agree - we could - but the more we dig into this, the more I think that's overkill | 15:26 |
| lucasagomes | trying to find a node by it's name instead of uuid with an old server the same | 15:26 |
| lucasagomes | or old client* | 15:26 |
| lucasagomes | the 406 is returned in other places which wasn't returned before | 15:27 |
| lucasagomes | but I agree I don't see anything major here | 15:27 |
| devananda | :) | 15:28 |
| lucasagomes | and my idea of microversioning maybe was wrong, I thought it was be a layer to make the API feels like an old API | 15:28 |
| lucasagomes | which is not | 15:28 |
| *** mtanino has joined #openstack-ironic | 15:32 | |
| devananda | JoshNang: i'm trying to understand https://bugs.launchpad.net/ironic/+bug/1442810 | 15:36 |
| openstack | Launchpad bug 1442810 in Ironic "Cleaning results in deadlocks" [High,In progress] - Assigned to Josh Gachnang (joshnang) | 15:36 |
| JoshNang | devananda: yeah, i'm frankly not clear on the root cause here. | 15:37 |
| NobodyCam | He was hitting that whrn attempting to put 20+ node in to cleaning at one time | 15:37 |
| NobodyCam | oh morning JoshNang | 15:37 |
| JoshNang | good morning :) | 15:38 |
| openstackgerrit | Brad P. Crochet proposed openstack/python-ironicclient: Implementation of set and unset https://review.openstack.org/172963 | 15:38 |
| devananda | JoshNang: you say "deadlock" - this has a very specific meaning in DB terms. do you, in fact, mean database deadlock? or something else? | 15:39 |
| NobodyCam | I understood it more as Hung or Stuck nodes? | 15:40 |
| JoshNang | yeah, i think that was a poor word choice. it was a stuck lock | 15:41 |
| devananda | JoshNang: stuck waiting on what? | 15:41 |
| devananda | JoshNang: converting from RPC call -> cast shouldn't change how or when a TaskManager lock is reelased | 15:42 |
| JoshNang | when it tried to acquire the lock in the conductor's continue_node_clean, it would almost always hit a nodelocked exception | 15:42 |
| devananda | nor should it change how or when the lock is acquired ... | 15:42 |
| JoshNang | right, it didn't seem sensible looking at the code | 15:42 |
| lucasagomes | JoshNang, after the error. The "resevation" field of the node was still set? | 15:43 |
| lucasagomes | devananda, I think it doesn't... It just call cast and return. If it fails to acquire the lock it will try again on the next heartbeat | 15:43 |
| JoshNang | lucasagomes: yeah. later i manually cleared them but they came back the next time it got to that method | 15:43 |
| lucasagomes | so it will eventually work | 15:43 |
| lucasagomes | that's why I think that the Closes-Bug tag there is wrong, cause it just work it around | 15:44 |
| devananda | lucasagomes: ah, i see | 15:44 |
| lucasagomes | it doesn't fix the problem with the locks | 15:44 |
| lucasagomes | which can be a race in between there | 15:44 |
| devananda | idea! | 15:44 |
| lucasagomes | JoshNang, gotcha | 15:44 |
| devananda | JoshNang: the continue-clean is passed by the agent on every heartbeat, right? | 15:44 |
| JoshNang | yeah, i think it has to be a race. some nodes finished cleaning. i got a handful of the 20 to finish | 15:44 |
| JoshNang | devananda: yes | 15:44 |
| devananda | JoshNang: so what if the conductor is in the middle of one cleaning operation when the agent sends a heartbeat? | 15:44 |
| devananda | JoshNang: it's going to error. because it can't continue cleaning while it's already cleaning | 15:45 |
| *** jistr has quit IRC | 15:45 | |
| JoshNang | let me look. i think it only continues when it sees the current command is complete, or no command has started (when the agent first boots) | 15:45 |
| BadCub | brb | 15:47 |
| JoshNang | yeah, the rpc only gets called on success or clean_version_mismatch (which restarts it) https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py#L139 | 15:47 |
| lucasagomes | yeah it will just return if the command still running and do nothing | 15:47 |
| lucasagomes | https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py#L322-L323 | 15:48 |
| lucasagomes | which goes here https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py#L155-L157 | 15:48 |
| devananda | conductorManager._do_next_clean_step starts off by changing node.clean_step and calling node.save() | 15:49 |
| devananda | so it's going to mismatch on the next heartbeat, won't it? | 15:49 |
| * lucasagomes diggs | 15:49 | |
| JoshNang | right | 15:50 |
| *** romcheg has quit IRC | 15:51 | |
| devananda | yea, on any heartbeat, if node.provision_state == states.CLEANING and node.clean_step: it's going to try to start cleaning again | 15:51 |
| devananda | even if cleaning is already in progress | 15:52 |
| lucasagomes | hmm so on heartbeat, if the node is cleaning we should perhaps test the progress | 15:54 |
| JoshNang | no, it'll check the agent, see the clean_step is different than the last command on the agent, and stop | 15:54 |
| lucasagomes | if it's running just do nothing | 15:54 |
| * lucasagomes still needs to read more the code | 15:54 | |
| lucasagomes | oh yeah... | 15:54 |
| JoshNang | the first thing it does if its CLEANING and has a clean step is run through checks for failed, running, not the right step, etc here: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py#L308 | 15:55 |
| lucasagomes | we never call continue_cleaning from the api, it's only called from the agent_base_vendor | 15:55 |
| devananda | JoshNang: shouldn't all this just check that the node isn't already locked by another thread and, if so, stop? | 15:56 |
| JoshNang | devananda: doesn't vendor passthru grab the lock before starting any of this code? | 15:56 |
| devananda | JoshNang: the heartbeat -> continue_cleaning() -> _notify_conductor_resume_clean path does not use an exclusive lock | 15:56 |
| openstackgerrit | Merged openstack/ironic: iLO driver documentation for UEFI secure boot https://review.openstack.org/168844 | 15:56 |
| devananda | JoshNang: /me double cehcks | 15:57 |
| kbs1 | Anyone have a good install/config/setup guide for IPA somewhere? agent accesses to temp_urls are throwing authorization errors so I'd like to see if something basic in my setup is busted. | 15:57 |
| *** igordcard_ has quit IRC | 15:57 | |
| *** mgoddard has quit IRC | 15:57 | |
| *** mgoddard1 has joined #openstack-ironic | 15:57 | |
| jlvillal | rloo, Instead of doing an assert for the value not being one of the valid states, do you have a suggestion? I could raise a ValueError? | 15:59 |
| NobodyCam | kbs1: are you using devstack? | 15:59 |
| rloo | jlvillal: wondering what will be useful for the user. | 16:00 |
| kbs1 | No. Using a pure bare metal deployment. | 16:00 |
| JoshNang | kbs1: our docs for that are definitely lacking there. maybe a mismatched swift user or temp url key? https://github.com/openstack/ironic/blob/master/etc/ironic/ironic.conf.sample#L754 | 16:00 |
| rloo | jlvillal: need to look at the code again. sec. | 16:00 |
| jlvillal | rloo, I'm not sure. It would indicate a bug in the code. | 16:00 |
| devananda | JoshNang: hm, yea. vendor_passthru should take an exclusive lock. driver_vendor_passthru does not, which is what I was thinking of | 16:00 |
| rloo | jlvillal: internal error, go on irc and ask? :-) | 16:01 |
| kbs1 | NobodyCam: Using a devstack setup as a cross-reference, though. | 16:01 |
| jlvillal | rloo, I'm just trying to be a little defensive to make sure we don't run along and have a huge error message if it goes recursive until the stack fills up | 16:01 |
| rloo | jlvillal: can we default to something instead of asserting? | 16:02 |
| NobodyCam | kbs1: so you've seem http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html#deploying-ironic-with-devstack | 16:02 |
| NobodyCam | seen even | 16:02 |
| rloo | jlvillal: or just leave the assert in there :-) | 16:02 |
| jlvillal | rloo, So try to cover up a bug in the code? | 16:02 |
| jlvillal | rloo, I vote for leave the assert there as it should only trigger if it is a bug in the code. | 16:03 |
| *** dttocs has joined #openstack-ironic | 16:03 | |
| rloo | jlvillal: so in general, if code is being used by others, it just seems odd to have asserts. developers always say, "it won't trigger in production", but what if it does?... | 16:03 |
| rloo | jlvillal: but it isn't a big deal cuz it won't trigger when the user is using it. | 16:04 |
| jlvillal | rloo, I would assume if it does trigger in product developers should be told because it is a bug. I can change it to a an if statement and raise a ValueError() if that makes it more appealing. | 16:05 |
| jlvillal | rloo, I would prefer to not ignore the case though. | 16:05 |
| rloo | jlvillal: it also seems to me that self.os_ironic_api_version should have a value too? and we aren't asserting on that? | 16:05 |
| jlvillal | rloo, Agreed. | 16:06 |
| *** ifarkas has quit IRC | 16:07 | |
| rloo | jlvillal: this is the first assert being added to the client. it opens the door for more asserts. just want to make sure we think it through. it is a personal thing for me, that i don't like asserts in 'production' code. | 16:09 |
| jlvillal | rloo, Okay. I will just change it to an if statement and raise an Exception. Works out to the same thing :) | 16:10 |
| kbs1 | NobodyCam: Yes. I've seen this. | 16:10 |
| rloo | jlvillal: but with hopefully a more meaningful msg :-) | 16:10 |
| jlvillal | rloo, A rose by a different name.... :) | 16:10 |
| rloo | jlvillal: think of the poor user or developer:-) | 16:11 |
| jroll | kbs1: might be worth checking for clock skew between ironic-conductor and swift | 16:11 |
| jroll | lucasagomes: I'm back now, you pinged? | 16:11 |
| lucasagomes | jroll, welcome back :-) yeah, trying to remember what was it | 16:11 |
| lucasagomes | ah | 16:11 |
| kbs1 | jroll: will do. | 16:12 |
| lucasagomes | jroll, do you guys have some downstream code that keeps the node powered on in case the deployment fail? | 16:12 |
| kbs1 | JoshNang: will check that out as well. | 16:12 |
| jroll | lucasagomes: yeah, we do | 16:12 |
| jroll | lucasagomes: mostly just hacks, nothing interesting, but I can share if that's helpful | 16:12 |
| lucasagomes | jroll, right I put some code up for that so I can access the logs from IPA when it fails | 16:12 |
| lucasagomes | jroll, no it's ok | 16:12 |
| JayF | I thought it behaved that way upstream :( | 16:13 |
| lucasagomes | I was more wondering... what's the action in nova | 16:13 |
| JayF | cleaning/cleanfail works that way | 16:13 |
| JoshNang | yeah that's on my list of "need to upstream soonish" | 16:13 |
| lucasagomes | JayF, not really :-( I added a config option that so we can do it | 16:13 |
| jroll | JayF: I don't think anything like that is upstream | 16:13 |
| jroll | lucasagomes: what do you mean the action in nova | 16:13 |
| lucasagomes | jroll, JayF https://review.openstack.org/#/c/172932/ | 16:13 |
| JoshNang | cleanfail definitely doesn't power off the node | 16:13 |
| lucasagomes | jroll, so see the Depends-On tag in that patch, it links to a patch in nova | 16:14 |
| lucasagomes | which I prevent the destroy() if the node is in maintenance | 16:14 |
| lucasagomes | I wonder if you do same, or you have a smarter way of doing it | 16:14 |
| JayF | oooh | 16:14 |
| jroll | lucasagomes: aha, I see, now that I think about it we do let nova kill it | 16:14 |
| JayF | our downstream doesn't behave that way | 16:14 |
| JayF | ==jroll | 16:15 |
| jroll | so we don't leave it on (though we power on after destroy is done) | 16:15 |
| lucasagomes | right, /me is glad I asked | 16:15 |
| lucasagomes | but if you power off you lose the logs | 16:15 |
| lucasagomes | cause when the deploy fails, nova automatically calls destroy() | 16:15 |
| jroll | right, so in general we don't have issues that the ramdisk logs will help with | 16:15 |
| *** saripurigopi has quit IRC | 16:15 | |
| jroll | almost never | 16:15 |
| NobodyCam | morning jroll JayF :) | 16:15 |
| lucasagomes | right | 16:16 |
| JayF | morning | 16:16 |
| * JayF not around long, will be on a plane in an hour | 16:16 | |
| lucasagomes | jroll, I find it specially useful the logs to debug IPA itself when implementing a new extension or something | 16:16 |
| NobodyCam | JayF: where you off to this fine monday | 16:16 |
| jroll | lucasagomes: the error passed back to ironic is usually enough to tell, e.g. we had a bad load balancer in front of swift one day, was fairly obvious what the issue is | 16:16 |
| lucasagomes | I worked it around ismilar to what I put in the patch when doing the iscsi for ipa | 16:16 |
| jroll | lucasagomes: indeed, for dev it's useful, we haven't found it useful in prod | 16:16 |
| lucasagomes | so I could see what went wrong in IPA | 16:16 |
| JayF | NobodyCam: training at our home office in Texas. Part of our "Emerging Leaders" program | 16:17 |
| jroll | lucasagomes: I'm +1 on the idea of these patches fwiw | 16:17 |
| lucasagomes | jroll, ack JayF | 16:17 |
| lucasagomes | yeah I want it to also be useful for you guys downstream | 16:17 |
| devananda | JoshNang: your change from call -> cast assumes that no other driver will need a response from continue_node_clean() | 16:17 |
| lucasagomes | perhaps we should do differently on the nova side | 16:17 |
| jroll | lucasagomes: yeah, dunno if we'd use it in prod, but we might | 16:17 |
| lucasagomes | but maybe we have different use cases for that | 16:17 |
| lucasagomes | ok | 16:18 |
| lucasagomes | cheers for the infos/ideas | 16:18 |
| NobodyCam | JayF: awesome! enjoy! | 16:18 |
| jroll | lucasagomes: quickly looking, both patches seem sane :) | 16:18 |
| JoshNang | devananda: correct, but that conductor function isn't returning anything ATM | 16:18 |
| lucasagomes | yeah I've tested it here locally | 16:18 |
| lucasagomes | it's simple, I default it to continue doing what it does today | 16:19 |
| devananda | JoshNang: except exceptions :) | 16:19 |
| lucasagomes | so it won't break anything | 16:19 |
| JoshNang | devananda: :) true | 16:19 |
| lucasagomes | it's more to help me and others with debugging IPA when needed | 16:19 |
| JayF | I bet it'd be interesting to run production that way | 16:19 |
| JayF | what I'd love to see is a way to say "was the falure ironic or external" and decide to leave on / destroy accordingly | 16:20 |
| devananda | JoshNang: anyway, this change is just ignoring the errors that continue_node_clean() may o rmay not raise, but I agree that there is absolutely no need for the calling thread to wait on the receiving thread in the same conductor in this case | 16:20 |
| JayF | for instnace, we have failures to neutron occassionally. There's no value in leaving that Agent up because Neutron did everything right, but it's a failure all the same | 16:20 |
| lucasagomes | JayF, oh right, yeah right now it just do for any failures. It would be interesting indeed to maybe classify how it failed and act upon that | 16:21 |
| JoshNang | devananda: right. for errors you get a vague 'expected exception' log message, and because the heartbeats keep coming, it'll try again later. certainly not ideal | 16:21 |
| devananda | JoshNang: so I'm good with the change. I'm still not sold on the whole driver-sends-rcp-to-the-conductor approach, though | 16:21 |
| devananda | JoshNang: mind if I clean it up a bit? | 16:22 |
| JoshNang | devananda: go ahead! | 16:22 |
| lucasagomes | JayF, I can try to take a look at it. The idea here now was just a "hey, deployment fails. You still have all logs so go and figure out" | 16:22 |
| JayF | lucasagomes: just a thought; you doing what you're doing is better than what we have now :) | 16:23 |
| JoshNang | agreed, but it was the least bad way i had come up with for getting back to the conductor from the driver in an async fashion. i'd love for a better solution | 16:23 |
| *** a1exhughe5 has quit IRC | 16:23 | |
| JayF | lucasagomes: just specifically finding agent failures and keeping the agent on would be something I'd be +1 to running in my prod | 16:23 |
| lucasagomes | nice, thanks for the feedback. Yeah I will take a look at it | 16:23 |
| *** Sukhdev has joined #openstack-ironic | 16:24 | |
| JoshNang | JayF: there's a spec up to send the agent logs back on heartbeats and after deployments (presumably failures as well). | 16:25 |
| *** lsmola_ has quit IRC | 16:25 | |
| openstackgerrit | John L. Villalovos proposed openstack/python-ironicclient: Client should fall back to the lower versions if necessary https://review.openstack.org/171334 | 16:32 |
| jlvillal | rloo, devananda ^^^^ | 16:32 |
| jlvillal | And anyone else who would like to review :) | 16:32 |
| JayF | JoshNang: I saw that, and am currently pondering it | 16:32 |
| devananda | JoshNang: yanno what, nvm. leaving comments on the review. I think this needs one more pass | 16:34 |
| *** ukalifon1 has quit IRC | 16:34 | |
| JoshNang | devananda: kk | 16:34 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Update ilo drivers documentation for inspection https://review.openstack.org/170065 | 16:37 |
| devananda | JoshNang: posted | 16:40 |
| JoshNang | devananda: thanks | 16:40 |
| devananda | JoshNang: the change i just suggested breaks 2 unit tests | 16:41 |
| openstackgerrit | John L. Villalovos proposed openstack/python-ironicclient: Client should fall back to the lower versions if necessary https://review.openstack.org/171334 | 16:41 |
| *** subscope_ has joined #openstack-ironic | 16:41 | |
| jlvillal | ^^^^ very minor change to make sure test both 'user' and 'negotiated' failure paths. | 16:41 |
| JoshNang | devananda: is partial-bug the better tag here? not sure if there's a 'workaround' tag | 16:42 |
| JoshNang | or related-bug | 16:43 |
| * NobodyCam thinks partial-bug | 16:43 | |
| JoshNang | wfm | 16:43 |
| jlvillal | JoshNang, https://wiki.openstack.org/wiki/GitCommitMessages There is 'Related-Bug' | 16:43 |
| jlvillal | JoshNang, I think you listed all the options they talk about | 16:44 |
| JoshNang | jlvillal: thanks | 16:44 |
| *** MattMan has left #openstack-ironic | 16:47 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Open Liberty development https://review.openstack.org/171274 | 16:48 |
| devananda | both related and partial will leave the bug open | 16:48 |
| devananda | if we're saying "this bug is fixed in kilo" we should actually use closes-bug | 16:48 |
| devananda | (this is me thinking through it) the bug is "cleaning results in deadlocks". assuming this patch results in no more deadlocks, then it closes it. | 16:49 |
| devananda | JoshNang: ^ | 16:50 |
| lucasagomes | devananda, oh good catch on the RPC_VERSION bump | 16:51 |
| *** ramineni has joined #openstack-ironic | 16:51 | |
| * lucasagomes should have catched it | 16:51 | |
| JoshNang | alright, i'll stick with closes. | 16:52 |
| devananda | JoshNang: also, would the bug description be better as "cleaning may leave nodes locked and require manual intervention to unlock" ? | 16:53 |
| JoshNang | yes, definitely | 16:54 |
| * devananda edits | 16:54 | |
| *** stendulker has joined #openstack-ironic | 16:55 | |
| devananda | anyone tested https://review.openstack.org/#/c/171743/ ? | 16:56 |
| devananda | that's the DIB patch for https://bugs.launchpad.net/ironic/+bug/1441556 | 16:56 |
| openstack | Launchpad bug 1441556 in diskimage-builder "Deploying whole disk image with local boot fails" [Medium,Triaged] - Assigned to Lucas Alvares Gomes (lucasagomes) | 16:56 |
| * lucasagomes looks | 16:57 | |
| *** derekh has quit IRC | 16:57 | |
| lucasagomes | oh perhaps greg should be assigned to that bug | 16:57 |
| lucasagomes | he did the workd | 16:57 |
| *** dtantsur is now known as dtantsur|afk | 16:57 | |
| devananda | I see a comment from NobodyCam that it needs to be tested, but no furhter comments from any of us that it *was* tested | 16:58 |
| devananda | which seems to be why the DIB team hasn't approved it yet (and reasonably so) | 16:58 |
| lucasagomes | I can try give it a go | 16:59 |
| lucasagomes | lemme see if I have an env handy | 16:59 |
| NobodyCam | oh the dib patch. I was going to see if I could fix my broken env today to test that | 16:59 |
| NobodyCam | lucasagomes: that would be awesome@ | 16:59 |
| NobodyCam | s/@/!/ | 16:59 |
| *** stendulker has quit IRC | 16:59 | |
| rloo | jlvillal: thx, see comments. | 17:03 |
| *** Marga_ has quit IRC | 17:04 | |
| *** Marga_ has joined #openstack-ironic | 17:05 | |
| jlvillal | rloo, Thanks. Yeah I'm not exactly sure what to do for the wording. I will think.... :) | 17:05 |
| *** Marga_ has quit IRC | 17:05 | |
| *** Marga_ has joined #openstack-ironic | 17:06 | |
| jlvillal | rloo, I like your suggestion | 17:07 |
| *** subscope_ has quit IRC | 17:08 | |
| NobodyCam | brb | 17:09 |
| lucasagomes | NobodyCam, devananda testing that local boot patch for DIB | 17:11 |
| openstackgerrit | Anusha Ramineni proposed openstack/ironic: iLO driver documentation for node cleaning https://review.openstack.org/169162 | 17:11 |
| openstackgerrit | Josh Gachnang proposed openstack/ironic: Convert internal RPC continue_node_cleaning to a "cast" https://review.openstack.org/172582 | 17:13 |
| rloo | jlvillal: the 'negotiated version' stuff may not be clear. I wonder if we should mention something like 'no API version was specified and the version used by the client didn't work.' | 17:14 |
| devananda | lucasagomes: cheers | 17:18 |
| *** achanda has joined #openstack-ironic | 17:18 | |
| devananda | jlvillal: yea, the wording is really close but could be a bit better | 17:18 |
| *** davideagnello has joined #openstack-ironic | 17:18 | |
| devananda | other than that, this looks good. /me tests locally | 17:19 |
| *** ramineni has quit IRC | 17:20 | |
| devananda | jlvillal: also, the message is much longer than my command line ... perhaps wrapping it would help :) | 17:20 |
| BadCub | anyone have a moment to look at: https://review.openstack.org/#/c/170065/ Needs +2/+A | 17:21 |
| jlvillal | devananda, Okay :) | 17:21 |
| *** edcragg has quit IRC | 17:21 | |
| BadCub | also - https://review.openstack.org/#/c/169162/ - ILO Driver doc for node cleaning Need to have +2s/+A reapplied. Was rebased | 17:21 |
| * BadCub kicks etherpad | 17:24 | |
| NobodyCam | BadCub: https://review.openstack.org/#/c/169162 re +a'd | 17:25 |
| BadCub | NobodyCam: ty :-) | 17:25 |
| *** jcoufal has quit IRC | 17:26 | |
| *** zhenguo has quit IRC | 17:27 | |
| NobodyCam | brb | 17:28 |
| openstackgerrit | John L. Villalovos proposed openstack/python-ironicclient: Client should fall back to the lower versions if necessary https://review.openstack.org/171334 | 17:32 |
| jlvillal | rloo, devananda: Feed back on the messages appreciated. Thanks. | 17:32 |
| rloo | NobodyCam: wrt 170065, are we going to change things so that python-ironicclient has a kilo version? | 17:33 |
| rloo | NobodyCam: vs version 0.4 or whatever it is. | 17:33 |
| *** Marga_ has quit IRC | 17:33 | |
| *** meghal has joined #openstack-ironic | 17:34 | |
| *** harlowja_away is now known as harlowja | 17:37 | |
| lucasagomes | devananda, http://paste.openstack.org/show/203666/ | 17:40 |
| *** Marga_ has joined #openstack-ironic | 17:40 | |
| lucasagomes | doens't seem to work, node is stuck in the deploywait even when it's finished the bootloader installation | 17:40 |
| lucasagomes | and the pass_bootloader_install_info is present at methods | 17:40 |
| *** trown is now known as trown|lunch | 17:41 | |
| * lucasagomes will test without the change see if it works fine (could be his env) | 17:42 | |
| *** Marga_ has quit IRC | 17:44 | |
| lucasagomes | devananda, apparently that patch is not passing any credentials to the curl command | 17:45 |
| NobodyCam | rloo: let me take a look. I just addressed comments in the last rev of that review | 17:45 |
| devananda | lucasagomes: huh. good catch | 17:45 |
| rloo | NobodyCam: sorry, I didn't check to see who added what in that patch. Might not be you. | 17:48 |
| NobodyCam | :) | 17:49 |
| *** r-daneel has joined #openstack-ironic | 17:51 | |
| *** athomas has quit IRC | 17:55 | |
| devananda | jlvillal: LGTM | 17:56 |
| devananda | lucasagomes: you working on the DIB patch? | 17:57 |
| lucasagomes | devananda, I'm just testing it w/o to see if it's my env (but I doubt) | 17:57 |
| lucasagomes | it's late here :/ | 17:57 |
| lucasagomes | unless nobody else can do it I will | 17:57 |
| devananda | lucasagomes: ah, of course | 17:57 |
| BadCub | anyone want to look at https://review.openstack.org/#/c/172582/ & https://review.openstack.org/#/c/171334/ ? | 17:58 |
| lucasagomes | do DIB has a cut? | 17:58 |
| *** mgagne_ is now known as mgagne | 17:58 | |
| lucasagomes | cause since it's not in the Ironic code databse should we wait on it ? | 17:58 |
| devananda | DIB isn't going to do a stable branch, as far as I know | 17:58 |
| lucasagomes | right, so we could cut our RC-1 without relying on that patch right? | 17:58 |
| devananda | but if we can land a fix in DIB, at least we can document in our release notes that this bug is fixed with DIB after some point in time | 17:59 |
| jlvillal | devananda, I like your plan :) | 17:59 |
| lucasagomes | right | 17:59 |
| devananda | correct me if i'm wrong, but without fixing this in DIB, we have to say that it is currently not supported in Kilo | 17:59 |
| lucasagomes | yeah that would be nice :/ | 17:59 |
| devananda | at this point, i'm really tempted to just do that | 17:59 |
| lucasagomes | | f30027a7-4e1a-4d70-b4f8-9b5619714a80 | testing | ACTIVE | - | Running | private=10.1.0.13 | | 18:00 |
| lucasagomes | it works without the patch | 18:00 |
| lucasagomes | devananda, well it's only supported on Kilo | 18:00 |
| *** pelix has quit IRC | 18:00 | |
| jlvillal | rloo, I'm used to two spaces after a period. But I have a co-worker of mine who hates that ;) | 18:00 |
| *** mgoddard1 has quit IRC | 18:00 | |
| lucasagomes | devananda, there's 2 problems here, if we use an old ramdisk with new server | 18:00 |
| lucasagomes | it won't call the pass_bootloader_info etc... so Ironic will get stuck on DEPLOYWAIT waiting for it | 18:01 |
| jroll | jlvillal: the never-ending battle | 18:01 |
| rloo | jlvillal: not having met you, my guess is that you're older than the younger folks ;) | 18:01 |
| lucasagomes | that patch is trying to fix the old server (which doesn't contain the pass_bootloader_info method) with a new ramdisk | 18:01 |
| rloo | jlvillal: so it shouldn't matter, if something is pretty-fying your text. I don't really care, just thought I'd point it out for consistency blah. | 18:02 |
| *** romcheg has joined #openstack-ironic | 18:03 | |
| lucasagomes | devananda, for the problem #1 I think we talked about passing a flag as part of the pass_deploy_info, and if not set Ironic would not wait on pass_bootloader_info | 18:03 |
| devananda | lucasagomes: right. we've introduced a new feature which breaks deployments if you try to use the new feature after upgrading, without ALSO updating the dib ramdisk. | 18:03 |
| rloo | jlvillal: why, back in the good old days when I used a typewriter... | 18:03 |
| lucasagomes | but apparently nobody did that | 18:03 |
| lucasagomes | devananda, exactly | 18:03 |
| devananda | lucasagomes: there was not enough thought put into the upgrade path for people, and as such, we're trying (still) to work around that with some very late fixes | 18:03 |
| lucasagomes | so we can put in our docs, that for local boot it also requires updating the deploy ramdisk | 18:03 |
| jroll | uh | 18:04 |
| jroll | lucasagomes: devananda: I thought we fixed problem #1 | 18:04 |
| jroll | old ramdisk with new server | 18:04 |
| * lucasagomes looks at the code | 18:04 | |
| jroll | never mind me, I'm thinking of whole dism image | 18:04 |
| jlvillal | rloo, I am older than the younger folks :) | 18:04 |
| devananda | jroll: right. whole disk image + netboot landed | 18:04 |
| jroll | yeah | 18:04 |
| rloo | jlvillal: :D | 18:05 |
| jroll | ignore me, too many things on my mind :) | 18:05 |
| NobodyCam | rloo: looking at your comemnt in ilo.rst on 170065, are you looking for a break down of essential vs extra there? | 18:05 |
| jlvillal | rloo, And I actually did take a typewriter class in high school :) | 18:05 |
| devananda | jroll: this is the ramdisk waiting or not waiting on an extra callback from ironic | 18:05 |
| jroll | right. | 18:05 |
| jlvillal | Which helped me type faster on my computer... | 18:05 |
| devananda | we found a way to detect that in newer ramdisks | 18:05 |
| jroll | devananda: I typed before I thought, really, ignore me | 18:05 |
| devananda | jroll: :) | 18:05 |
| jroll | :) | 18:05 |
| lucasagomes | yeah what devananda said :-) | 18:05 |
| lucasagomes | the whole_disk_image is fixed yeah | 18:06 |
| jlvillal | rloo, I took the typewriter class while I owned a computer. TRS-80 | 18:06 |
| rloo | NobodyCam: from the user's point of view, what is the difference, what does it matter? | 18:06 |
| lucasagomes | devananda, so, maybe document that to use local boot we require also to update the deploy ramdisk with a new version | 18:06 |
| lucasagomes | if used with DIB | 18:06 |
| NobodyCam | rloo: ack... let me see what I can comeup with | 18:06 |
| lucasagomes | with DIB ramdisk* | 18:06 |
| rloo | jlvillal: TRS-80 rings a bell ;) | 18:06 |
| devananda | lucasagomes: does that obviate the need for this patch? | 18:07 |
| NobodyCam | model i ii or III | 18:07 |
| jlvillal | rloo, NobodyCam: I started with the Model I | 18:07 |
| rloo | NobodyCam: not a big deal. It can be added later. but I thought there was some diff, that eg essential were/had to be all discovered. | 18:07 |
| NobodyCam | :) nice | 18:07 |
| lucasagomes | devananda, well that patch only solves the old server with new ramdisk | 18:07 |
| devananda | lucasagomes: no ... i think this patch is making the newer DIB ramdisk compatible with older ironic services | 18:07 |
| devananda | right | 18:07 |
| devananda | so we still need that | 18:07 |
| devananda | but it doesn' tneed to block our RC1 | 18:08 |
| lucasagomes | yeah, exactly | 18:08 |
| NobodyCam | jlvillal: remember https://www.youtube.com/watch?v=FQizYzw27FY | 18:08 |
| lucasagomes | it would be nice to have the other bits old ramdisk with new server in place... but seems late | 18:08 |
| lucasagomes | that will also require another change in DIB to set that flag and so on :/ | 18:09 |
| *** Marga_ has joined #openstack-ironic | 18:09 | |
| jlvillal | NobodyCam, I remember LDOS :) | 18:09 |
| NobodyCam | keep watching | 18:09 |
| lucasagomes | but overall I think it's ok... local boot is a Kilo feature | 18:09 |
| lucasagomes | so requiring a new DIB to work with it doesn't seems crazy | 18:09 |
| lucasagomes | it's not awesome, but not bad either | 18:09 |
| jlvillal | NobodyCam, Oh yeah I remember the dancing demon :D | 18:10 |
| * lucasagomes we should try to be awesome... | 18:10 | |
| NobodyCam | lol | 18:10 |
| jroll | lucasagomes: ++ | 18:10 |
| jroll | if you're using juno-era ramdisk, you need to update | 18:10 |
| jroll | so it doesn't seem all that terrible | 18:11 |
| lucasagomes | yes | 18:11 |
| lucasagomes | to use that specific feature | 18:11 |
| jroll | right | 18:11 |
| lucasagomes | it's not that we are breaking deployments or anything | 18:11 |
| lucasagomes | only if you want to use local boot | 18:11 |
| jroll | yeah | 18:11 |
| lucasagomes | right, so I will relax a bit if people agree on that | 18:12 |
| lucasagomes | and get a break | 18:12 |
| devananda | lucasagomes: requiring a new DIB to use the new feature is fine. requiring a new DIB just because I upgraded to Kilo is not fine, and breaking deploys if I choose t ouse a newer ramdisk and NOT update my service -- also not fine. | 18:12 |
| *** ijw has joined #openstack-ironic | 18:12 | |
| lucasagomes | devananda, right yeah we don't do that | 18:12 |
| devananda | lucasagomes: well, without this DIB patch, we have broken deployments of stable/juno, right? | 18:12 |
| *** Marga_ has quit IRC | 18:12 | |
| devananda | because the new ramdisk is going to expect a second callback tht will never come | 18:13 |
| lucasagomes | devananda, no we don't, because we check if we passed the boot_option kernel cmdline in the ramdisk | 18:13 |
| *** Marga_ has joined #openstack-ironic | 18:13 | |
| devananda | oh | 18:13 |
| lucasagomes | and old servers won't do that | 18:13 |
| devananda | yea | 18:13 |
| lucasagomes | and for new servers we check if boot_option = local before waiting for the call back | 18:13 |
| *** Marga_ has quit IRC | 18:13 | |
| devananda | right. too many different race conditions in my brain right now :-/ | 18:13 |
| lucasagomes | no problem | 18:13 |
| lucasagomes | so it's not terrible | 18:13 |
| *** Marga_ has joined #openstack-ironic | 18:13 | |
| lucasagomes | there's room for making it awesome, but we failed on that :-( | 18:14 |
| devananda | ok - so lets recap - what are the release blockers at this point | 18:15 |
| openstackgerrit | Merged openstack/python-ironicclient: Upgrade hacking to latest release https://review.openstack.org/172072 | 18:15 |
| devananda | https://bugs.launchpad.net/ironic/+bug/1442810 | 18:15 |
| openstack | Launchpad bug 1442810 in Ironic "Cleaning may leave nodes locked and require manual intervention to unlock" [High,In progress] - Assigned to Josh Gachnang (joshnang) | 18:15 |
| BadCub | https://review.openstack.org/#/c/171334/ | 18:16 |
| devananda | https://review.openstack.org/#/c/172582/ << fix cleaning | 18:16 |
| devananda | I think ^ is the only server-side patch that's blocking the RC | 18:16 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Update ilo drivers documentation for inspection https://review.openstack.org/170065 | 18:16 |
| NobodyCam | rloo: take a look and let me know what you think | 18:17 |
| rloo | BadCub: 171334 doesn't block rc1, does it? that's the client. | 18:17 |
| devananda | 171334 is a client side change which we need to unbreak stable/juno, and I'll handle getting that into stable/kilo after RC1 | 18:17 |
| rloo | NobodyCam: looking... | 18:17 |
| devananda | but that doesn't block our RC1 | 18:17 |
| devananda | rloo: you're correct | 18:17 |
| BadCub | devananda: that works | 18:17 |
| BadCub | so: https://review.openstack.org/#/c/172582/ should be our only blocker for rc1 release | 18:18 |
| devananda | yah | 18:19 |
| BadCub | unless the Tempest patch is an issue as well? | 18:19 |
| BadCub | but shouldn;t be | 18:19 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Open Liberty development https://review.openstack.org/171274 | 18:19 |
| devananda | BadCub: does not block our RC, but again, something we need to (work with tempest team to) land | 18:20 |
| BadCub | devananda: pinged them up several times last week :( | 18:21 |
| devananda | i dont see any feedback on the review from them | 18:21 |
| BadCub | nope | 18:21 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Local boot note about updated deploy ramdisk https://review.openstack.org/173029 | 18:22 |
| lucasagomes | devananda, jroll ^ | 18:22 |
| * lucasagomes not sure if that's the best way to word it | 18:22 | |
| jroll | woot | 18:22 |
| lucasagomes | feel free to update the patch | 18:22 |
| *** Sukhdev has quit IRC | 18:23 | |
| jroll | lucasagomes: maybe "since 2015/04/01" or something like that? | 18:23 |
| lucasagomes | jroll, "... in the kilo-era (since 2015/04/01)" ? | 18:24 |
| BadCub | brb | 18:24 |
| jroll | lucasagomes: yeah, something like that, don't know the actual date | 18:25 |
| devananda | dates are not that useful. how about pointing at a commit SHA ? | 18:25 |
| devananda | or a release # | 18:25 |
| lucasagomes | that was the patch https://review.openstack.org/#/c/169419/ | 18:25 |
| devananda | dib does releases to PIP | 18:25 |
| *** coolsvap|afk is now known as coolsvap | 18:26 | |
| *** trown|lunch is now known as trown | 18:26 | |
| lucasagomes | that's in ironic | 18:26 |
| lucasagomes | hmm in dib is | 18:26 |
| devananda | so I think it's reasonable to say "built with dib >= x.y.z" or some such | 18:26 |
| lucasagomes | https://review.openstack.org/#/c/169354/ | 18:26 |
| * lucasagomes checks last dib release | 18:27 | |
| lucasagomes | ok the last release seems to incoorporate this changes | 18:27 |
| lucasagomes | https://pypi.python.org/pypi/diskimage-builder | 18:27 |
| openstackgerrit | Alex Weeks proposed openstack/ironic-specs: Add implement-rescue-mode spec https://review.openstack.org/171878 | 18:27 |
| lucasagomes | 2015-04-09 | 18:27 |
| * lucasagomes updates the patch | 18:27 | |
| jroll | nice | 18:28 |
| devananda | JoshNang: only issue with https://review.openstack.org/#/c/172582/2/ironic/tests/conductor/test_manager.py,cm is that these are going to be logged as exceptions | 18:29 |
| devananda | JoshNang: given thta you seemed to be hitting those locks quite often when this was call(), I would imagine you'll start getting a lot of exceptions logged now | 18:29 |
| aweeks | jroll: JoshNang: I think I've addressed the comments on this spec https://review.openstack.org/#/c/171878/ | 18:29 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Local boot note about updated deploy ramdisk https://review.openstack.org/173029 | 18:29 |
| lucasagomes | here we go | 18:29 |
| devananda | JoshNang: even though they're being ignored and mostly benign, becaues the agent just retries | 18:29 |
| lucasagomes | I will take a break, have a good night everyone! | 18:29 |
| devananda | lucasagomes: cheers, gnight! | 18:30 |
| lucasagomes | devananda, jroll feel free to update that patch if needed | 18:30 |
| JoshNang | devananda: actually, i saw almost no lock conflicts with that change | 18:30 |
| *** lucasagomes is now known as lucas-dinner | 18:30 | |
| devananda | JoshNang: interesting | 18:30 |
| jroll | aweeks: lucas-dinner awesome, thanks | 18:30 |
| devananda | JoshNang: i'm really surprised. i wouldn't expect that to change the lock behavior *at all* | 18:31 |
| JoshNang | devananda: right. i think there's something deeper wrong here that i couldn't find | 18:32 |
| *** ijw_ has joined #openstack-ironic | 18:32 | |
| devananda | JoshNang: yea.... :-/ | 18:32 |
| * devananda +2's the change | 18:32 | |
| * NobodyCam +2 but hold off on +A so others can also look | 18:33 | |
| rloo | does anyone else want to look at 172582? Otherwise I'll +2. | 18:33 |
| rloo | ah, NobodyCam beat me to it. I'm going to +A then. | 18:34 |
| NobodyCam | rloo: ++ | 18:34 |
| BadCub | awesome | 18:34 |
| rloo | so we good now, for RC1? (After that merges of course.) | 18:34 |
| BadCub | rloo: we should be :-) | 18:34 |
| *** ijw has quit IRC | 18:34 | |
| NobodyCam | lol: http://www.merriam-webster.com/top-ten-lists/top-10-rare-and-amusing-insults-vol-2/loblolly.html | 18:35 |
| devananda | rloo: I believe so, yes | 18:35 |
| devananda | we still need to land the client fix ASAP | 18:35 |
| *** dttocs has quit IRC | 18:36 | |
| devananda | and I need to sort out how we're supposed to handle stable branches of the client in the next day or two | 18:36 |
| *** jxiaobin has joined #openstack-ironic | 18:36 | |
| devananda | (that fix needs to be on it) | 18:36 |
| *** coolsvap is now known as coolsvap|afk | 18:36 | |
| rloo | devananda: great. I've already +2'd the client. Need one more... | 18:36 |
| devananda | i held off on +2'ing my own code ... though at this point jlvillal has rewritten most of it :) | 18:36 |
| * NobodyCam I'll look now | 18:37 | |
| rloo | devananda: I dare you to +2 it! | 18:37 |
| BadCub | lol | 18:37 |
| jlvillal | I enjoyed working on it :) | 18:37 |
| * BadCub waves the shiny button at devananda | 18:38 | |
| *** Marga_ has quit IRC | 18:38 | |
| *** Marga_ has joined #openstack-ironic | 18:38 | |
| devananda | rloo: done | 18:38 |
| rloo | devananda: ha ha. awesome. thx! | 18:39 |
| * NobodyCam stops reviewing | 18:39 | |
| BadCub | awesome | 18:39 |
| *** dprince has quit IRC | 18:43 | |
| *** dttocs has joined #openstack-ironic | 18:44 | |
| *** meghal has quit IRC | 18:44 | |
| *** dttocs has quit IRC | 18:52 | |
| *** dttocs has joined #openstack-ironic | 18:53 | |
| NobodyCam | brb | 18:55 |
| *** dprince has joined #openstack-ironic | 18:59 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Update ilo drivers documentation for inspection https://review.openstack.org/170065 | 19:05 |
| *** ijw has joined #openstack-ironic | 19:07 | |
| *** meghal has joined #openstack-ironic | 19:07 | |
| *** ijw_ has quit IRC | 19:09 | |
| openstackgerrit | Jim Mankovich proposed openstack/ironic-specs: Send iLO health metrics to ceilometer https://review.openstack.org/127378 | 19:13 |
| *** jmank has joined #openstack-ironic | 19:13 | |
| openstackgerrit | Victor Lowther proposed openstack/ironic: Fix drac implementetation of set_boot_device. https://review.openstack.org/173047 | 19:19 |
| openstackgerrit | Merged openstack/ironic: Fix type of value in error middleware response header https://review.openstack.org/171544 | 19:21 |
| *** coolsvap|afk is now known as coolsvap | 19:21 | |
| devananda | odd - https://review.openstack.org/#/c/172036 | 19:25 |
| NobodyCam | whats the error there? | 19:27 |
| devananda | ahhh shit ... | 19:28 |
| *** Sukhdev has joined #openstack-ironic | 19:28 | |
| * devananda digs more | 19:28 | |
| NobodyCam | http://logs.openstack.org/36/172036/2/gate/gate-tempest-dsvm-ironic-agent_ssh/14006ac/logs/screen-n-cpu.txt.gz#_2015-04-13_18_37_10_125 ??? | 19:29 |
| NobodyCam | this looks strange ... thou I do not beleieve it is releated to the failure: http://logs.openstack.org/36/172036/2/gate/gate-tempest-dsvm-ironic-agent_ssh/14006ac/logs/ironic-bm-logs/baremetalbrbm_0_console.txt.gz | 19:33 |
| *** kkoski has quit IRC | 19:34 | |
| devananda | it looks like tempest failed to delete after successfully deleting a nova instance | 19:37 |
| devananda | but perhaps i'm reading something wrong | 19:37 |
| NobodyCam | devananda: here : http://logs.openstack.org/36/172036/2/gate/gate-tempest-dsvm-ironic-agent_ssh/14006ac/logs/screen-n-api.txt.gz#_2015-04-13_18_56_20_405 | 19:38 |
| devananda | there *does* appear to be an odd interaction around cleaning. or perhaps just one i haven't seen before | 19:38 |
| devananda | yea, that | 19:38 |
| NobodyCam | ya | 19:38 |
| NobodyCam | was looking at that | 19:38 |
| devananda | that delete was being retried internally by Nova at the time | 19:39 |
| devananda | actually nvm, not at that time | 19:39 |
| jroll | well, doesn't tempest try to delete, then poll the instance to watch for it disappearing? | 19:39 |
| devananda | http://logs.openstack.org/36/172036/2/gate/gate-tempest-dsvm-ironic-agent_ssh/14006ac/logs/screen-n-cpu.txt.gz?level=INFO#_2015-04-13_18_51_15_160 | 19:39 |
| jroll | 404 at the end seems fine | 19:39 |
| *** bradjones has quit IRC | 19:40 | |
| devananda | http://logs.openstack.org/36/172036/2/gate/gate-tempest-dsvm-ironic-agent_ssh/14006ac/console.html#_2015-04-13_18_56_21_743 | 19:41 |
| rloo | is it possible it timed out cuz cleaning took too long? although there wasn't anything to clean, right? | 19:42 |
| NobodyCam | the patch to turn off disk eraseing landed no? | 19:43 |
| JoshNang | it did | 19:43 |
| JoshNang | it took 5 mins for the agent to boot and heartbeat | 19:43 |
| JoshNang | http://logs.openstack.org/36/172036/2/gate/gate-tempest-dsvm-ironic-agent_ssh/14006ac/logs/screen-ir-cond.txt.gz#_2015-04-13_18_56_22_779 | 19:43 |
| JoshNang | powered on at http://logs.openstack.org/36/172036/2/gate/gate-tempest-dsvm-ironic-agent_ssh/14006ac/logs/screen-ir-cond.txt.gz#_2015-04-13_18_51_39_230 | 19:43 |
| devananda | yea -- http://logs.openstack.org/36/172036/2/gate/gate-tempest-dsvm-ironic-agent_ssh/14006ac/logs/screen-ir-cond.txt.gz#_2015-04-13_18_56_22_779 | 19:43 |
| devananda | wow :( | 19:43 |
| devananda | during that time tempest is still waiting for the node to free up, as it should | 19:44 |
| BadCub | ick | 19:44 |
| NobodyCam | so is that just more of: https://bugs.launchpad.net/openstack-ci/+bug/1311066 | 19:46 |
| openstack | Launchpad bug 1311066 in OpenStack-Gate "Some nodes allocated in node pool are very very slow" [High,Confirmed] | 19:46 |
| openstackgerrit | Merged openstack/python-ironicclient: Client should fall back to the lower versions if necessary https://review.openstack.org/171334 | 19:48 |
| NobodyCam | woo hoo | 19:48 |
| *** coolsvap is now known as coolsvap|afk | 19:48 | |
| BadCub | yeah | 19:49 |
| BadCub | the last one approved, first one merged lol | 19:49 |
| devananda | http://paste.openstack.org/show/203712/ | 19:49 |
| NobodyCam | wow lines #6 and #9 show the same amount of time | 19:51 |
| NobodyCam | almost | 19:52 |
| devananda | so our current tempest test goes through 3 PXE boot cycles, each of which takes ~ 5 minutes | 19:53 |
| devananda | before cleaning, it only did 2 cycles | 19:53 |
| NobodyCam | yep | 19:53 |
| devananda | did anyone bump the timeout in tempest test for the delete? | 19:54 |
| NobodyCam | I though JoshNang put up a patch to change the time for state transations in genrenal | 19:54 |
| JoshNang | yeah, sec, looking for the patch | 19:54 |
| devananda | it's still 60 | 19:55 |
| JoshNang | https://review.openstack.org/#/c/168175/ | 19:55 |
| JoshNang | ah, this is the unprovision timeout | 19:55 |
| devananda | yes | 19:55 |
| *** Marga_ has quit IRC | 19:56 | |
| devananda | http://git.openstack.org/cgit/openstack/tempest/tree/tempest/config.py#n1047 | 19:56 |
| *** achanda has quit IRC | 19:57 | |
| *** coolsvap has joined #openstack-ironic | 19:57 | |
| NobodyCam | JoshNang: are you tossing up a patch for that? | 19:57 |
| devananda | I got it | 19:57 |
| NobodyCam | TY devananda | 19:57 |
| NobodyCam | :) | 19:58 |
| JoshNang | is that not what 168175 is doing? | 19:58 |
| openstackgerrit | Merged openstack/ironic: iLO driver documentation for node cleaning https://review.openstack.org/169162 | 19:58 |
| NobodyCam | JoshNang: humm: https://review.openstack.org/#/c/168175/5/lib/tempest | 20:00 |
| NobodyCam | maybe 300 is not enough? rev four had it at 600 | 20:01 |
| NobodyCam | devananda: JoshNang: https://review.openstack.org/#/c/168175/4..5/lib/tempest | 20:02 |
| devananda | JoshNang: ah, you're right - that is raising the value | 20:02 |
| JoshNang | i put it at 600 because i had no idea how long it would take | 20:02 |
| openstackgerrit | ayoung proposed openstack/ironic: Fix drac implementation of set_boot_device. https://review.openstack.org/173047 | 20:02 |
| JoshNang | i thought 10 mins seemed ridiculous...but maybe not :( | 20:02 |
| * NobodyCam has worked on some hardware that took > 10 minutes to post :-p | 20:04 | |
| devananda | apparently not. in this gate job, cleaning didn't start until 5:03 after the DELETE call was finished | 20:04 |
| devananda | 2015-04-13 18:56:21.743 | Details: Timed out waiting for node 03794d6d-47af-4f3e-8932-67be02bce9ed to reach provision_state state(s) [None] | 20:04 |
| devananda | 2015-04-13 18:56:22.779 16346 INFO ironic.conductor.manager [-] Executing cleaning on node 03794d6d-47af-4f3e-8932-67be02bce9ed, remaining steps: [] | 20:05 |
| devananda | should we raise all the timeouts from 5 to 10 minutes? | 20:07 |
| devananda | this one just happened to take a tiny bit longer than most ... | 20:07 |
| NobodyCam | thats prob the safe thing to do. | 20:08 |
| JoshNang | ++, it was definitely quite close to finishing. | 20:09 |
| rloo | that test didn't just start timing out today though? | 20:09 |
| devananda | rloo: i have not checked logstash | 20:09 |
| JoshNang | its 3-4ish minutes usually, and the timeout is 5 | 20:09 |
| NobodyCam | could it be load on the infa vm's? didn't most projects open L today | 20:09 |
| devananda | NobodyCam: those are just rax / hpcloud instances ... | 20:10 |
| rloo | i've seen a few 'recheck' comments in some patches, but I didn't look to see why. | 20:10 |
| *** lucas-dinner has quit IRC | 20:10 | |
| NobodyCam | ahh | 20:10 |
| * devananda needs to step away soon and get lunch | 20:11 | |
| NobodyCam | mmmm food | 20:11 |
| openstackgerrit | Merged openstack/ironic: Convert internal RPC continue_node_cleaning to a "cast" https://review.openstack.org/172582 | 20:11 |
| BadCub | food sounds yummy | 20:11 |
| * NobodyCam should go pick up his glasses | 20:12 | |
| BadCub | looks like we are good for rc1 :) | 20:12 |
| NobodyCam | nice! | 20:12 |
| devananda | rloo: looks like only one hit in the last 7d for timed out waiting on delete | 20:13 |
| NobodyCam | awesome woke EVERYONE! | 20:13 |
| BadCub | ^ was the last pending | 20:13 |
| NobodyCam | work even :-p | 20:13 |
| JoshNang | \o/ | 20:13 |
| devananda | however, we've got 16 hits for ironic jobs failing due to timeout waiting to reach some state or other, in the last 7 days | 20:13 |
| rloo | devananda: good to know. thx. | 20:13 |
| rloo | devananda: so reaching some state that isn't due to delete :-( | 20:14 |
| rloo | I thought there was a bug about timeouts. | 20:15 |
| devananda | rloo: however, some of those logstash entries are probably other errors masquerading as timeouts | 20:15 |
| devananda | i picked one at random -- it's actually a failure in a change in DIB. ironic never got the callback, sothe deploy timed out | 20:15 |
| rloo | https://bugs.launchpad.net/ironic/+bug/1408067 | 20:16 |
| openstack | Launchpad bug 1408067 in Ironic "ironic jobs sometimes fail with node callback timeout" [Medium,Triaged] | 20:16 |
| devananda | well ... that actually could the the same thing, reallly | 20:16 |
| devananda | yea | 20:16 |
| devananda | things like this -- http://logs.openstack.org/15/172715/2/check/check-tempest-dsvm-ironic-pxe_ssh/efef2e0/logs/ironic-bm-logs/baremetalbrbm_0_console.txt.gz | 20:16 |
| devananda | anyone create an e-r query for that? | 20:18 |
| rloo | hmm, i wonder if this is also related: https://bugs.launchpad.net/ironic/+bug/1441007 | 20:19 |
| openstack | Launchpad bug 1441007 in devstack "Gate failure: Timed out waiting for Nova hypervisor-stats count >= 1" [Undecided,In progress] | 20:19 |
| devananda | rloo: no, i think that's related to the order that services are started in devstack | 20:20 |
| *** kkoski has joined #openstack-ironic | 20:20 | |
| devananda | adam_g was working on something there last week IIRC | 20:20 |
| devananda | if ironic-api is started too late, nova-compute will bail. has nothing to do with deploys timing out, if it's the bug I'm thinking of | 20:21 |
| * devananda goes in search of lunch | 20:21 | |
| *** Marga_ has joined #openstack-ironic | 20:24 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Update ilo drivers documentation for inspection https://review.openstack.org/170065 | 20:24 |
| *** achanda has joined #openstack-ironic | 20:25 | |
| NobodyCam | rloo: and that one will render correctlly :-p | 20:26 |
| openstackgerrit | Victor Lowther proposed openstack/ironic: Fix drac implementation of set_boot_device. https://review.openstack.org/173047 | 20:27 |
| devananda | commented on https://bugs.launchpad.net/ironic/+bug/1408067 and changed to High | 20:27 |
| openstack | Launchpad bug 1408067 in Ironic "ironic jobs sometimes fail with node callback timeout" [High,Triaged] | 20:27 |
| devananda | BadCub: ^ something to start tracking, given that I've seen 16 failures matching this in 7 days .... | 20:28 |
| * devananda really lunches now | 20:30 | |
| NobodyCam | :) | 20:30 |
| NobodyCam | go eat! | 20:30 |
| *** kozhukalov has quit IRC | 20:30 | |
| *** jgrimm is now known as zz_jgrimm | 20:33 | |
| *** HenryG has quit IRC | 20:33 | |
| rloo | NobodyCam: ok, i'll wait for it to render first, before looking ;) | 20:35 |
| NobodyCam | rloo: ++ :) Thank you | 20:35 |
| BadCub | devananda: got it | 20:36 |
| *** HenryG has joined #openstack-ironic | 20:37 | |
| *** kkoski has quit IRC | 20:38 | |
| *** kkoski has joined #openstack-ironic | 20:39 | |
| *** mjturek1 has quit IRC | 20:52 | |
| *** andreykurilin has quit IRC | 20:55 | |
| trown | do LOG.debug messages need internationalization? | 20:56 |
| rloo | trown: don't need | 20:58 |
| rloo | trown: this might help: http://docs.openstack.org/developer/oslo.i18n/guidelines.html | 21:00 |
| trown | thanks rloo | 21:03 |
| *** trown is now known as trown|outttypeww | 21:04 | |
| *** Marga_ has quit IRC | 21:09 | |
| *** Marga_ has joined #openstack-ironic | 21:10 | |
| *** andreykurilin has joined #openstack-ironic | 21:20 | |
| *** Sukhdev has quit IRC | 21:28 | |
| *** mrda-away is now known as mrda | 21:33 | |
| mrda | Morning | 21:33 |
| *** mjturek1 has joined #openstack-ironic | 21:34 | |
| jlvillal | mrda, Good morning | 21:35 |
| mrda | o/ | 21:36 |
| devananda | victor_lowther: looks like you proposed 173047 without a bug tagged. mind filing one so we can track that? | 21:45 |
| devananda | victor_lowther: also, how serious is that? | 21:45 |
| devananda | I'm about to cut Kilo, so if there's some critical bug in the DRAC driver ... it'd be nice to know that now :) | 21:46 |
| NobodyCam | morning mrda | 21:47 |
| BadCub | hiya mrda | 21:47 |
| victor_lowther | devananda: It will cause Ironic to error out if set_boot_device is called against the pxe_drac driver with the device that the system is already set to. | 21:48 |
| devananda | victor_lowther: I don't know off hand how to guage the impact of that. Does it happen during every deploy? or only when the user manually invokes set-boot-device? | 21:49 |
| victor_lowther | I know that tripleo does that. | 21:49 |
| victor_lowther | It happens in every deploy that the scripts do not check the boot device to see if it needs changing before trying to change the boot device. :) | 21:49 |
| devananda | I don't believe the Nova driver does that at all | 21:50 |
| victor_lowther | The non-randisk parts of idonic-discoverd do. | 21:51 |
| pshige | mrda: morning | 21:51 |
| victor_lowther | er, ironic | 21:51 |
| victor_lowther | if I remember the sequence of events that led me to find the patch properly. | 21:51 |
| devananda | that would make sense | 21:52 |
| *** Marga_ has quit IRC | 21:52 | |
| devananda | victor_lowther: seems like a reasonable backport candidate to me. quite small and contained within the driver. also needs unit tests :) | 21:53 |
| victor_lowther | I wouldn't call it critical, but i knew exactly what was failing in the drac driver, and I am less familiar with the higher-level codepaths. | 21:53 |
| victor_lowther | The IPMI and other drivers probably just don't care, but the peculiarities of the drac and the lifecycle controller make it a little more brittle about accepting non-changes like that. | 21:54 |
| devananda | victor_lowther: as you can imagine, every driver seems to have some unique quirks in this area ;) | 21:55 |
| devananda | i think AMT only accepts the command the first time you send it | 21:55 |
| NobodyCam | devananda: yep | 21:55 |
| devananda | any request to change boot device after the first one, and before the next reboot, is ignored | 21:55 |
| victor_lowther | AMT uses sorta the same WSMAN braindamange that drac does, so that makes sense. | 21:55 |
| victor_lowther | IIRC. | 21:56 |
| *** Marga_ has joined #openstack-ironic | 21:56 | |
| *** Marga_ has quit IRC | 21:57 | |
| *** kkoski has quit IRC | 22:03 | |
| *** andreykurilin has quit IRC | 22:05 | |
| rloo | NobodyCam: wrt 170065. Sorry. Is this something we wanted to get into rc1? (I don't think it is necessary.) | 22:07 |
| NobodyCam | rloo: it a doc, not a blocker, but nice to have I would think | 22:08 |
| rloo | NobodyCam: well, your or someone else's call. use what is in the patch now even though it isn't great, or fix, or defer til after rc1. | 22:09 |
| devananda | the patch to open liberty is in the gate, btw | 22:09 |
| BadCub | nice | 22:09 |
| devananda | once 171274 merges, RC is official and lierty is open | 22:10 |
| NobodyCam | devananda: w00t | 22:10 |
| NobodyCam | rloo: humm let me take a look | 22:10 |
| devananda | time for us to start writing some real release notes | 22:14 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Update ilo drivers documentation for inspection https://review.openstack.org/170065 | 22:35 |
| *** rfchapman has joined #openstack-ironic | 22:47 | |
| *** Marga_ has joined #openstack-ironic | 22:49 | |
| *** mtanino has quit IRC | 22:51 | |
| *** coolsvap has quit IRC | 22:51 | |
| *** dprince has quit IRC | 22:56 | |
| *** mtanino has joined #openstack-ironic | 23:04 | |
| *** chlong has joined #openstack-ironic | 23:15 | |
| openstackgerrit | Pierre Riteau proposed openstack/python-ironicclient: Add support for filtering nodes by chassis UUID https://review.openstack.org/167964 | 23:21 |
| openstackgerrit | Pierre Riteau proposed openstack/python-ironicclient: Add support for filtering nodes by chassis UUID https://review.openstack.org/167964 | 23:21 |
| *** jamielennox|away is now known as jamielennox | 23:22 | |
| openstackgerrit | Merged openstack/ironic: Cleanup DHCPFactory._dhcp_provider after tests https://review.openstack.org/172036 | 23:25 |
| openstackgerrit | Merged openstack/ironic: Local boot note about updated deploy ramdisk https://review.openstack.org/173029 | 23:25 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Update ilo drivers documentation for inspection https://review.openstack.org/170065 | 23:31 |
| *** yuanying has joined #openstack-ironic | 23:39 | |
| *** r-daneel has quit IRC | 23:43 | |
| *** saripurigopi has joined #openstack-ironic | 23:51 | |
| *** saripurigopi has quit IRC | 23:53 | |
| *** Haomeng has joined #openstack-ironic | 23:56 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Check temp dir is usable for ipmitool driver https://review.openstack.org/160383 | 23:58 |
| *** Haomeng|2 has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!