| *** igordcard has quit IRC | 00:21 | |
| *** matsuhashi has joined #openstack-ironic | 00:35 | |
| *** matsuhashi has quit IRC | 00:35 | |
| *** matsuhashi has joined #openstack-ironic | 00:35 | |
| *** igordcard has joined #openstack-ironic | 00:48 | |
| *** nosnos has joined #openstack-ironic | 00:55 | |
| *** igordcard has quit IRC | 01:19 | |
| *** Haomeng has quit IRC | 01:36 | |
| *** blamar has joined #openstack-ironic | 01:45 | |
| *** pcrews has quit IRC | 01:45 | |
| *** blamar has quit IRC | 01:47 | |
| *** Haomeng has joined #openstack-ironic | 01:52 | |
| *** nosnos has quit IRC | 03:36 | |
| lifeless | NobodyCam: any chance you could take over 98050 for me ? | 03:52 |
|---|---|---|
| openstackgerrit | lifeless proposed a change to openstack/ironic: Draft a 'pxe_logpower' driver https://review.openstack.org/98950 | 03:54 |
| *** pcrews has joined #openstack-ironic | 04:03 | |
| *** pcrews has quit IRC | 04:09 | |
| *** nosnos has joined #openstack-ironic | 04:15 | |
| *** k4n0 has joined #openstack-ironic | 04:19 | |
| openstackgerrit | A change was merged to openstack/python-ironicclient: Add UTF-8 coding lines to all Python files https://review.openstack.org/96109 | 04:31 |
| *** eghobo has joined #openstack-ironic | 04:31 | |
| *** rakesh_hs has joined #openstack-ironic | 04:40 | |
| *** coolsvap|afk is now known as coolsvap | 04:43 | |
| *** coolsvap is now known as coolsvap|afk | 04:58 | |
| *** shausy has joined #openstack-ironic | 05:01 | |
| *** sabah has joined #openstack-ironic | 05:05 | |
| *** coolsvap|afk is now known as coolsvap | 05:09 | |
| *** lazy_prince has joined #openstack-ironic | 05:10 | |
| *** Nisha has joined #openstack-ironic | 05:11 | |
| *** Nisha has quit IRC | 05:13 | |
| *** rakesh_hs has quit IRC | 05:13 | |
| *** Nisha has joined #openstack-ironic | 05:13 | |
| *** rakesh_hs has joined #openstack-ironic | 05:20 | |
| *** radsy has quit IRC | 05:22 | |
| *** matsuhashi has quit IRC | 06:00 | |
| *** matsuhashi has joined #openstack-ironic | 06:01 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/96063 | 06:02 |
| *** lazy_prince has quit IRC | 06:03 | |
| *** rakesh_hs has quit IRC | 06:04 | |
| *** matsuhas_ has joined #openstack-ironic | 06:38 | |
| *** coolsvap is now known as coolsvap|afk | 06:38 | |
| *** coolsvap|afk is now known as coolsvap | 06:38 | |
| *** jcoufal has joined #openstack-ironic | 06:38 | |
| *** sabah has quit IRC | 06:38 | |
| *** matsuhashi has quit IRC | 06:38 | |
| *** k4n0 has quit IRC | 06:55 | |
| *** eghobo has quit IRC | 06:57 | |
| *** matsuhas_ has quit IRC | 07:01 | |
| *** Mikhail_D_ltp has joined #openstack-ironic | 07:01 | |
| *** matsuhashi has joined #openstack-ironic | 07:04 | |
| *** Mikhail_D_ltp has left #openstack-ironic | 07:05 | |
| *** k4n0 has joined #openstack-ironic | 07:05 | |
| *** coolsvap is now known as coolsvap|afk | 07:06 | |
| *** ifarkas has joined #openstack-ironic | 07:07 | |
| *** Mikhail_D_ltp has joined #openstack-ironic | 07:07 | |
| *** rameshg87 has joined #openstack-ironic | 07:09 | |
| *** Mikhail_D_ltp has quit IRC | 07:13 | |
| *** Mikhail_D_ltp has joined #openstack-ironic | 07:13 | |
| *** Mikhail_D_ltp has quit IRC | 07:15 | |
| *** Faizan has joined #openstack-ironic | 07:16 | |
| *** coolsvap|afk is now known as coolsvap | 07:17 | |
| *** Mikhail_D_ltp has joined #openstack-ironic | 07:27 | |
| *** ndipanov has joined #openstack-ironic | 07:32 | |
| *** athomas has joined #openstack-ironic | 07:32 | |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic: Fix workaround for the "device is busy" problem https://review.openstack.org/90126 | 07:39 |
| *** dtantsur|pto is now known as dtantsur | 07:42 | |
| dtantsur | Morning, Ironic \o/ How were things last week? | 07:42 |
| ifarkas | morning dtantsur and all | 07:43 |
| ifarkas | dtantsur, it was just as usual ;-) how was your vacation? | 07:43 |
| dtantsur | ifarkas, very good! though I didn't expect Prague to be so crowded | 07:44 |
| ifarkas | dtantsur, oh yeah, it's always very packed | 07:45 |
| *** coolsvap is now known as coolsvap|afk | 07:46 | |
| *** Haomeng has quit IRC | 07:48 | |
| *** Haomeng has joined #openstack-ironic | 07:50 | |
| *** rakesh_hs has joined #openstack-ironic | 07:53 | |
| openstackgerrit | Dmitry Tantsur proposed a change to openstack/ironic: Add some real-world testing on DiskPartitioner https://review.openstack.org/94620 | 07:54 |
| GheRivero | morning all | 07:54 |
| *** coolsvap|afk is now known as coolsvap | 07:54 | |
| *** lazy_prince has joined #openstack-ironic | 07:54 | |
| dtantsur | GheRivero, morning | 07:55 |
| Haomeng | morning all:) | 07:56 |
| *** Mikhail_D_ltp has quit IRC | 07:56 | |
| *** jistr has joined #openstack-ironic | 07:57 | |
| dtantsur | Haomeng, morning | 07:58 |
| Haomeng | dtantsur: :) | 07:59 |
| openstackgerrit | Dmitry Tantsur proposed a change to openstack/ironic: Add some real-world testing on DiskPartitioner https://review.openstack.org/94620 | 08:01 |
| *** foexle has joined #openstack-ironic | 08:03 | |
| *** martyntaylor has joined #openstack-ironic | 08:21 | |
| *** lucasagomes has joined #openstack-ironic | 08:28 | |
| Mikhail_D_wk | Good morning all :) | 08:30 |
| *** loki184 has joined #openstack-ironic | 08:31 | |
| dtantsur | Mikhail_D_wk, morning | 08:32 |
| *** vinbs has joined #openstack-ironic | 08:42 | |
| *** coolsvap is now known as coolsvap|afk | 08:47 | |
| *** pelix has joined #openstack-ironic | 08:48 | |
| *** sysexit has joined #openstack-ironic | 08:51 | |
| *** viktors has joined #openstack-ironic | 08:53 | |
| *** k4n0 has quit IRC | 08:53 | |
| *** k4n0 has joined #openstack-ironic | 08:53 | |
| *** coolsvap|afk is now known as coolsvap | 08:54 | |
| *** athomas has quit IRC | 08:58 | |
| *** athomas has joined #openstack-ironic | 09:02 | |
| *** rakesh_hs has quit IRC | 09:05 | |
| *** rakesh_hs has joined #openstack-ironic | 09:07 | |
| *** loki184 has quit IRC | 09:07 | |
| openstackgerrit | A change was merged to openstack/ironic: Driver interface's validate should return nothing https://review.openstack.org/97855 | 09:15 |
| mrda | hey Ironic - just confirming, are we expecting a team meeting in 10 hours or so time? | 09:16 |
| mrda | the wiki hasn't been updated, that's all | 09:16 |
| dtantsur | mrda, hi! I'm expecting, but I'm not the best source of information right now :) | 09:17 |
| mrda | sure, I underastand | 09:17 |
| mrda | I'll take the punt and wake up early :) | 09:17 |
| openstackgerrit | Phil Day proposed a change to openstack/ironic: Remove node paramter from power_off() method in virt driver https://review.openstack.org/100197 | 09:18 |
| openstackgerrit | Phil Day proposed a change to openstack/ironic: Add clean shutdown parameters to power_off() https://review.openstack.org/99932 | 09:20 |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Perform manual error checks instead of set -e https://review.openstack.org/97860 | 09:27 |
| *** romcheg has joined #openstack-ironic | 09:38 | |
| *** Haomeng has quit IRC | 09:44 | |
| openstackgerrit | Ghe Rivero proposed a change to openstack/ironic-python-agent: Remove duplicated test for caching image https://review.openstack.org/99230 | 09:46 |
| *** Haomeng has joined #openstack-ironic | 09:46 | |
| mrda | Night ironic | 09:47 |
| romcheg | g'night mrda! | 09:47 |
| *** mrda is now known as mrda-away | 09:47 | |
| romcheg | See you today :D | 09:47 |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ironic: Updated from global requirements https://review.openstack.org/96228 | 09:47 |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ironic-python-agent: Updated from global requirements https://review.openstack.org/88722 | 09:47 |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/python-ironicclient: Updated from global requirements https://review.openstack.org/96263 | 09:52 |
| *** sabah has joined #openstack-ironic | 09:52 | |
| vinbs | Hello Ironic! | 09:59 |
| *** coolsvap is now known as coolsvap|afk | 09:59 | |
| *** Haomeng|2 has joined #openstack-ironic | 10:00 | |
| *** Haomeng has quit IRC | 10:01 | |
| lucasagomes | :) morning all | 10:02 |
| lucasagomes | dtantsur, welcome back | 10:02 |
| *** openstackgerrit has quit IRC | 10:06 | |
| romcheg | Morning lucasagomes, dtantsur! | 10:06 |
| *** openstackgerrit has joined #openstack-ironic | 10:07 | |
| openstackgerrit | Ghe Rivero proposed a change to openstack/ironic-python-agent: Improve cache_image testing https://review.openstack.org/99301 | 10:07 |
| openstackgerrit | Mikhail Durnosvistov proposed a change to openstack/ironic: Cleanup mock patch without `with` part 3 https://review.openstack.org/73256 | 10:07 |
| openstackgerrit | Mikhail Durnosvistov proposed a change to openstack/ironic: Get rid of the newline "\" https://review.openstack.org/66793 | 10:07 |
| *** Nisha has quit IRC | 10:09 | |
| *** coolsvap|afk is now known as coolsvap | 10:13 | |
| *** rakesh_hs has quit IRC | 10:19 | |
| *** matsuhashi has quit IRC | 10:24 | |
| *** matsuhashi has joined #openstack-ironic | 10:24 | |
| lucasagomes | lifeless, thanks for the review, I left a comment at #99426 | 10:26 |
| *** matsuhashi has quit IRC | 10:33 | |
| *** matsuhashi has joined #openstack-ironic | 10:34 | |
| *** openstackgerrit has quit IRC | 10:35 | |
| *** openstackgerrit has joined #openstack-ironic | 10:36 | |
| *** max_lobur has joined #openstack-ironic | 10:36 | |
| *** matsuhashi has quit IRC | 10:41 | |
| *** matsuhashi has joined #openstack-ironic | 10:41 | |
| openstackgerrit | A change was merged to openstack/ironic: Let ipmitool natively retry commands https://review.openstack.org/96902 | 10:47 |
| *** Haomeng|2 has quit IRC | 10:47 | |
| *** igordcard has joined #openstack-ironic | 10:48 | |
| dtantsur | morning, lucasagomes, romcheg, vinbs! | 10:49 |
| *** Haomeng|2 has joined #openstack-ironic | 10:49 | |
| *** shausy has quit IRC | 10:50 | |
| *** rakesh_hs has joined #openstack-ironic | 10:51 | |
| *** shausy has joined #openstack-ironic | 10:51 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic-specs: Promote set_boot_device() from vendor_passtrhu https://review.openstack.org/100218 | 10:59 |
| openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Replace mknod() with chmod() https://review.openstack.org/99484 | 11:14 |
| *** nosnos has quit IRC | 11:15 | |
| *** sabah has quit IRC | 11:16 | |
| openstackgerrit | Dmitry Tantsur proposed a change to openstack/ironic: Test for membership should be 'not in' https://review.openstack.org/97412 | 11:19 |
| *** matsuhashi has quit IRC | 11:29 | |
| *** matsuhashi has joined #openstack-ironic | 11:29 | |
| *** matsuhashi has quit IRC | 11:33 | |
| vinbs | Hello Ironic, I'm coming across an error at pxe deploy stage | 11:37 |
| vinbs | I'm trying to launch an instance on real hardware with openstack ironic | 11:38 |
| vinbs | my baremetal server gets an ip address assigned and also I can see all the devices on the server getting initialized | 11:39 |
| *** lucasagomes is now known as lucas-hungry | 11:40 | |
| vinbs | but after this stage "request Ironic API to deploy image" | 11:40 |
| vinbs | which gets accepted | 11:40 |
| vinbs | I get this error on the baremetal server: | 11:40 |
| vinbs | daemon.err tgtd: iser_ib_init(3349) Failed to intialize RDMA | 11:41 |
| *** coolsvap is now known as coolsvap|afk | 11:41 | |
| vinbs | after this the server goes into reboot and on the openstack side the launch instance state goes into running | 11:44 |
| vinbs | and openstack sees it as successful deploy of OS | 11:44 |
| vinbs | but my server after reboot goes into pxe boot again since OS wasn't deployed | 11:45 |
| rameshg87 | vinbs: after pxe deploy, the node does pxe boot again. so node doing pxe boot again after deployment is expected | 11:50 |
| rameshg87 | vinbs, do you see the error "Failed to intialize RDMA" on the baremetal ? | 11:51 |
| *** lazy_prince has quit IRC | 11:53 | |
| dtantsur | GheRivero, could you give your updates on https://bugs.launchpad.net/ironic/+bug/1198813 and https://bugs.launchpad.net/ironic/+bug/1199532 ? | 11:54 |
| *** jdob has joined #openstack-ironic | 11:54 | |
| dtantsur | You've been assigned for quite a while | 11:54 |
| dtantsur | GheRivero, ah, and https://bugs.launchpad.net/ironic/+bug/1199536 | 11:54 |
| vinbs | rameshg87, yes I see the error "Failed to intialize RDMA" on the baremetal server | 11:54 |
| vinbs | and here's the ironic conductor log http://paste.openstack.org/show/84158/ | 11:55 |
| vinbs | I can see errors related to ISCSI, but I don't completely understand what the error means | 11:55 |
| rameshg87 | vinbs, lines 26-38 on the copied output | 11:59 |
| rameshg87 | vinbs, it seems sfdisk command failed on the ironic conductor node. | 12:00 |
| rameshg87 | vinbs: is sfdisk installed on the conductor node, btw ? | 12:00 |
| vinbs | Rameshg87, yes it is installed | 12:01 |
| rameshg87 | vinbs, not saying that is the reason, but you might be running an older version of ironic. the new ironic has changed to use the parted utility | 12:02 |
| rameshg87 | instead of sfdisk | 12:02 |
| vinbs | rameshg87, is there a way I can share a screenshot? I can show you the error which is thrown on the baremetal server? | 12:03 |
| vinbs | rameshg87, oh.. didn't know that. But it should work fine with sfdisk utility too right? | 12:03 |
| rameshg87 | vinbs, yes, it should work fine with sfdisk utility as well | 12:03 |
| rameshg87 | vinbs, can you just run this command on your conductor node to see if everything is proper with rootwrap and sfdisk: sudo ironic-rootwrap /etc/ironic/rootwrap.conf sfdisk -l | 12:05 |
| vinbs | rameshg87, I'm logged in as root, so I would be running it without the sudo. | 12:06 |
| rameshg87 | vinbs, hmm okay give a try as root first first, but i would suggest to run as non-root user and see if ironic-rootwrap succeeds | 12:07 |
| *** rakesh_hs has quit IRC | 12:07 | |
| GheRivero | dtantsur: sure. will do later today | 12:08 |
| vinbs | rameshg87, here's the output from that command http://paste.openstack.org/show/84160/ | 12:08 |
| dtantsur | GheRivero, thanks! | 12:08 |
| vinbs | I'm getting an error | 12:08 |
| *** rakesh_hs has joined #openstack-ironic | 12:10 | |
| rameshg87 | vinbs, that error seems fine .. i was just wondering if ironic-conductor wasn't able to execute sfdisk at all | 12:11 |
| rameshg87 | vinbs, does it fail only for this machine ? | 12:11 |
| vinbs | rameshg87, I haven't tried with any other machine yet | 12:11 |
| rameshg87 | vinbs, i suggest to try with the latest code, but i can't assure if it will make any difference .. it might be a disk read/write error or something else that causes sfdisk to fail | 12:13 |
| *** rakesh_hs has quit IRC | 12:14 | |
| vinbs | rameshg87, okay.. let me try with the latest code | 12:14 |
| rameshg87 | vinbs, may be it will give a better clue | 12:14 |
| *** rakesh_hs has joined #openstack-ironic | 12:14 | |
| *** romcheg has left #openstack-ironic | 12:15 | |
| *** romcheg has joined #openstack-ironic | 12:15 | |
| rameshg87 | vinbs, now it just says sfdisk returned 1 | 12:15 |
| vinbs | rameshg87, I'm using openstack setup not a devstack setup | 12:15 |
| *** bmahalakshmi has joined #openstack-ironic | 12:15 | |
| vinbs | rameshg87, how would I be upgrading my code? | 12:15 |
| vinbs | rameshg87, a very basic thing which I don't know yet :) | 12:16 |
| rameshg87 | vinbs, how did you install ironic btw | 12:18 |
| vinbs | rameshg87, I have an all-in-one openstack icehouse setup | 12:19 |
| *** bmahalakshmi has quit IRC | 12:19 | |
| vinbs | rameshg87, and i followed this guide to setup ironic http://docs.openstack.org/developer/ironic/install/ironic-install.html#configure-compute-service-to-use-the-bare-metal-service | 12:19 |
| *** igordcard has quit IRC | 12:19 | |
| *** bmahalakshmi has joined #openstack-ironic | 12:20 | |
| *** bmahalakshmi has quit IRC | 12:20 | |
| rameshg87 | vinbs, i am not sure - you might want to check with others as well, but you can download the tarball from here that was released last week juno-1: https://launchpad.net/ironic | 12:20 |
| vinbs | rameshg87, ok I will check that out | 12:21 |
| vinbs | rameshg87, thanks a bunch :) | 12:21 |
| rameshg87 | vinbs, :-) | 12:21 |
| rameshg87 | lucasagomes, dtantsur, a request to a look at the review https://review.openstack.org/#/c/89500/ | 12:22 |
| *** maha2 has joined #openstack-ironic | 12:22 | |
| *** maha2 has quit IRC | 12:23 | |
| dtantsur | will look in a few minutes | 12:23 |
| *** maha2 has joined #openstack-ironic | 12:23 | |
| *** romcheg has left #openstack-ironic | 12:23 | |
| *** romcheg has joined #openstack-ironic | 12:23 | |
| *** maha2 has quit IRC | 12:23 | |
| *** bmaha has joined #openstack-ironic | 12:24 | |
| *** vinbs has quit IRC | 12:26 | |
| *** martyntaylor has quit IRC | 12:30 | |
| *** linggao has joined #openstack-ironic | 12:37 | |
| *** lucas-hungry is now known as lucasagomes | 12:39 | |
| *** k4n0 has quit IRC | 12:40 | |
| *** shausy has quit IRC | 12:41 | |
| *** petertoft has joined #openstack-ironic | 12:43 | |
| *** shausy has joined #openstack-ironic | 12:45 | |
| *** martyntaylor has joined #openstack-ironic | 12:45 | |
| *** bmaha has quit IRC | 12:50 | |
| rameshg87 | hello dtantsur | 12:53 |
| rameshg87 | dtantsur: thanks for the review. i wanted clarification on one comment | 12:53 |
| *** k4n0 has joined #openstack-ironic | 12:54 | |
| dtantsur | rameshg87, sure | 12:58 |
| rameshg87 | dtantsur, this is regarding the comment on line 54 on this file: https://review.openstack.org/#/c/89500/16/ironic/tests/drivers/ilo/test_common.py | 13:02 |
| NobodyCam | Good Morning iRonic | 13:03 |
| rameshg87 | dtantsur, this was done to avoid skipping tests as per on the previous patch, PS15 | 13:03 |
| rameshg87 | good morning NobodyCam :-) | 13:03 |
| dtantsur | rameshg87, is is good, but mocking should not be done this way | 13:03 |
| NobodyCam | :) | 13:03 |
| dtantsur | NobodyCam, morning :) | 13:03 |
| NobodyCam | morning rameshg87 and dtantsur :) | 13:04 |
| *** jdob has quit IRC | 13:04 | |
| *** jdob has joined #openstack-ironic | 13:04 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic-specs: Promote set_boot_device() from vendor_passtrhu https://review.openstack.org/100218 | 13:04 |
| *** rloo has joined #openstack-ironic | 13:04 | |
| *** rloo has quit IRC | 13:04 | |
| rameshg87 | dtantsur, mainly it was to mock ilo_client.IloError | 13:05 |
| *** rloo has joined #openstack-ironic | 13:05 | |
| rameshg87 | for example, the usage at https://review.openstack.org/#/c/89500/16/ironic/drivers/modules/ilo/common.py, L124-127 | 13:05 |
| dtantsur | I remember, yes. | 13:05 |
| dtantsur | it's the correct idea. The implementation is incorrect | 13:06 |
| rameshg87 | dtantsur, oh okay. now i get it. :-) | 13:06 |
| rameshg87 | dtantsur, i didn't notice it was to add "@mock.patch.object(ilo_common, 'ilo_client')" | 13:07 |
| rameshg87 | dtantsur, i never thought about trying this out :-) | 13:07 |
| dtantsur | yes :) | 13:07 |
| rameshg87 | dtantsur, let me give a try ... | 13:07 |
| rameshg87 | dtantsur, thanks | 13:07 |
| rameshg87 | :-) | 13:07 |
| *** rakesh_hs has quit IRC | 13:08 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the validate() methods https://review.openstack.org/91104 | 13:15 |
| lucasagomes | ^ had a 2 +2 before, needed rebase... | 13:15 |
| NobodyCam | oh such sleepy time hold music | 13:15 |
| *** coolsvap|afk is now known as coolsvap | 13:17 | |
| NobodyCam | beethoven's lullaby to callers | 13:18 |
| *** krtaylor has joined #openstack-ironic | 13:18 | |
| lucasagomes | NobodyCam, when u get some time can u take a look at 91104? | 13:18 |
| NobodyCam | otp atm | 13:18 |
| rloo | morning everyone! | 13:18 |
| lucasagomes | NobodyCam, :D hah | 13:18 |
| NobodyCam | good mornig rloo | 13:18 |
| lucasagomes | rloo, morning | 13:18 |
| rloo | lucasagomes: if it is only a rebase, I'll +2 91104 after Jenkins approves. Oh, I guess I can do it now. | 13:19 |
| lucasagomes | rloo, yeah it was only a rebase, we removed some code from the validate() in the ipmitool (https://review.openstack.org/#/c/99121/) and it was conflicting with that change | 13:20 |
| *** jdob_ has joined #openstack-ironic | 13:20 | |
| *** Faizan has quit IRC | 13:21 | |
| rushiagr | rameshg87: is this driver required for working with an HP iLO gen8 machine? | 13:22 |
| rushiagr | I was trying Ironic with that machine, and following this blog http://ma.ttwagner.com/bare-metal-deploys-with-devstack-and-ironic/ with some modifications | 13:22 |
| rushiagr | and the provisioning failed, the prominent error was 'timout happened waiting for callback' or something on these lines | 13:23 |
| rameshg87 | rushiagr: the current ironic driver works with hp proliant gen8 machines | 13:23 |
| *** martyntaylor1 has joined #openstack-ironic | 13:23 | |
| *** martyntaylor has quit IRC | 13:23 | |
| rushiagr | I was thinking I made some networking configuration error, but then saw the conversation above | 13:23 |
| *** martyntaylor has joined #openstack-ironic | 13:23 | |
| rloo | lucasagomes: done ;) | 13:24 |
| rameshg87 | rushiagr, the newly proposed driver only tries to some other capabilities (like avoiding pxe in deploy, etc) | 13:24 |
| rushiagr | rameshg87: okay, so it looks like the issue is something else. I was using the 'pxe_ipmitool' driver btw, just to be more clear | 13:24 |
| rameshg87 | rushiagr, yeah issue must be something else .. :-) | 13:24 |
| *** martyntaylor has quit IRC | 13:24 | |
| rushiagr | rameshg87: oh, okay. | 13:25 |
| rushiagr | rameshg87: okay, thanks. | 13:25 |
| *** martyntaylor has joined #openstack-ironic | 13:25 | |
| lucasagomes | rloo, ta much! | 13:25 |
| *** jdob has quit IRC | 13:26 | |
| *** jdob_ is now known as jdob | 13:26 | |
| dtantsur | rloo, morning | 13:32 |
| rloo | afternoon dtantsur :) | 13:33 |
| *** jgrimm_ has joined #openstack-ironic | 13:35 | |
| openstackgerrit | jan grant proposed a change to openstack/ironic: WIP: Serialise paramiko requests https://review.openstack.org/99626 | 13:35 |
| openstackgerrit | Ghe Rivero proposed a change to openstack/ironic: Migration to oslo.db https://review.openstack.org/92138 | 13:36 |
| openstackgerrit | A change was merged to openstack/ironic: Test for membership should be 'not in' https://review.openstack.org/97412 | 13:36 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: SSH virsh to use the new ManagementInterface https://review.openstack.org/89884 | 13:37 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: IPMITool to use the new ManagementInterface https://review.openstack.org/86092 | 13:37 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: SeaMicro to use the new ManagementInterface https://review.openstack.org/86328 | 13:37 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: IPMINative to use the new ManagementInterface https://review.openstack.org/86588 | 13:37 |
| *** shausy has quit IRC | 13:39 | |
| NobodyCam | lucasagomes: got a second to chat about 84631? | 13:42 |
| lucasagomes | NobodyCam, sure, 84631 ?! | 13:42 |
| lucasagomes | https://review.openstack.org/#/c/84631 ? | 13:42 |
| NobodyCam | :-p 84361 | 13:43 |
| NobodyCam | doh | 13:43 |
| lucasagomes | ah :) | 13:43 |
| lucasagomes | I think someone already proposed something for that no? | 13:43 |
| lucasagomes | https://review.openstack.org/#/c/91719/ | 13:44 |
| NobodyCam | ya but that is ssh only, and I think I am hitting that on real hw | 13:44 |
| NobodyCam | (thou not 100% sure) | 13:44 |
| NobodyCam | was the thinking that its not needed with real hw because of post time? | 13:45 |
| lucasagomes | +1 | 13:45 |
| lucasagomes | but I think that 91719 would cover that | 13:46 |
| lucasagomes | it's doing it in the PXE module | 13:46 |
| lucasagomes | which might cover pxe_ipmitool, pxe_ipminative, pxe_ssh etc... | 13:46 |
| NobodyCam | 91719 is checking the power driver | 13:46 |
| NobodyCam | seems what we need to leverage is 78041 | 13:47 |
| lucasagomes | ohh I see | 13:47 |
| lucasagomes | yeah I think it might worth to wait it idependent of the power driver | 13:47 |
| *** martyntaylor has quit IRC | 13:48 | |
| NobodyCam | ya | 13:49 |
| *** martyntaylor has joined #openstack-ironic | 13:50 | |
| NobodyCam | oh I need to go walkies in hte rain... brb | 13:50 |
| *** k4n0 has quit IRC | 13:50 | |
| lucasagomes | :) | 13:51 |
| lucasagomes | NobodyCam, mind adding a comment on that patch saying that ur hitting it also in the baremetal? so adam_g can adapt the patch | 13:51 |
| jroll | mornin y'all | 13:53 |
| lucasagomes | jroll, good morning | 13:53 |
| jroll | hey lucasagomes :) 2 things for you - jfyi, I rebased your instance_info stuff on top of https://review.openstack.org/#/c/90233/ (working on a chain for the agent driver) - also did a bit more refactoring on 90233 after talking to deva, if you'd like to look | 13:55 |
| NobodyCam | wow that was one of the quickest walkies I had.. | 13:55 |
| jroll | NobodyCam: joggies? | 13:55 |
| jroll | :P | 13:55 |
| jroll | dtantsur: welcome back! when you have time today, we're waiting on your vote for these two IPA patches: https://review.openstack.org/#/c/91202/ https://review.openstack.org/#/c/98678/ :) | 13:56 |
| dtantsur | jroll, hi, sure, will look in 5-10 minutes :) | 13:57 |
| *** igordcard has joined #openstack-ironic | 13:57 | |
| jroll | dtantsur: no rush at all, thanks! :) | 13:57 |
| *** k4n0 has joined #openstack-ironic | 13:59 | |
| lucasagomes | jroll, thank you for that | 13:59 |
| *** jistr has quit IRC | 13:59 | |
| jroll | np | 14:00 |
| lucasagomes | will review 90233 again, it's important to get it merged asap | 14:00 |
| NobodyCam | good morning jroll | 14:00 |
| jroll | yes | 14:00 |
| NobodyCam | added lucasagomes | 14:00 |
| jroll | lucasagomes: I just pulled out the cleanup stuff | 14:00 |
| jroll | lucasagomes: deleting the pxe config from disk etc | 14:00 |
| lucasagomes | right | 14:00 |
| *** jistr has joined #openstack-ironic | 14:01 | |
| *** dwalleck has joined #openstack-ironic | 14:07 | |
| *** dwalleck has quit IRC | 14:08 | |
| *** dwalleck has joined #openstack-ironic | 14:08 | |
| openstackgerrit | Chris Krelle proposed a change to openstack/ironic: Enforce a minimum time between all IPMI commands https://review.openstack.org/96558 | 14:10 |
| *** dhellmann has quit IRC | 14:12 | |
| *** dhellmann has joined #openstack-ironic | 14:13 | |
| agordeev | good morning Ironic! | 14:15 |
| NobodyCam | good morning agordeev | 14:16 |
| agordeev | does anybody know where is better to ask about cloud-init in context of images built by disk-imagebuilder? | 14:16 |
| agordeev | NobodyCam: morning :) | 14:16 |
| NobodyCam | agordeev: #tripleo | 14:16 |
| agordeev | NobodyCam: ty! | 14:16 |
| *** romcheg1 has joined #openstack-ironic | 14:19 | |
| *** romcheg has quit IRC | 14:20 | |
| dtantsur | agordeev, morning | 14:21 |
| agordeev | dtantsur: morning | 14:22 |
| *** romcheg has joined #openstack-ironic | 14:23 | |
| *** dwalleck has quit IRC | 14:23 | |
| *** romcheg1 has quit IRC | 14:25 | |
| *** k4n0 has quit IRC | 14:27 | |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Use processutils properly https://review.openstack.org/91202 | 14:27 |
| *** dwalleck has joined #openstack-ironic | 14:28 | |
| *** rwsu has joined #openstack-ironic | 14:29 | |
| openstackgerrit | Ramakrishnan G proposed a change to openstack/ironic: Add IloDriver and its IloPower module https://review.openstack.org/89500 | 14:29 |
| *** coolsvap is now known as coolsvap|afk | 14:30 | |
| *** rameshg87 has left #openstack-ironic | 14:30 | |
| NobodyCam | next call time.. back in a bit | 14:55 |
| *** foexle has quit IRC | 14:55 | |
| *** jistr has quit IRC | 14:58 | |
| NobodyCam | more sleepy time hold music :( | 14:58 |
| *** jistr has joined #openstack-ironic | 14:59 | |
| * NobodyCam always thinks of this when waiting for conf calls: https://www.youtube.com/watch?v=fGGWrJp4JHA&feature=kp | 15:00 | |
| *** rloo has quit IRC | 15:02 | |
| *** rloo has joined #openstack-ironic | 15:03 | |
| *** datajerk has joined #openstack-ironic | 15:06 | |
| lucasagomes | NobodyCam, dtantsur https://review.openstack.org/#/c/91104/ | 15:06 |
| lucasagomes | had a 2 +2 before, and was rebased | 15:06 |
| * dtantsur is looking | 15:06 | |
| *** dwalleck has quit IRC | 15:07 | |
| *** pcrews has joined #openstack-ironic | 15:07 | |
| *** Guest8031 is now known as mgagne | 15:08 | |
| *** mgagne has joined #openstack-ironic | 15:08 | |
| *** mdorman has joined #openstack-ironic | 15:09 | |
| rloo | lucasagomes: https://review.openstack.org/#/c/98372/. Do you know any reason why it wasn't approved? | 15:11 |
| lucasagomes | rloo, gate problems I think | 15:11 |
| lucasagomes | gate was very slow before so we were approving | 15:11 |
| lucasagomes | only important bug fixes etc | 15:11 |
| rloo | lucasagomes: so I shouldn't approve? oh, this is from last week. are we ok approving things this week? | 15:12 |
| lucasagomes | hmm I'm approving stuff today | 15:12 |
| lucasagomes | gate seems fast enough | 15:12 |
| dtantsur | rloo, gate seems to work ok, I'm approving things as well | 15:12 |
| *** devananda changes topic to "Bare Metal Provisioning | Status: http://bit.ly/ironic-whiteboard | Docs: http://docs.openstack.org/developer/ironic/ | Bugs: https://bugs.launchpad.net/ironic" | 15:12 | |
| rloo | dtantsur, lucasagomes: ok thx. I approved it. | 15:13 |
| devananda | morning, all | 15:13 |
| lucasagomes | devananda, morning | 15:13 |
| dtantsur | devananda, morning | 15:13 |
| rloo | morning devananda | 15:13 |
| devananda | sdague changed a few things late last week, ironic is no longer in the integrated gate | 15:14 |
| dtantsur | what does it mean for us? | 15:14 |
| devananda | meaning, essentially, we have our own queue and landing patches is now must faster for us | 15:14 |
| dtantsur | oh, wow | 15:14 |
| devananda | non-integratd projects (like ones still in incubation) shouldn't be part of the integrated gate | 15:14 |
| devananda | when ironic graduates, we'll become part of that gate | 15:15 |
| devananda | some changes to the way oslo gets tested caused it to happen prematurely (that's what he fixed) | 15:15 |
| lucasagomes | seems reasonable | 15:16 |
| *** rloo has quit IRC | 15:16 | |
| NobodyCam | good morning devananda | 15:16 |
| romcheg | That sounds good | 15:16 |
| romcheg | Morning devananda! | 15:16 |
| *** rloo has joined #openstack-ironic | 15:17 | |
| NobodyCam | morning romcheg | 15:18 |
| romcheg | g'morning NobodyCam! | 15:18 |
| *** datajerk has quit IRC | 15:18 | |
| *** zul has quit IRC | 15:22 | |
| dtantsur | folks, are we having a meeting today? I already started filling agenda :) | 15:24 |
| devananda | dtantsur: yep | 15:26 |
| openstackgerrit | Ruby Loo proposed a change to openstack/ironic: VendorPassthru.validate()s call _parse_driver_info https://review.openstack.org/97558 | 15:27 |
| dtantsur | cool! than see you later | 15:27 |
| *** dtantsur is now known as dtantsur|afk | 15:27 | |
| lucasagomes | devananda, for the mgmt interface spec... you think it's ok to make the REST API as a different spec? | 15:32 |
| lucasagomes | devananda, since we don't want to expose any function accessing the BMC directly (sync) | 15:32 |
| *** viktors is now known as viktors|afk | 15:32 | |
| lucasagomes | we need to change the datamodel etc | 15:32 |
| devananda | lucasagomes: no, i think the changes to the REST API that are necessatedby the new management interface should be part of the spec | 15:34 |
| lucasagomes | devananda, right ok, gotta include those in the spec then | 15:35 |
| devananda | lucasagomes: given that the async api is still being discussed, i wouldn't pin the mgmtinterface work on that, just descripbe the API structure and base on what can be implemented now | 15:36 |
| lucasagomes | devananda, right, the get_boot_device will access the BMC directly, is that fine? | 15:36 |
| lucasagomes | I know we can rework that after, I'm ok with that as well | 15:37 |
| lucasagomes | if u say it's cool | 15:37 |
| devananda | lucasagomes: ugh | 15:37 |
| devananda | lucasagomes: then the same for set_boot_device? | 15:37 |
| lucasagomes | devananda, that's why I was thinking about making it async :) | 15:37 |
| lucasagomes | devananda, the set we can make async | 15:37 |
| lucasagomes | no bothers | 15:37 |
| devananda | k | 15:37 |
| lucasagomes | devananda, I can make both async, but for that I would need to include a "boot_device" field on the nodes | 15:38 |
| lucasagomes | devananda, + a periodic task to sync it | 15:38 |
| devananda | lucasagomes: right | 15:38 |
| lucasagomes | which seems fine for me as well | 15:38 |
| devananda | that's closer to what we've been discussing in the async work | 15:38 |
| lucasagomes | (looking into it now, I'm first trying to find a way to get the boot device from ipmitool) | 15:38 |
| lucasagomes | devananda, yes | 15:38 |
| lucasagomes | tho periodic tasks still needs some tune | 15:39 |
| lucasagomes | having another function accessing the BMC to sync state might slow down the other periodic tasks | 15:39 |
| lucasagomes | syncing power state etc | 15:39 |
| devananda | ya. there's a lot we will need to do w.r.t. that eventually | 15:39 |
| lucasagomes | sure | 15:39 |
| *** coolsvap|afk is now known as coolsvap | 15:39 | |
| lucasagomes | devananda, btw, i was looking into | 15:40 |
| * lucasagomes grab the link | 15:40 | |
| lucasagomes | devananda, https://pypi.python.org/pypi/APScheduler/2.1.2 | 15:40 |
| lucasagomes | to see if we could replace the periodic tasks with that ^ | 15:40 |
| lucasagomes | it's possible... but idk whether we should do that, or change the periodic task in oslo | 15:41 |
| devananda | incremental improvements are better here, i think, than waiting for The Perfect Solution | 15:41 |
| lucasagomes | to be able to run concurrent | 15:41 |
| *** ifarkas has quit IRC | 15:41 | |
| lucasagomes | devananda, and... last thing... can I have ur opnion on https://review.openstack.org/#/c/99426/ ? | 15:43 |
| lucasagomes | check out lifeless comment and my answer please | 15:43 |
| devananda | so with the IPMI fragility fix we landed last week | 15:44 |
| devananda | running periodic tasks that touch the BMC in parallel -- without some better coordination -- doesn't seem likea good thing | 15:44 |
| devananda | IOW, I think we'll need to coalesce "things we need to do periodiccally" based on node | 15:45 |
| devananda | and in run parallel against N nodes | 15:45 |
| devananda | rather than run separate /types of jobs/ in parallel | 15:46 |
| lucasagomes | right yeah I agree it's not a great idea to run a lot of periodic tasks, becasue hmm I don't think we want to bombard the BMC with a lot of requests | 15:47 |
| openstackgerrit | A change was merged to openstack/ironic: Remove 'node' parameter from the validate() methods https://review.openstack.org/91104 | 15:47 |
| lucasagomes | yeah we need to come up with a plan | 15:47 |
| lucasagomes | devananda, not the perfect solution... so maybe we should leave the get_boot_devices requesting it directly to the BMC then, until we come up with a plan | 15:50 |
| devananda | lucasagomes: not just "we dont want to" -- we will prevent it once https://review.openstack.org/#/c/96558/ lands | 15:52 |
| lucasagomes | ah, I see | 15:52 |
| *** igordcard has quit IRC | 15:52 | |
| *** bearhands is now known as comstud | 15:53 | |
| *** rloo has quit IRC | 15:54 | |
| *** rloo has joined #openstack-ironic | 15:55 | |
| *** max_lobur has quit IRC | 15:56 | |
| *** martyntaylor has quit IRC | 15:56 | |
| *** jcoufal has quit IRC | 15:56 | |
| *** dividehex has joined #openstack-ironic | 15:57 | |
| *** ellenh has joined #openstack-ironic | 16:00 | |
| linggao | morning Ironic. | 16:01 |
| lucasagomes | morning linggao | 16:01 |
| NobodyCam | mornign linggao | 16:01 |
| linggao | Hey lucasagomes NobodyCam | 16:02 |
| linggao | question: in a large BM pool where there are many ironic-conductors, | 16:03 |
| linggao | do we assign one ironic nova driver to each conductor? | 16:03 |
| devananda | linggao: no - at the moment, you should probably just use one nova-compute | 16:04 |
| devananda | linggao: for the whole cluster | 16:04 |
| lucasagomes | no it's totally transparent for the n-cpu | 16:04 |
| lucasagomes | yeah that ^ | 16:04 |
| devananda | linggao: there has been work done to allow multiple nova-compute processes to manage the same pool in an active-active fashion | 16:05 |
| devananda | thus providing high-availablility for the n-cpu layer | 16:05 |
| devananda | I dont think much testing has been done with that, and I'm fairlysure there are some race conditions that we haven't fixed yet, just from my knowledge of the old nova code | 16:06 |
| linggao | so for now one nova-comoute works with many nodes where ironic-conductors are on? | 16:06 |
| devananda | yes | 16:06 |
| *** eghobo has joined #openstack-ironic | 16:06 | |
| linggao | ok, thanks devananda. | 16:06 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: SSH virsh to use the new ManagementInterface https://review.openstack.org/89884 | 16:06 |
| devananda | and you can try multiple nova-compute, but you will definitely need to use this class | 16:06 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: IPMITool to use the new ManagementInterface https://review.openstack.org/86092 | 16:06 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: SeaMicro to use the new ManagementInterface https://review.openstack.org/86328 | 16:06 |
| devananda | https://github.com/openstack/ironic/blob/master/ironic/nova/compute/manager.py | 16:06 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: IPMINative to use the new ManagementInterface https://review.openstack.org/86588 | 16:06 |
| devananda | instead of default ComputeManager | 16:07 |
| *** hemna_ is now known as hemna | 16:08 | |
| *** athomas has quit IRC | 16:13 | |
| linggao | devananda, FYI, we are still working on setting up the third party ci, moving on slowly. Had some issues, hope it can be up soon. | 16:15 |
| *** blamar has joined #openstack-ironic | 16:16 | |
| NobodyCam | linggao: awesome-ness | 16:18 |
| linggao | NobodyCam, thanks. We are still play with the basic setup. Have not gotten to the real Ironic testing part yet. | 16:19 |
| linggao | Huge learming cuve to clump up for us. | 16:20 |
| linggao | learning curve | 16:20 |
| linggao | I cannot spell today :-) | 16:20 |
| linggao | climb? | 16:20 |
| romcheg | linggao: Yeah, learning curve for Infra's stuff is very hard | 16:20 |
| linggao | romcheg, true. | 16:21 |
| romcheg | linggao: You know either nothing or the most of the things :) | 16:21 |
| linggao | :) | 16:21 |
| romcheg | Again, I have no idea how to change that so just complaining :) | 16:22 |
| linggao | romcheg, I found out that working in community projects makes you (I mean me) feel very stupid. Maybe just me. It gets rid of all your pride. But good thing is that you learn a lot. | 16:23 |
| *** dividehex has quit IRC | 16:24 | |
| romcheg | linggao: +1! That's exactly what I feel and think :) | 16:25 |
| linggao | romcheg, I feel better now :-). | 16:26 |
| *** blamar has quit IRC | 16:31 | |
| devananda | lucasagomes: in looking at your iPXE patch 99318 | 16:34 |
| *** ifarkas has joined #openstack-ironic | 16:34 | |
| devananda | lucasagomes: it appears to assume that iPXE loader will be fetching images from a local apache server | 16:34 |
| lucasagomes | devananda, yes | 16:34 |
| lucasagomes | you can configure the http_root and http_server if u want | 16:35 |
| lucasagomes | there's also a devstack patch that you can use to test it | 16:35 |
| * lucasagomes lemme find the patch no | 16:35 | |
| lucasagomes | #99677 I see u reviewed it, thanks | 16:36 |
| devananda | just reviewed the devstak patch :) | 16:36 |
| romcheg | Guys, is dict thread-safe? | 16:36 |
| devananda | lucasagomes: so one of the neat things about making boot more scalable with ipxe will be being able to fetch iamges directly from glance | 16:36 |
| devananda | romcheg: dict? | 16:36 |
| devananda | romcheg: the data type? | 16:36 |
| romcheg | Yes | 16:36 |
| devananda | romcheg: it's a python built-in. i'm going to assume its threadsafe until someone shows me evidence otherwise. | 16:37 |
| lucasagomes | devananda, yes! I thought bout it and we have a spec right now | 16:37 |
| devananda | romcheg: actually, how are you using a dict between threads? | 16:37 |
| romcheg | I just realised, I do not know whether adding/changing data in the dict in parallel is going to make any problems | 16:37 |
| romcheg | devananda: https://review.openstack.org/#/c/96558/22/ironic/drivers/modules/ipmitool.py | 16:37 |
| lucasagomes | devananda, adding support to the swift urls | 16:37 |
| romcheg | Like this | 16:38 |
| devananda | lucasagomes: right. so at some point we'll need pxe driver t ounderstand /not/ downloading images locally | 16:38 |
| lucasagomes | devananda, yes, can be configurable whether u want it to downloading directly from glance+swift or locally | 16:38 |
| lucasagomes | we just need a flag for it to make it pluggable | 16:39 |
| romcheg | devananda: Probably I'm just panicking | 16:39 |
| devananda | romcheg: so we aren't using real threads here | 16:40 |
| devananda | romcheg: this shouldn't even run in separate PHYSICAL threads. just separate greenthreads | 16:40 |
| romcheg | devananda: Yeah. Should work | 16:40 |
| lucasagomes | romcheg, we can't change a dict when it's being interacted | 16:40 |
| lucasagomes | http://stackoverflow.com/questions/3358770/python-dictionary-is-thread-safe | 16:40 |
| devananda | romcheg: if we were using real threads here, then you'd be correct, and we should use a threadsafe data type instead of a dict | 16:41 |
| romcheg | Sometimes I forget about green threads | 16:41 |
| romcheg | Sorry for making panic :) | 16:41 |
| devananda | :) | 16:41 |
| lucasagomes | devananda, so, one thing about downloading it from glance is that we would need to use temporary urls... but they should not be temporary u know | 16:42 |
| lucasagomes | devananda, once it's in the config file, the URL should exist until we tear down that node or something | 16:42 |
| lucasagomes | devananda, I have to investigate it because I'm not familiar with swift temp urls | 16:43 |
| NobodyCam | brb | 16:43 |
| devananda | lucasagomes: that node would only need to download deploy k&r and user image once | 16:43 |
| devananda | if it were set to lcoal boot, that'd be it. othwerise, yea, it' could need to download user k&r repeatedly | 16:43 |
| lucasagomes | devananda, and the image kernel/ramdisk will be local? | 16:44 |
| lucasagomes | I thought about leaving it in glance as well | 16:44 |
| devananda | i have a draft spec up for supporting different types of boot state | 16:44 |
| devananda | we do need to support local boot | 16:44 |
| lucasagomes | ah right, yeah with localboot that would work | 16:44 |
| lucasagomes | but cool I will keep it in mind once the swift stuff gets in | 16:46 |
| lucasagomes | I will see if I can add HTTPS support to it as well | 16:46 |
| lucasagomes | gotta play with the certificate creation and stuff | 16:46 |
| *** jistr has quit IRC | 16:46 | |
| *** eghobo has quit IRC | 16:47 | |
| *** eghobo has joined #openstack-ironic | 16:47 | |
| lucasagomes | devananda, I will add a note in the scability part of the spec about fetching images directly from glance so we don't forget | 16:49 |
| lucasagomes | (as a future work) | 16:49 |
| *** ifarkas has quit IRC | 16:50 | |
| devananda | ++ | 16:53 |
| *** ifarkas has joined #openstack-ironic | 16:53 | |
| lucasagomes | aight, I will eat something will be back for the meeting later on | 16:54 |
| *** lucasagomes is now known as lucas-dinner | 16:54 | |
| openstackgerrit | Ellen Hui proposed a change to openstack/ironic: Add more INFO logging to ironic/common https://review.openstack.org/99495 | 16:57 |
| JayF | dtantsur|afk: thanks for the approval on the agent catch | 16:58 |
| JayF | *patch | 16:59 |
| *** rameshg87 has joined #openstack-ironic | 17:04 | |
| jroll | devananda: fwiw, we have an environment running two nova-compute nodes. we'll definitely make noise if things break :) | 17:05 |
| jroll | linggao: ^ | 17:06 |
| devananda | jroll: ooh, good to know | 17:06 |
| JayF | We'd strongly prefer to not have to make noise :P | 17:07 |
| *** rloo has quit IRC | 17:07 | |
| *** rloo has joined #openstack-ironic | 17:07 | |
| jroll | this has 2 +2's on it :) https://review.openstack.org/#/c/90233 | 17:13 |
| JayF | jroll: it's approved | 17:18 |
| JayF | \o/ | 17:18 |
| JayF | JoshNang should appreciate that ^^ | 17:18 |
| jroll | sick | 17:18 |
| jroll | thanks y'all | 17:18 |
| *** rloo has quit IRC | 17:23 | |
| openstackgerrit | Ellen Hui proposed a change to openstack/ironic: Add more INFO logging to ironic/common https://review.openstack.org/99495 | 17:26 |
| *** martyntaylor has joined #openstack-ironic | 17:26 | |
| openstackgerrit | Ghe Rivero proposed a change to openstack/ironic: Migration to oslo.db https://review.openstack.org/92138 | 17:27 |
| *** ndipanov has quit IRC | 17:27 | |
| JayF | jroll: it's double-obnoxious. Looks like the comment 72h after the last one appears to have kicked off another check... so it has to pass check, then go into the gate for merge | 17:28 |
| JayF | jroll: so for that to merge we're going to be queuing twice | 17:28 |
| JayF | huzzah | 17:28 |
| * jroll shrug | 17:28 | |
| jroll | we have our own gate now | 17:28 |
| jroll | is quite fast | 17:29 |
| JayF | yeah but we don't have our own check queue | 17:29 |
| JayF | and that is /not/ quite fast | 17:29 |
| * jroll still shrugs | 17:30 | |
| *** martyntaylor has left #openstack-ironic | 17:31 | |
| *** rloo has joined #openstack-ironic | 17:32 | |
| jroll | I hate to plug patches, but it's been almost a month since a core reviewed this patch: https://review.openstack.org/#/c/86744 | 17:32 |
| devananda | GheRivero: have you followed the discussion around db migration test bugs in oslo.db? | 17:33 |
| devananda | GheRivero: I think we will need to do some cleanup in our code after https://review.openstack.org/#/c/93424/ lands to be able to benefit from it | 17:34 |
| devananda | GheRivero: and probably rip out more once https://review.openstack.org/#/c/99965/2 lands (alembic migrations in oslo.db) | 17:34 |
| *** pelix has quit IRC | 17:34 | |
| jroll | devananda: quick(?) question: what's blocking boot from local disk today? | 17:36 |
| jroll | devananda: just that we are not writing a boot partition? | 17:36 |
| devananda | jroll: technically? yes. | 17:36 |
| devananda | jroll: have you seen the draft spec I put up describing several different needs thereabouts? | 17:37 |
| devananda | jroll: also, pls dont hesitate to ping folks if you feel a patch has fallen off the review queue. like that one. | 17:37 |
| jroll | devananda: I've seen that it exists :P | 17:37 |
| devananda | jroll: fwiw, i appreciate the pings | 17:37 |
| * jroll will read | 17:37 | |
| jroll | devananda: ok, great :) | 17:37 |
| devananda | jroll: thanks. it's probably incomplete. if taht's a feature ya'll want to work on, i'd be happy to have someone else run with the spec, too :) | 17:38 |
| JayF | I think the agent already supports local boot ;) | 17:38 |
| jroll | devananda: context for my question: IPA today only supports whole disk images and boot from local disk | 17:38 |
| *** romcheg has left #openstack-ironic | 17:38 | |
| *** romcheg has joined #openstack-ironic | 17:38 | |
| jroll | JayF: that's more because we don't support root/swap/ephemeral like we should :/ | 17:39 |
| devananda | so I think we need a programatic way to expose//determine what boot modes are supported//chosen//desired by what | 17:39 |
| JayF | I wish --preserve-ephemeral didn't exist. I don't like the concept especially in concert with hardware :X | 17:39 |
| *** ChanServ changes topic to "OpenStack Bare Metal Provisioning | Docs: http://docs.openstack.org/developer/ironic/ | Bugs: https://bugs.launchpad.net/ironic | Status: https://etherpad.openstack.org/p/IronicWhiteBoard" | 17:39 | |
| devananda | eg, some images may only be suitable for use with local boot. some nodes may be required to only netboot. etc | 17:40 |
| devananda | JayF: ditto, but it was added to nova-baremetal and thus ironic is required to support it | 17:40 |
| JayF | Can we have a secret plan to deprecate it as soon as we graduate? | 17:41 |
| JayF | :P | 17:41 |
| devananda | JayF: though my dislike for it aside, I do see hwo it's useful for certain workloads (eg, tripleo) | 17:41 |
| devananda | "upgrade the image on this machine without destroying my /data volume" | 17:41 |
| JayF | I just think in a world where we want to treat servers like cattle, the idea of 'preserving' data from the node seems silly | 17:41 |
| devananda | is what it wants to be | 17:41 |
| devananda | JayF: exactly | 17:41 |
| devananda | except ^ | 17:41 |
| JayF | I would be of the opinion something like that should be done with operating system / config management tools, not Openstack, but I'm too late to the party to have this argument :( | 17:42 |
| devananda | JayF: CM tools don't work at fleet scale | 17:42 |
| devananda | i'm generalizing, but generally speaking, they don't | 17:43 |
| JayF | I've run fleets of servers with CM tools, but okay :) | 17:43 |
| JayF | aka multi-thousand server environments. | 17:43 |
| devananda | that's the tipping point | 17:43 |
| devananda | 10^3 is workable for today's CM tools. entropy is low enough that divergence in the fleet is manageable by humans | 17:44 |
| devananda | at 10^6 the entropy and divergence becomes unmanageable for humans, and the CM systems I've seen don't handle it well either | 17:44 |
| *** ellenh has quit IRC | 17:45 | |
| devananda | increased homogeneity across the fleet is needed. hence, golden images, and doing upgrades by pushing out new golden images, rather than using packages | 17:46 |
| *** rloo_ has joined #openstack-ironic | 17:46 | |
| devananda | at least that's my working theory :) | 17:46 |
| *** rloo__ has joined #openstack-ironic | 17:47 | |
| devananda | JayF: i'd be very interested in ya'll's experience with CMS if you've had success with them beyond ~10^3 nodes | 17:48 |
| *** rloo_ has quit IRC | 17:50 | |
| *** max_lobur has joined #openstack-ironic | 17:51 | |
| *** blamar has joined #openstack-ironic | 17:52 | |
| *** ellenh has joined #openstack-ironic | 17:54 | |
| *** rloo__ has quit IRC | 17:55 | |
| *** rloo_ has joined #openstack-ironic | 17:56 | |
| *** rloo_ has quit IRC | 17:56 | |
| *** rloo_ has joined #openstack-ironic | 17:56 | |
| *** rloo_ has quit IRC | 17:56 | |
| *** rloo_ has joined #openstack-ironic | 17:57 | |
| JayF | devananda: My experience mainly would say that at that level of scaling, you need to talk about /replacing/ services rather than upgrading inline. | 17:59 |
| JayF | Which is why the argument as that being the use case for --preserve-ephemeral is a little perplexing to me | 17:59 |
| JayF | I'd never run an inline upgrade on a server where data mattered; I'd deploy a copy of it and migrate over data | 18:00 |
| Shrews | adam_g: I *think* I might have a handle on waiting for the ironic node between tempest tests: https://review.openstack.org/#/c/94439/7/tempest/scenario/manager.py | 18:00 |
| *** harlowja has joined #openstack-ironic | 18:05 | |
| openstackgerrit | A change was merged to openstack/ironic: Reuse validate_instance_and_node() Nova ironic Driver https://review.openstack.org/98372 | 18:13 |
| *** rloo_ has quit IRC | 18:13 | |
| *** rloo_ has joined #openstack-ironic | 18:13 | |
| openstackgerrit | Russell Haering proposed a change to openstack/ironic-python-agent: Fix HardwareManager loading order bug https://review.openstack.org/98678 | 18:14 |
| NobodyCam | I know late notice, but any one have changes / updates for the agenda? | 18:15 |
| JayF | let me look at it | 18:15 |
| JayF | Any objection to me adding a link to the meeting agenda on the whiteboard? | 18:16 |
| JayF | I just want it accessible via /topic somehow :) | 18:16 |
| JayF | lgtm as it sits | 18:17 |
| *** harlowja has quit IRC | 18:19 | |
| *** harlowja has joined #openstack-ironic | 18:19 | |
| lifeless | lucas-dinner: devananda: replied | 18:19 |
| devananda | hmm, looks like the bot changed the room topic to the old one | 18:21 |
| NobodyCam | devananda: that bot broke over the weekend | 18:22 |
| *** devananda changes topic to "Bare Metal Provisioning | Status: http://bit.ly/ironic-whiteboard | Docs: http://docs.openstack.org/developer/ironic/ | Bugs: https://bugs.launchpad.net/ironic" | 18:22 | |
| *** dtantsur|afk has quit IRC | 18:23 | |
| *** dtantsur has joined #openstack-ironic | 18:23 | |
| NobodyCam | lifeless: sorry I missed the ping RE 98050... sure | 18:25 |
| dtantsur | g'evening. If anyone asked me something, please retry, as my client lost the history :( | 18:26 |
| JayF | I just told you thanks for getting that fix in :) | 18:27 |
| adam_g | Shrews, nice! | 18:28 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Drivers determine acceptable power states https://review.openstack.org/86744 | 18:30 |
| lifeless | NobodyCam: thanks | 18:30 |
| jroll | that's a rebase, should be good to go now ^ | 18:30 |
| devananda | JoshNang: comments on https://review.openstack.org/#/c/86744/ | 18:34 |
| devananda | or jroll ^ | 18:35 |
| jroll | devananda: thanks | 18:37 |
| *** rloo_ has quit IRC | 18:37 | |
| *** rloo_ has joined #openstack-ironic | 18:38 | |
| NobodyCam | lifeless: side was there any one patch that help with the tftp timeout issue the hp ce guys were running into? ES team seems to be runing into the same issue where nodes power on get ipaddress then timeout with tftp error | 18:40 |
| jroll | devananda: totally agree - although I think the check for changing the instance<->node association of an already-deployed node should not be based on power state (which is part of this change) | 18:40 |
| *** rloo_ has quit IRC | 18:40 | |
| *** rloo_ has joined #openstack-ironic | 18:40 | |
| jroll | devananda: essentially, the goal is that the driver can determine if the power state is acceptable for deploys, which that bit plays a role in | 18:41 |
| jroll | devananda: I agree that the change does more than it should, though, will fix | 18:41 |
| jroll | thanks! | 18:41 |
| lifeless | NobodyCam: they had the wrong eth card in their br-ctlplane | 18:41 |
| lifeless | NobodyCam: the logpower driver worked fine once we identified that and corrected it | 18:41 |
| devananda | jroll: welcome! I agree with the parts that the commit message covers | 18:42 |
| lifeless | well, modulo my bug with assuming macs was present | 18:42 |
| jroll | devananda: great :) there will need to be changes in that section, just different than what is currently there | 18:42 |
| NobodyCam | ahh... lifeless would you have a second to take a look at teh ironic "hipChat" channel I was trying to help them on Sunday | 18:43 |
| lifeless | k | 18:44 |
| dtantsur | lifeless, could you have some agreement with devananda on https://bugs.launchpad.net/python-ironicclient/+bug/1327517 if you insist? | 18:44 |
| jroll | we have a hipchat channel? O.o | 18:44 |
| devananda | jroll: also, think about how this will interact with making the API async. we'll need to be able to perform the validation in the API layer as well, I believe | 18:44 |
| devananda | jroll: no, they're talking about internal things ... | 18:44 |
| jroll | oh, heh | 18:45 |
| jroll | and yeah, wuill consider the async work | 18:45 |
| jroll | will, even | 18:45 |
| Shrews | adam_g: I'll reply in the review, but also here... so, 'sync' sends the flush request to the kernel, but the kernel may not do it right away. So the sleep() is necessary to give it time. | 18:49 |
| Shrews | adam_g: which is why they used to (still do?) recommend 3 syncs. The time it took to type the extra 2 syncs gave the kernel time to flush :) | 18:50 |
| *** rameshg87 has left #openstack-ironic | 18:50 | |
| GheRivero | morning all | 18:51 |
| lifeless | Shrews: huh, sync is a blocking clll | 18:52 |
| adam_g | Shrews, fer real? i always thought 'sync' was atomic. | 18:52 |
| lifeless | its not atomic | 18:52 |
| lifeless | but it is blocking | 18:52 |
| lifeless | when it returns, the kernel has successfully synced all the pages that were dirty at the start of the call, as well (IIRC) triggering a barrier on all fs's. | 18:53 |
| russell_h | if so, that's new | 18:54 |
| russell_h | and by "new" I mean less than 5 years old | 18:54 |
| *** dwalleck has joined #openstack-ironic | 18:54 | |
| russell_h | http://man7.org/linux/man-pages/man8/sync.8.html | 18:55 |
| russell_h | "On Linux, sync is guaranteed only to schedule the dirty blocks for writing" | 18:55 |
| russell_h | er, I guess the actual syscall docs are more authoritative | 18:56 |
| russell_h | but say the same thing | 18:56 |
| russell_h | ha | 18:56 |
| russell_h | "However, since version 1.3.20 Linux does actually wait" | 18:56 |
| russell_h | so I guess a lot more than 5 years | 18:56 |
| lifeless | yup | 18:57 |
| lifeless | and per Linus this ain't going to change. | 18:57 |
| lifeless | also because on modern systems 3 seconds is no where near enough to flush dirty pages. | 18:57 |
| Shrews | lifeless: well, if that's NOT true, then there is some other reason a single sync isn't actually sync'ing a newly created file immediately | 18:57 |
| lifeless | (1GB of ram is ~ 8 seconds IO for most consumer devices) | 18:58 |
| lifeless | Shrews: so I missed the beginning of the conversation. Whats the exact thing being checked ? | 18:58 |
| *** mrda-away is now known as mrda | 18:58 | |
| mrda | morning ironic | 18:59 |
| *** dwalleck has quit IRC | 18:59 | |
| Shrews | lifeless: a tempest test that touches a file on the ephemeral partition, then a 'nova rebuild' is performed, then we validate the touched file still exists | 18:59 |
| *** lucasagomes has joined #openstack-ironic | 18:59 | |
| *** dwalleck has joined #openstack-ironic | 18:59 | |
| dtantsur | oh, we're approaching a meeting! | 18:59 |
| romcheg | Good morning mrda! | 18:59 |
| Shrews | lifeless: b/c we don't politely shutdown a node on rebuild, it sometimes does not exist, unless we force sync | 19:00 |
| mrda | romcheg: Good later on today :) | 19:00 |
| *** lucas-dinner has quit IRC | 19:00 | |
| romcheg | mrda: :-P | 19:00 |
| NobodyCam | meeting? | 19:01 |
| jroll | meeting! | 19:01 |
| lifeless | Shrews: how are we making the file? ssh in ? | 19:01 |
| Shrews | lifeless: yes | 19:01 |
| lifeless | Shrews: and you have a sync(), but it doesn't always work ? | 19:01 |
| Shrews | lifeless: right (/bin/sync, to be precise) | 19:02 |
| devananda | woops! | 19:02 |
| devananda | meeting! | 19:02 |
| devananda | sorry | 19:02 |
| *** dwalleck_ has joined #openstack-ironic | 19:03 | |
| *** dwalleck has quit IRC | 19:03 | |
| *** sysexit has quit IRC | 19:04 | |
| *** Penick has joined #openstack-ironic | 19:05 | |
| *** Mikhail_D_ltp has joined #openstack-ironic | 19:07 | |
| *** ellenh has quit IRC | 19:12 | |
| openstackgerrit | A change was merged to openstack/ironic: VendorPassthru.validate()s call _parse_driver_info https://review.openstack.org/97558 | 19:22 |
| *** blamar has quit IRC | 19:23 | |
| *** ellenh has joined #openstack-ironic | 19:41 | |
| *** max_lobur has quit IRC | 19:46 | |
| openstackgerrit | A change was merged to openstack/ironic: Factoring out PXE and TFTP functions https://review.openstack.org/90233 | 19:49 |
| jroll | JoshNang: ^^ | 19:49 |
| JayF | \o/ | 19:49 |
| *** sysexit has joined #openstack-ironic | 19:55 | |
| jroll | lucasagomes: I may be optimistic, but I think IPA will be production-ready for juno release, personally | 19:58 |
| dtantsur | sounds amazing :) | 19:58 |
| JayF | Probably already is now, depending on how smart hte people are running it and what features you need. | 19:59 |
| dtantsur | my concern is: does it make sense to implement some trivial nodes discovery for PXE driver in case it won't be in IPA by J | 19:59 |
| dtantsur | devananda, what do you think? ^^^ | 19:59 |
| lucasagomes | jroll, nice | 20:00 |
| jroll | dtantsur: so, what do you actually mean by node discovery? like, the operator does not put nodes in ironic's db, rather that is automatic? | 20:00 |
| lucasagomes | jroll, btw, you guys might want to take a look at https://review.openstack.org/#/c/98822/ as well | 20:00 |
| lucasagomes | I got patches on ironic and devstack for that ^ | 20:00 |
| jroll | lucasagomes: nice :D | 20:00 |
| JayF | dtantsur: I mean, with the hardwaremanager stuff agordeev is working on, I think we're mainly missing plumbing and infrastructure for discovery | 20:01 |
| JayF | dtantsur: i.e. the agent can get all the info, and tell ironic about it, just right now the agent is wired to not do much if it can't find a ndoe | 20:01 |
| lucasagomes | https://review.openstack.org/99677 and https://review.openstack.org/99318 | 20:01 |
| JayF | dtantsur: I'd think duplicating that effort somewhere else would be a bad use of time. | 20:01 |
| NobodyCam | ok a couple of minutes then the next call | 20:01 |
| romcheg | That was a nice meeting | 20:01 |
| lucasagomes | the discovery thing is a feature that a lot of ppl _really_ wants to have | 20:02 |
| devananda | discovery ++ | 20:02 |
| lucasagomes | idk how hard it would be to get a simply discovery working outside IPA | 20:02 |
| lucasagomes | but seems that it's not much | 20:02 |
| jroll | lucasagomes: including us :) | 20:02 |
| lucasagomes | jroll, yup | 20:02 |
| jroll | we have most of what's needed in flight to IPA right now | 20:03 |
| devananda | there are a few challenges with discovery | 20:03 |
| dtantsur | btw, how hard is tl | 20:03 |
| devananda | taht aren't obvious in the simple case | 20:03 |
| jroll | dtantsur: tl? | 20:03 |
| dtantsur | btw, how hard is to use IPA in RHEL/Fedora word? | 20:03 |
| devananda | such as authenticating the enrollment data | 20:03 |
| dtantsur | (I'm pressing random buttons already, sorry) | 20:03 |
| lucasagomes | devananda, same as we do for the deploy? token in a file on the tftp folder | 20:03 |
| lucasagomes | I know it's not great | 20:03 |
| JayF | dtantsur: Our agent works with full-disk images at the moment, so really as long as you can get a configdrive working for it it can image any OS you want :) | 20:03 |
| lucasagomes | but that's a v0.1 | 20:03 |
| devananda | lucasagomes: that's essentialy public info | 20:04 |
| jroll | dtantsur: we have support today for running in a linux container, fedora supports those right? | 20:04 |
| devananda | lucasagomes: anyone with network access has it | 20:04 |
| lucasagomes | devananda, yeah, I understand the security issues there | 20:04 |
| lucasagomes | devananda, but we do now have it in Ironic | 20:04 |
| jroll | devananda: we were talking about client-side certs the other day | 20:04 |
| dtantsur | jroll, should :) I'm really looking forward to some docs and will try on my lab machine with F20 | 20:04 |
| lucasagomes | devananda, maybe a way to do it would be having trust working in Ironic | 20:04 |
| JayF | dtantsur: if you're talking about running the agent itself inside RH or Fedora, that's not something any of us have put any effort in | 20:04 |
| jroll | devananda: for ironic -> IPA communication, at least | 20:04 |
| lucasagomes | so we can fix both, the deployment + discovery | 20:04 |
| devananda | lucasagomes: running with a fixed set of hardware on a controlled network is one thing. allowing hardware that "appears" on the network to add itself to the system automatically is scary for most people | 20:05 |
| JayF | jroll: that works better for already-enrolled agents | 20:05 |
| romcheg | Have to go now guys. Have a nice rest of the day! | 20:05 |
| JayF | jroll: but for discovery you have to boot an unauthenticated agent on an unauthenticated network | 20:05 |
| devananda | romcheg: cheers! have a good night | 20:05 |
| dtantsur | romcheg, buy | 20:05 |
| JayF | aka everything boots an agent | 20:05 |
| romcheg | dtantsur: please poke me tomorrow regarding to the spec | 20:05 |
| mrda | romcheg: bye! | 20:05 |
| JayF | at which point if you give them a cert, the authentication doesn't really help much | 20:05 |
| jroll | JayF: sure | 20:05 |
| lucasagomes | devananda, yeah well, thats true | 20:05 |
| JayF | at least in the way we run it in our lab, we only boot agents on known boxes, so sending it a client cert and trusting network-level security is more reasonable | 20:05 |
| *** yuriyz has quit IRC | 20:06 | |
| jroll | indeed | 20:06 |
| JayF | If you're using the agent for /discovery/, you have to at some level have: 1) Out of band information to validate (aka I know I have boxes with serial numbers 1001-1023) | 20:06 |
| JayF | or 2) Sufficient network security that you can trust ability to pxe == authorization to be added to ironic | 20:07 |
| JayF | neither of which is a great answer for upstream :) | 20:07 |
| adam_g | NobodyCam, hey! on what hw setup / driver were you seeing issues similar to the one addressed in https://review.openstack.org/#/c/91719/ | 20:07 |
| devananda | in more secure sustems, we would have authentication (the client is known to the system), authorization (the client is allowed to do $this), encryption (the data wasn't tampered with in transit), and attestation of each of the prior items by a seaprate trusted source | 20:07 |
| *** martyntaylor has joined #openstack-ironic | 20:07 | |
| *** martyntaylor has quit IRC | 20:07 | |
| NobodyCam | adam_g: ipmitool_pxe | 20:08 |
| NobodyCam | im not 100% sure about the cause | 20:08 |
| * jroll brb | 20:08 | |
| * devananda needs food | 20:08 | |
| devananda | bbiab | 20:08 |
| lucasagomes | off-topic: good news, the tempest test that is blocking the instance_info stuff just got a +2 | 20:08 |
| NobodyCam | but lifeless' patch did the slepp for all pxe nit just ssh | 20:08 |
| adam_g | NobodyCam, hmm. could it have been that race, or some other issue with neutron agents being down? | 20:09 |
| *** dwalleck_ has quit IRC | 20:09 | |
| NobodyCam | or even a bad param passed to neutron | 20:09 |
| NobodyCam | still working it | 20:09 |
| JayF | lucasagomes: nice | 20:09 |
| adam_g | lucasagomes, \o/ | 20:09 |
| adam_g | NobodyCam, is this happening on one of the racks we were banging on, or elsewhere? | 20:09 |
| dtantsur | ok folks, hope to continue on discovery later, now going to sleep | 20:10 |
| dtantsur | g'night | 20:10 |
| *** dtantsur is now known as dtantsur|afk | 20:10 | |
| lucasagomes | yeah it's late here too | 20:10 |
| lucasagomes | dtantsur|afk, g'night | 20:10 |
| NobodyCam | adam_g: it was with hte ES tests | 20:10 |
| *** dhellmann has quit IRC | 20:10 | |
| lucasagomes | we can talk about the discovery later, but makes sense to put an effort on IPA | 20:10 |
| NobodyCam | night dtantsur|afk | 20:10 |
| lucasagomes | night everybody | 20:11 |
| adam_g | NobodyCam, ES tests? | 20:11 |
| *** ifarkas has quit IRC | 20:11 | |
| *** lucasagomes is now known as lucas-sleepy | 20:11 | |
| NobodyCam | es team ... not sure what rack they are on | 20:11 |
| NobodyCam | night lucas-sleepy | 20:11 |
| lifeless | NobodyCam: yes because races are races, the sleep is only a pessimisation when something is a) faster than that already and b) neutron is guaranteed to be faster... e.g. nowhere. | 20:11 |
| *** dhellmann has joined #openstack-ironic | 20:12 | |
| *** martyntaylor has joined #openstack-ironic | 20:14 | |
| *** martyntaylor has quit IRC | 20:22 | |
| *** sysexit has quit IRC | 20:22 | |
| *** linggao has quit IRC | 20:23 | |
| *** dwalleck has joined #openstack-ironic | 20:26 | |
| *** petertoft has quit IRC | 20:32 | |
| *** rloo has quit IRC | 20:33 | |
| *** foexle has joined #openstack-ironic | 20:43 | |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Fix HardwareManager loading order bug https://review.openstack.org/98678 | 20:44 |
| mgagne | questions for people using ironic in their shop, how do you integrate your existing inventory system to ironic? | 20:47 |
| *** Penick has quit IRC | 20:50 | |
| *** jdob has quit IRC | 21:02 | |
| openstackgerrit | Chris Krelle proposed a change to openstack/ironic: Add in text for text mode on trusty https://review.openstack.org/98050 | 21:05 |
| NobodyCam | lifeless: tests should now pass on ^^^ ... I'll keep an eye on it | 21:05 |
| NobodyCam | brb | 21:08 |
| *** foexle has quit IRC | 21:10 | |
| *** jrist has quit IRC | 21:11 | |
| *** dhellmann has quit IRC | 21:11 | |
| *** jrist has joined #openstack-ironic | 21:12 | |
| *** dhellmann has joined #openstack-ironic | 21:13 | |
| *** ryanpetrello has joined #openstack-ironic | 21:15 | |
| ryanpetrello | hey Ironic folks | 21:15 |
| ryanpetrello | anybody have any insight into what’s plaguing me re: the failure on https://review.openstack.org/#/c/97475/? | 21:15 |
| ryanpetrello | if I can’t get this backported into Ironic at some point, I’m going to have to remove voting for ironic stable gating from pecan, because it’s holding up other patches | 21:16 |
| openstackgerrit | Ellen Hui proposed a change to openstack/ironic: Add more INFO logging to ironic/nova https://review.openstack.org/100360 | 21:26 |
| *** Penick has joined #openstack-ironic | 21:28 | |
| openstackgerrit | Ellen Hui proposed a change to openstack/ironic: Add more INFO logging to ironic/conductor https://review.openstack.org/99744 | 21:30 |
| devananda | ryanpetrello: looking | 21:33 |
| *** Mikhail_D_ltp has quit IRC | 21:33 | |
| devananda | ryanpetrello: oh, i think you pinged me on this a while ago. sorry for dropping it | 21:33 |
| ryanpetrello | np :) | 21:34 |
| devananda | ryanpetrello: backported? is this supposed to land on stable/icehouse branch or on master? | 21:34 |
| openstackgerrit | Chris Krelle proposed a change to openstack/ironic: Add in text for text mode on trusty https://review.openstack.org/98050 | 21:35 |
| ryanpetrello | devananda: backported to icehouse, yes? | 21:35 |
| ryanpetrello | pecan tests are failing for me against stable/icehouse because of this | 21:35 |
| ryanpetrello | it has been fixed in master | 21:35 |
| ryanpetrello | (and tests pass there) | 21:35 |
| ryanpetrello | but for the backport itself, I can’t get Ironic tests to pass | 21:35 |
| devananda | ryanpetrello: gotcha. so I may be the cause of this, since I didn't open stable/icehouse | 21:36 |
| ryanpetrello | they keep failing with what seems to be this bug: https://bugs.launchpad.net/openstack-ci/+bug/1300208 | 21:36 |
| devananda | oh, hmm | 21:38 |
| devananda | so https://review.openstack.org/#/c/89554/ is an earlier attempt to resolve the ksclient config changes | 21:38 |
| devananda | ryanpetrello: but your patch looks much more complete AND passes unit and api tests | 21:38 |
| devananda | ryanpetrello: your test failed not due to that CI bug, but because of this: http://logs.openstack.org/75/97475/1/check/check-tempest-dsvm-virtual-ironic/088c23b/console.html.gz#_2014-06-12_15_03_54_955 | 21:39 |
| jroll | devananda: I have a patch adding three methods to ipmitool's vendor interface. should those be three separate patches, or is one ok? | 21:40 |
| ryanpetrello | devananda: ah | 21:41 |
| ryanpetrello | blurgh | 21:41 |
| openstackgerrit | Adam Gandelman proposed a change to openstack/ironic: Wait for Neutron port updates when using SSHPower https://review.openstack.org/91719 | 21:42 |
| devananda | ryanpetrello: i'm digging through the debug log now | 21:42 |
| devananda | ryanpetrello: my initial guess is, our functional test won't pass on icehouse | 21:42 |
| ryanpetrello | okay | 21:43 |
| devananda | ryanpetrello: so we'll need to disable that job for stable/icehouse to be able to land patches there | 21:43 |
| devananda | yea | 21:46 |
| devananda | do_node_deploy is called but never finishes, afaict | 21:46 |
| devananda | http://logs.openstack.org/75/97475/1/check/check-tempest-dsvm-virtual-ironic/088c23b/logs/screen-ir-cond.txt.gz#_2014-06-12_14_57_45_403 | 21:46 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Add methods to ipmitool driver https://review.openstack.org/100364 | 21:47 |
| *** dwalleck has quit IRC | 21:51 | |
| rloo_ | are we serious about -2'ing reviews linked to non-approved blueprints? If we -2, only the -2'er and devananda can remove it, right? | 21:55 |
| mrda | I'm not sure even devananda can remove -2's rloo_ - I think it's only the person placing the -2 | 21:58 |
| rloo_ | mrda: I think devananda has more powers than you're aware of ;) | 21:58 |
| mrda | ...as I think another project had this issue with a core being absent for a while | 21:58 |
| devananda | ok, oddity. i can't +2/+A things on the stable/icehouse branch | 21:58 |
| mrda | rloo_: I know he can bend things with his mind, but can he remove -2's? | 21:59 |
| rloo_ | mrda: guess devananda has less power than I thought :-( | 21:59 |
| mrda | lol | 21:59 |
| *** romcheg has quit IRC | 22:00 | |
| devananda | rloo_: mrda I'm unaware of a means for me to remove other's -2's | 22:00 |
| devananda | it probably takes an action by infra | 22:00 |
| mrda | And that operation returns -ETOOHARD | 22:00 |
| devananda | rloo_: it sounded as though you were volunteering to help with managing feature proposals | 22:00 |
| devananda | rloo_: but maybe i read too much into it :) | 22:01 |
| rloo_ | devananda: in that case, I don't think we should -2 those reviews associated with unapproved blueprints. too harsh and too nasty to the user. | 22:01 |
| mrda | ...so anyone who -2's should follow up and un -2 things when appropriate | 22:01 |
| rloo_ | devananda: huh? i was just looking at at review and noticed it was associated with an unapproved BP :-) | 22:01 |
| *** foexle has joined #openstack-ironic | 22:01 | |
| devananda | for reference, nova has johnthetubaguy managing feature proposals | 22:01 |
| rloo_ | mrda: I don't want to be a blocker for anyone if I'm not around that minute/hour/day when someone wants something changed from -2... | 22:01 |
| devananda | so he's tracking bp and spec and adjusting status as needed | 22:01 |
| devananda | rloo_: if anything is that urgent, it's not a feature | 22:02 |
| rloo_ | devananda: oh, is that what you meant. a sort of managerial position ;) | 22:02 |
| devananda | rloo_: most days, that's all I do now :p | 22:02 |
| devananda | and i'm terrible at it | 22:02 |
| devananda | would much rather be writing code | 22:03 |
| rloo_ | devananda: I'd also rather be writing code. But then, I didn't apply for PTL ;) | 22:03 |
| devananda | some days, anyway. other days i'd rather be at the beach | 22:03 |
| devananda | rloo_: careful ... i got here by writing code ;) | 22:03 |
| rloo_ | devananda: I don't mind managing feature proposals if it is as you described it above. I'm just not sure about the -2 part. is that what mrtuba does? | 22:03 |
| devananda | rloo_: yep | 22:04 |
| rloo_ | devananda: is there some way to give all cores a way to override -2's? | 22:04 |
| devananda | rloo_: for example, https://review.openstack.org/#/c/68942 | 22:04 |
| devananda | rloo_: i would definitely not want to do that | 22:04 |
| devananda | -2 is meant to be a very strong objection. we can override each other's -1's easily (though we should avoid that unless there's a strong reason) | 22:05 |
| rloo_ | devananda: so you think a -1 is sufficient? the problem with -1 or work-in-progress is that it disappears with a new revision. | 22:06 |
| devananda | "the feature proposal for this code is not yet approved" is worth a -2 in my opinion, so that other cores, who may not have read the spec/bp, and look at the code, and think it's good, dont merge it | 22:06 |
| mrda | rloo_: I don't think it's possible without changing gerrit code to do that | 22:06 |
| * devananda steps back to look at the topic differently | 22:06 | |
| devananda | we have a lot of code that has been, and will be, proposed | 22:07 |
| devananda | enough that we often can't review it in a reasonable time | 22:07 |
| rloo_ | devananda: ok, if people want -2 for such things, if mrtuba can do it, I can do it too... i'll just need to make sure I check in once every 2 or so days :-) | 22:07 |
| rloo_ | devananda: I'll let you sleep on it and if you still want that tomorrow, I'll start. | 22:08 |
| rloo_ | devananda: gotta go. dinner bell has rung. (might be back later. or not.) | 22:08 |
| devananda | rloo_: ciao! | 22:08 |
| *** rloo_ is now known as rloo_out | 22:09 | |
| *** datajerk has joined #openstack-ironic | 22:09 | |
| NobodyCam | brb | 22:10 |
| devananda | mrda: hi! how's things? | 22:10 |
| mrda | devananda: good! | 22:11 |
| mrda | looking forward to midcycle | 22:11 |
| mrda | I have flights now | 22:11 |
| devananda | awesome | 22:12 |
| mrda | how are we traveling towards J? Are you happy with feature progress? bug fixes? | 22:13 |
| *** datajerk has quit IRC | 22:13 | |
| *** datajerk has joined #openstack-ironic | 22:14 | |
| devananda | bug fixes have been awesome | 22:14 |
| devananda | j1 milestone had 62 bug fixes | 22:14 |
| devananda | stability is definitely improving | 22:14 |
| mrda | that's pretty good | 22:14 |
| devananda | and Shrews has been working on our tempest test coverage. looks like it may land soon | 22:14 |
| devananda | i think you don't overlap with him at all, but dtantsur|afk has been doing a lot to clean up our bug list | 22:15 |
| devananda | basically finding in-progress bugs that got stale // had no progress, and updating them or poking people | 22:15 |
| devananda | we don't have enough discussion happening around specs right now | 22:16 |
| mrda | I'll take a look today - after the reminder in the meeting earlier | 22:16 |
| devananda | so i'm concerned about our ability to land features // the perception that the process is not being followed | 22:17 |
| mrda | I think people are still getting used to specs | 22:19 |
| devananda | we've had plenty of spec proposals, which is good | 22:19 |
| *** datajerk has quit IRC | 22:20 | |
| mrda | I guess getting them approved, and tagged for J, will help people see the roadmap forward | 22:20 |
| jroll | what version of keystone is our config built from? | 22:21 |
| devananda | mrda: right :) | 22:21 |
| devananda | jroll: our sample conf should be autogenerated from the latest keystoneclient | 22:21 |
| jroll | like, master? | 22:22 |
| devananda | jroll: oh, sorry. look in requirements.txt | 22:22 |
| devananda | jroll: that lists the version of keystoneclient. not "master" but that pip version (or current pip version) | 22:23 |
| jroll | I see that, but it seems to not match our master sample config | 22:23 |
| * jroll tries tox -r again | 22:23 | |
| devananda | fantastic | 22:23 |
| jroll | I thought we had this solved | 22:23 |
| devananda | if ksclient changes those opts again, it'll break us again | 22:24 |
| jroll | ah, I see | 22:24 |
| jroll | oh, maybe it is working | 22:24 |
| jroll | weird | 22:24 |
| jroll | it wasn't the other day for me | 22:24 |
| jroll | sorry | 22:24 |
| *** Penick has quit IRC | 22:32 | |
| adam_g | can anyone suggest a better way to mock whats being added in setUp() @ https://review.openstack.org/#/c/99518/3/ironic/nova/tests/virt/ironic/test_driver.py ? | 22:33 |
| *** datajerk has joined #openstack-ironic | 22:33 | |
| *** jgrimm_ has quit IRC | 22:34 | |
| NobodyCam | brb ... fish feed time... (fft) | 22:34 |
| NobodyCam | so nice to watch the fish getting fed | 22:40 |
| openstackgerrit | Russell Haering proposed a change to openstack/ironic-python-agent: Reduce requests connection pool size https://review.openstack.org/100375 | 22:47 |
| lifeless | NobodyCam: awesome | 22:48 |
| openstackgerrit | Chris Krelle proposed a change to openstack/ironic: Add in text for text mode on trusty https://review.openstack.org/98050 | 22:48 |
| NobodyCam | Shrews: is there a place to lists what tests fail for us (vs the rest of the world) <- in ref to https://bugs.launchpad.net/ironic/+bug/1329077 | 22:52 |
| NobodyCam | s/lists/list | 22:52 |
| *** datajerk has quit IRC | 22:58 | |
| *** Penick has joined #openstack-ironic | 23:05 | |
| * devananda does a blueprint/spec clean up | 23:08 | |
| JayF | the cleanest specs are the ones that are +2'd and merged | 23:08 |
| JayF | :P | 23:08 |
| devananda | heh | 23:09 |
| adam_g | devananda, another challenge in keeping the testing going on stable branches is there is no DIB stable branch. we're building ramdisks using elements from trunk, which may change in a way that is not backward compatible | 23:13 |
| *** lucas-sleepy has quit IRC | 23:15 | |
| *** radsy has joined #openstack-ironic | 23:16 | |
| *** radsy has joined #openstack-ironic | 23:16 | |
| devananda | adam_g: ugh, right | 23:19 |
| devananda | so if tripleo wants their tools, eg. dib, to be consumable by projects that do care about about stable branches, I think they need to support that | 23:20 |
| devananda | for instance, I would like to use DIB to build IPA images | 23:21 |
| devananda | and some day have a stable branch of both ironic and ipa | 23:21 |
| devananda | but if dib itself changes in a non-backwards-compatible way after we tag a stable release of ironic and ipa, that's not going to work | 23:21 |
| devananda | lifeless: thoughts ^ ? | 23:22 |
| adam_g | right | 23:23 |
| devananda | adam_g: do you think dib is, at least inthis case, transparent enough that it won't be an issue? | 23:23 |
| *** Penick has quit IRC | 23:23 | |
| devananda | adam_g: the problem is in the image elements, not dib itself, i think | 23:24 |
| devananda | adam_g: and that the ironic-deploy element isn't versioned right now | 23:24 |
| adam_g | devananda, maybe? one alternative would be to have a stable-$foo-dsvm-virtual-ironic job running in the DIB pipeline? | 23:24 |
| devananda | ** doesn't have a stable branch | 23:25 |
| adam_g | devananda, right. the elements. tripleo-image-elements does have a stable branch. | 23:25 |
| devananda | ah, good | 23:25 |
| adam_g | DIB doesn't, but DIB repo has some elements we need..specifically deploy-ironic | 23:25 |
| devananda | right - that's the problem, then | 23:25 |
| *** dividehex has joined #openstack-ironic | 23:25 | |
| lifeless | devananda: thats why dib does semver | 23:25 |
| lifeless | devananda: and releases once a week | 23:26 |
| devananda | what ever bits of code are interacting directly with ironic during a deploy need to be versionable | 23:26 |
| lifeless | devananda: have you found a backwards incompatible change? | 23:26 |
| devananda | what builds that image doesn't matter to our testing | 23:26 |
| devananda | lifeless: nope. adam was pointing out a potential issue with ironic stable branch maintenance | 23:26 |
| devananda | lifeless: which, it seems we already have (but not due to dib) | 23:26 |
| lifeless | right, so for stable branches there is an existing answer | 23:27 |
| lifeless | you put an upper version lock on the API you depend on on all your deps | 23:27 |
| lifeless | novaclient | 23:27 |
| lifeless | neutronclient | 23:27 |
| lifeless | dib | 23:27 |
| lifeless | etc | 23:27 |
| *** blamar has joined #openstack-ironic | 23:27 | |
| devananda | dib isn't a dependency of ironic's -- devstack uses it when preparing the environment, though | 23:27 |
| devananda | the deploy-ironic _element_ is a dependency, however | 23:28 |
| lifeless | its a testing dependency since there is code coupling. | 23:28 |
| lifeless | tie is a direct dependency | 23:28 |
| lifeless | dib is a dep of tie | 23:28 |
| devananda | lifeless: ah. dib now has pips. adam_g, we should add those, with version bounds, in devstack | 23:30 |
| devananda | thanks | 23:30 |
| *** Penick has joined #openstack-ironic | 23:31 | |
| adam_g | devananda, hmm. where? AFAIK devstack doesn't allow specifying requirements as such. it relies on the requirements described in each project's requirements.txt | 23:34 |
| adam_g | that would mean adding DIB to ironic's requirements.txt and having devstack run DIB from the ironic venv? | 23:35 |
| *** mdorman has quit IRC | 23:36 | |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Fix: partitions recently imaged not being detected https://review.openstack.org/99504 | 23:36 |
| devananda | adam_g: right | 23:38 |
| *** Penick has quit IRC | 23:38 | |
| devananda | that'd put dib in the global requirements file | 23:38 |
| devananda | *that'd take putting ... | 23:38 |
| adam_g | devananda, hmm. seems like that would work. | 23:42 |
| openstackgerrit | A change was merged to openstack/ironic-python-agent: Reduce requests connection pool size https://review.openstack.org/100375 | 23:45 |
| devananda | adam_g: looks like dib is already in global reqs. so should be a trivialc hange | 23:49 |
| devananda | just need a version boundary on it | 23:49 |
| adam_g | devananda, ack | 23:50 |
| *** hemna is now known as hemna_ | 23:50 | |
| Shrews | NobodyCam: not sure i follow. | 23:51 |
| NobodyCam | hey Shrews | 23:54 |
| Shrews | hi hi | 23:55 |
| NobodyCam | lol I just worked on tests.drivers.test_pxe.PXEPrivateMethodsTestCase.test__build_pxe_config and fixed the test so that they ran on my local env | 23:55 |
| NobodyCam | only to figure out that not what the gate wants. so I was just checking if you had a list of broken test anywhere | 23:56 |
| adam_g | devananda, this will start the process https://review.openstack.org/100385 | 23:56 |
| adam_g | devananda, we'll need to backport that to stable/icehouse of requirements, sync to both master stable of ironic, and adjust devstack to rely on that instead of a git clone | 23:57 |
| Shrews | NobodyCam: ah, no. i only reported what a full tox run reports atm | 23:57 |
| adam_g | er, sync to both master + stable/icehouse of ironic | 23:57 |
| NobodyCam | Shrews: ack :) | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!