| *** pcrews has joined #openstack-ironic | 00:15 | |
| *** radsy has quit IRC | 00:56 | |
| *** nosnos has joined #openstack-ironic | 01:02 | |
| *** Haomeng has quit IRC | 01:23 | |
| *** eghobo has quit IRC | 01:47 | |
| *** zz_naotok is now known as naotok | 02:09 | |
| *** naotok is now known as zz_naotok | 02:13 | |
| *** pcrews has quit IRC | 02:18 | |
| *** shakamunyi has joined #openstack-ironic | 02:41 | |
| *** Haomeng has joined #openstack-ironic | 02:50 | |
| *** shakamunyi has quit IRC | 03:00 | |
| *** mitz_ has quit IRC | 03:11 | |
| *** mitz has joined #openstack-ironic | 03:14 | |
| *** eghobo has joined #openstack-ironic | 03:23 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ironic: Updated from global requirements https://review.openstack.org/111524 | 03:23 |
|---|---|---|
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/python-ironicclient: Updated from global requirements https://review.openstack.org/111629 | 03:28 |
| *** nosnos has quit IRC | 03:45 | |
| *** Poornima has joined #openstack-ironic | 03:53 | |
| *** ramineni has joined #openstack-ironic | 04:11 | |
| *** eghobo has quit IRC | 04:27 | |
| *** eghobo has joined #openstack-ironic | 04:28 | |
| *** nosnos has joined #openstack-ironic | 04:28 | |
| *** sabah has joined #openstack-ironic | 04:38 | |
| *** k4n0 has joined #openstack-ironic | 04:54 | |
| *** nikunj2512 has joined #openstack-ironic | 05:06 | |
| *** pradipta_away is now known as pradipta | 05:09 | |
| *** eguz has joined #openstack-ironic | 05:09 | |
| *** eghobo has quit IRC | 05:10 | |
| *** bvivek has joined #openstack-ironic | 05:11 | |
| *** sirushti has quit IRC | 05:18 | |
| *** sirushti has joined #openstack-ironic | 05:18 | |
| *** killer_prince is now known as lazy_prince | 05:37 | |
| openstackgerrit | Anusha Ramineni proposed a change to openstack/ironic-specs: Management Interface to firmware update https://review.openstack.org/100842 | 05:39 |
| *** Mikhail_D_ltp has joined #openstack-ironic | 05:42 | |
| *** k4n0 has quit IRC | 05:44 | |
| *** Mikhail_D_ltp has quit IRC | 05:47 | |
| *** rakesh_hs has joined #openstack-ironic | 05:52 | |
| *** k4n0 has joined #openstack-ironic | 05:58 | |
| *** dtantsur|brb is now known as dtantsur | 06:06 | |
| dtantsur | Morning, Ironic! | 06:06 |
| Haomeng | dtantsur: morning:) | 06:07 |
| dtantsur | :) | 06:07 |
| Haomeng | dtantsur: one question - do you know where we can raise the defect for pyghmi issue? | 06:08 |
| dtantsur | hmmm | 06:08 |
| Haomeng | dtantsur: I am trying to test pyghmi to implement the 'send sensor data to ceilometer' bp for our ipminative driver, but encounteer bug for pyghmi | 06:09 |
| Haomeng | dtantsur: I just raised it into our ironic bug list - https://bugs.launchpad.net/ironic/+bug/1352079 | 06:09 |
| dtantsur | Haomeng, no idea :( | 06:10 |
| Haomeng | dtantsur: ok, np:) | 06:10 |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/110860 | 06:11 |
| Haomeng | dtantsur: looks like pyghmi is mantained by xcat team | 06:11 |
| Haomeng | dtantsur: from ibm:) | 06:11 |
| Haomeng | dtantsur: I will contact them:) | 06:11 |
| dtantsur | makes sense) | 06:11 |
| Haomeng | dtantsur: :) | 06:11 |
| Haomeng | dtantsur: for ipminative driver supporting 'send sensor data to ceiloemter', I understand no need to write new spec, it covers with existing spec already, right? | 06:13 |
| dtantsur | Haomeng, I would say yes, if the implementation is more-or-less straightforward, new spec is not required | 06:13 |
| Haomeng | dtantsur: ok | 06:14 |
| Haomeng | dtantsur: :) | 06:14 |
| *** SpamapS has quit IRC | 06:17 | |
| *** pquerna has quit IRC | 06:18 | |
| *** SpamapS has joined #openstack-ironic | 06:19 | |
| *** pquerna has joined #openstack-ironic | 06:19 | |
| *** pquerna has joined #openstack-ironic | 06:20 | |
| *** SpamapS has quit IRC | 06:20 | |
| *** SpamapS has joined #openstack-ironic | 06:20 | |
| dtantsur | FYI http://specs.openstack.org/openstack/ironic-specs/ :) | 06:33 |
| *** lsmola has joined #openstack-ironic | 06:35 | |
| *** ifarkas has joined #openstack-ironic | 06:43 | |
| *** Mikhail_D_ltp has joined #openstack-ironic | 07:07 | |
| *** eguz has quit IRC | 07:22 | |
| *** Poornima has quit IRC | 07:42 | |
| *** rameshg87 has joined #openstack-ironic | 07:48 | |
| *** ndipanov has joined #openstack-ironic | 07:50 | |
| *** sabah has quit IRC | 08:00 | |
| *** jistr has joined #openstack-ironic | 08:05 | |
| rameshg87 | dtantsur, hi | 08:24 |
| rameshg87 | dtantsur, for the comment at L75 on https://review.openstack.org/#/c/111232/1/ironic/drivers/modules/pxe.py | 08:25 |
| rameshg87 | dtantsur, the property will be already under [pxe] group in ironic.conf file. do you still feel we should rename to tftp_image_cache_ttl ? | 08:25 |
| dtantsur | rameshg87, hi! I don't like the same name under different group, especially with similar names. Especially provided that difference between PXE and iSCSI won't be too clear to a user, not knowing our internals. | 08:27 |
| rameshg87 | dtantsur, okay. make sense to me too. will change it :-) | 08:27 |
| dtantsur | thanks | 08:28 |
| *** viktors|afk has quit IRC | 08:30 | |
| rameshg87 | dtantsur, just one more question | 08:31 |
| rameshg87 | dtantsur, https://review.openstack.org/#/c/111232/1/ironic/drivers/modules/pxe.py L272 | 08:31 |
| rameshg87 | dtantsur, we can combine the two methods, but they doing two separate things, isn't it better to keep them separate - and also for the same reason why the initial author proposed not to put check_image_size() within cache_instance_image() | 08:32 |
| *** athomas has joined #openstack-ironic | 08:46 | |
| *** k4n0 has quit IRC | 08:46 | |
| dtantsur | rameshg87, I'm ok with leaving it as is | 08:47 |
| rameshg87 | dtantsur, okay | 08:47 |
| *** bvivek has quit IRC | 08:54 | |
| *** k4n0 has joined #openstack-ironic | 08:59 | |
| *** rakesh_hs has quit IRC | 09:14 | |
| *** bvivek has joined #openstack-ironic | 09:15 | |
| *** shausy has joined #openstack-ironic | 09:18 | |
| *** k4n0 has quit IRC | 09:22 | |
| *** enikanorov__ has quit IRC | 09:24 | |
| *** k4n0 has joined #openstack-ironic | 09:25 | |
| *** lynxman has quit IRC | 09:27 | |
| *** lynxman has joined #openstack-ironic | 09:27 | |
| *** lynxman has joined #openstack-ironic | 09:27 | |
| *** lazy_prince is now known as killer_prince | 09:33 | |
| openstackgerrit | Ramakrishnan G proposed a change to openstack/ironic: Take iSCSI deploy out of pxe driver https://review.openstack.org/111232 | 09:40 |
| openstackgerrit | Ramakrishnan G proposed a change to openstack/ironic: Take iSCSI deploy out of pxe driver https://review.openstack.org/111232 | 09:45 |
| *** killer_prince is now known as lazy_prince | 09:46 | |
| rameshg87 | dtantsur, just had a question regarding mocking | 09:47 |
| rameshg87 | dtantsur, for testcases | 09:47 |
| *** dtantsur is now known as dtantsur|lunch | 09:53 | |
| *** lucasagomes has joined #openstack-ironic | 09:54 | |
| lucasagomes | hi all, are we going to have a meeting tonight? (It's holiday here) | 09:55 |
| *** sbadia has joined #openstack-ironic | 10:05 | |
| *** k4n0 has quit IRC | 10:16 | |
| *** sabah has joined #openstack-ironic | 10:17 | |
| *** rameshg87 has quit IRC | 10:22 | |
| *** rameshg87 has joined #openstack-ironic | 10:25 | |
| *** sabah has quit IRC | 10:26 | |
| *** sabah has joined #openstack-ironic | 10:27 | |
| *** k4n0 has joined #openstack-ironic | 10:30 | |
| *** sabah has quit IRC | 10:34 | |
| romcheg | Morning lucasagomes! | 10:53 |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ironic: Updated from global requirements https://review.openstack.org/111524 | 10:58 |
| *** ramineni has quit IRC | 11:03 | |
| *** dtantsur|lunch is now known as dtantsur | 11:07 | |
| dtantsur | welcome back lucasagomes, hi romcheg! | 11:09 |
| dtantsur | lucasagomes, I believe we have it | 11:09 |
| dtantsur | lucasagomes, also have a look: https://etherpad.openstack.org/p/Ironic-PoCDiscovery-Juno | 11:09 |
| *** pradipta is now known as pradipta_away | 11:17 | |
| *** nikunj2512 has quit IRC | 11:18 | |
| romcheg | dtantsur: Morning | 11:33 |
| *** Haomeng|2 has joined #openstack-ironic | 11:33 | |
| romcheg | dtantsur: I've seen you reviewed 97731 and made some suggestions there | 11:34 |
| *** Haomeng has quit IRC | 11:34 | |
| romcheg | dtantsur: I think there is even better solution that should simplify it all, please to not approve if Yongli uploads a new patch set. I'm working on the proposal :) | 11:35 |
| dtantsur | romcheg, ack, but you probably should bring it to review, so that other folks do not approve | 11:35 |
| romcheg | I will make it fast | 11:36 |
| romcheg | I was the one who wrote that crappy code so I think I know how to make it better now :) | 11:36 |
| dtantsur | heh :) | 11:37 |
| *** nosnos has quit IRC | 11:40 | |
| *** dhellmann_ has quit IRC | 11:42 | |
| *** dhellmann has joined #openstack-ironic | 11:42 | |
| *** rakesh_hs has joined #openstack-ironic | 11:47 | |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic-specs: DRAC vendor passthru for RAID management https://review.openstack.org/107981 | 11:55 |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic-specs: DRAC vendor passthru for RAID management https://review.openstack.org/107981 | 11:59 |
| *** lucasagomes is now known as lucas-holiday | 12:08 | |
| *** cian_ has joined #openstack-ironic | 12:11 | |
| *** toabctl_ has joined #openstack-ironic | 12:11 | |
| *** christop1eraedo has joined #openstack-ironic | 12:13 | |
| *** SpamapS_ has joined #openstack-ironic | 12:14 | |
| *** LiveOne_ has joined #openstack-ironic | 12:15 | |
| *** tteggel_ has joined #openstack-ironic | 12:16 | |
| romcheg | dtantsur: Could you please take a look at my proposal https://review.openstack.org/#/c/97731/ ? | 12:18 |
| *** faizan has joined #openstack-ironic | 12:18 | |
| dtantsur | romcheg, sure | 12:18 |
| *** bvivek has quit IRC | 12:19 | |
| *** athomas has quit IRC | 12:19 | |
| *** SpamapS has quit IRC | 12:19 | |
| *** toabctl has quit IRC | 12:19 | |
| *** christopheraedo has quit IRC | 12:19 | |
| *** LiveOne has quit IRC | 12:19 | |
| *** tteggel has quit IRC | 12:19 | |
| *** dricco has quit IRC | 12:19 | |
| *** toabctl_ is now known as toabctl | 12:19 | |
| dtantsur | romcheg, looks good from the 1st sight | 12:19 |
| *** LiveOne_ is now known as LiveOne | 12:20 | |
| *** rameshg87 has quit IRC | 12:20 | |
| romcheg | dtantsur: I've just tried to do something like that on my code that's why it took so much time to propose it :) I didn't update all the testing stuff but it seems to be what we need | 12:21 |
| *** athomas has joined #openstack-ironic | 12:23 | |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic-specs: DRAC hardware discovery https://review.openstack.org/108324 | 12:25 |
| *** krtaylor has quit IRC | 12:26 | |
| *** linggao has joined #openstack-ironic | 12:39 | |
| faizan | Hi jroll | 13:00 |
| *** jasondotstar has joined #openstack-ironic | 13:01 | |
| *** k4n0 has quit IRC | 13:01 | |
| jbjohnso | hmm, as of friday xCAT tree has an ironic driver in it | 13:03 |
| *** bvivek has joined #openstack-ironic | 13:06 | |
| jroll | faizan: hi! | 13:08 |
| jroll | Haomeng|2: jbjohnso maintains pyghmi and is often here :) | 13:09 |
| Haomeng|2 | jroll: got, thank you:) | 13:09 |
| Haomeng|2 | Haomeng|2: :) | 13:09 |
| jroll | :) | 13:10 |
| jroll | jbjohnso: see this? https://bugs.launchpad.net/ironic/+bug/1352079 | 13:10 |
| jbjohnso | ok, looking | 13:18 |
| Haomeng|2 | jbjohnso: thank you very much:) | 13:19 |
| jbjohnso | hmmm | 13:20 |
| Haomeng|2 | jbjohnso: :) | 13:21 |
| jbjohnso | going to take me a few minutes... trying to figure out when a numeric_format is set but the decode_formula not called.. | 13:21 |
| Haomeng|2 | jbjohnso: I am confusing why there is some undefined vars in sdr.py, such as self.tolerance, self.m, etc | 13:21 |
| jbjohnso | it's been a while | 13:21 |
| jbjohnso | those are defined in 'decode_formula' | 13:22 |
| Haomeng|2 | jbjohnso: no rush, thanks | 13:22 |
| Haomeng|2 | jbjohnso: ok, got | 13:22 |
| jbjohnso | on my setup, decode_formula is tripped on all the ones of interest | 13:22 |
| Haomeng|2 | jbjohnso: how do you think if ipmitool sdr ouput is required for you to take as difference? | 13:22 |
| Haomeng|2 | jbjohnso: I can help if you want | 13:22 |
| jbjohnso | Haomeng|2, a raw sdr might be informative... | 13:23 |
| jbjohnso | I need to see if I hit bouth full and compact... | 13:23 |
| faizan | jroll: Regarding your review comment on uefi spec for ip-address based config file names | 13:24 |
| faizan | jroll: for elilo.efi bootloader | 13:24 |
| jroll | faizan: yes? I just read the explanation from victor, seems fine | 13:25 |
| jbjohnso | ok, I do have full and compact records in my setup.. | 13:25 |
| jbjohnso | Haomeng|2, quick question, does the identical thing happen if you do IPMIPASSWORD=whatever pyghmiutil bmcip whatevertheuseris sensors | 13:26 |
| openstackgerrit | Vladyslav Drok proposed a change to openstack/ironic: Remove gettextutils _ injection https://review.openstack.org/110634 | 13:26 |
| *** matty_dubs|gone is now known as matty_dubs | 13:26 | |
| faizan | jroll: thanks. | 13:26 |
| jroll | :) | 13:27 |
| Haomeng|2 | jbjohnso: you mean we can have a test with pyghmiutil command? | 13:27 |
| Haomeng|2 | jbjohnso: how to run this pyghmiutil? | 13:27 |
| faizan | jroll: shall I post a patch to address your other comments "nits" or wait for other reviewers to give their comments | 13:28 |
| jbjohnso | it's in the pyghmi pip install as a 'scripts' payload | 13:28 |
| Haomeng|2 | jbjohnso: let me have try with "pyghmiutil bmcip" | 13:28 |
| Haomeng|2 | jbjohnso: yes, got it | 13:28 |
| jbjohnso | I'm starting to think of what might cause it | 13:28 |
| jbjohnso | off to the ipmi spec... | 13:29 |
| jroll | faizan: I'm fine with merging it with those, but if you push another version please fix them | 13:29 |
| Haomeng|2 | jbjohnso: ok | 13:30 |
| faizan | jroll: sure I will do that. | 13:30 |
| jroll | :) | 13:30 |
| jbjohnso | Haomeng|2, ok, its' a full record with an interesting linearization... | 13:30 |
| jbjohnso | Halacs, in def full_decode | 13:30 |
| jbjohnso | can you add an else to the self.linearization | 13:30 |
| jbjohnso | and print the value of self.linearization | 13:30 |
| jbjohnso | and tell me what gets printed | 13:31 |
| Haomeng|2 | jbjohnso: http://paste.openstack.org/show/89993/ | 13:31 |
| faizan | Hi NobodyCam, dtantsur, devananda: If you have some time can you please review uefi spec - https://review.openstack.org/#/c/99850 | 13:32 |
| Haomeng|2 | jbjohnso: ok, let me add print all | 13:32 |
| jbjohnso | Haomeng|2, go ahead and use your test script | 13:32 |
| Haomeng|2 | jbjohnso: yes | 13:32 |
| Haomeng|2 | jbjohnso: how can I print all values for 'reading' object? | 13:34 |
| jbjohnso | I was saying go into the full_decode() def | 13:35 |
| jbjohnso | and where it says if linearization <= 11 | 13:36 |
| jbjohnso | add an else | 13:36 |
| jbjohnso | to print what self.linearization is | 13:36 |
| Haomeng|2 | jbjohnso: print repr(reading)? | 13:36 |
| jbjohnso | well, I just wanted self.linearization | 13:36 |
| jbjohnso | which is just a number | 13:36 |
| Haomeng|2 | jbjohnso: ok | 13:36 |
| *** krtaylor has joined #openstack-ironic | 13:36 | |
| jbjohnso | but repr(reading) will try to dump it all | 13:36 |
| jbjohnso | in your test script | 13:36 |
| jbjohnso | reading.value, states, units, imprecision, name, type, and health | 13:37 |
| Haomeng|2 | jbjohnso: http://paste.openstack.org/show/89995/ | 13:37 |
| Haomeng|2 | jbjohnso: ok | 13:37 |
| jbjohnso | health is something I think ipmitool shies away from but I'll be ambitious and judgemental | 13:38 |
| Haomeng|2 | jbjohnso: let me print reading.value, states, units, imprecision, name, type, and health | 13:38 |
| jbjohnso | yeah, Ijust need info about the nonlinear record | 13:38 |
| jbjohnso | that would bail out on that if self.linearization <= 11 | 13:39 |
| romcheg | Morning Haomeng|2 jbjohnso! | 13:40 |
| jbjohnso | romcheg, | 13:41 |
| jbjohnso | morning | 13:41 |
| Haomeng|2 | jbjohnso: http://paste.openstack.org/show/89998/ | 13:41 |
| Haomeng|2 | romcheg: morning:) | 13:42 |
| Haomeng|2 | jbjohnso: is this result you want? | 13:42 |
| jbjohnso | Haomeng|2, well, was looking for a modification of the instance of sdr.py | 13:42 |
| jbjohnso | edit sdr.py | 13:42 |
| Haomeng|2 | jbjohnso: ok | 13:42 |
| jbjohnso | if self.linearization <= 11 | 13:43 |
| jbjohnso | add an else to it | 13:43 |
| Haomeng|2 | jbjohnso: ok | 13:43 |
| jbjohnso | #TODO(jbjohnso): implement get sensor reading factors support for | 13:43 |
| jbjohnso | that's the todo I'e struct upon, but I need to obviously flag entry to that TODO better | 13:43 |
| Haomeng|2 | jbjohnso: ok | 13:44 |
| *** shausy has quit IRC | 13:44 | |
| *** faizan has quit IRC | 13:47 | |
| Haomeng|2 | jbjohnso: I add else branch under the line ' if self.linearization <= 11', but looks like not run into else and print information | 13:48 |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Do get sensor reading factors for unrecognized linearizations https://review.openstack.org/111731 | 13:49 |
| Haomeng|2 | jbjohnso: the chaged code is - http://paste.openstack.org/show/90000/ | 13:49 |
| jbjohnso | print self.linearization | 13:49 |
| jbjohnso | but other than that, perfect | 13:49 |
| jbjohnso | and after that, if you try the above change, you should get a *different* exception | 13:50 |
| jbjohnso | I guess I should marke it as WIP | 13:50 |
| jbjohnso | oh well, I'll try to get it done soon... | 13:50 |
| jbjohnso | basically I have to ask the BMC how to decode that lovely sensor reading I got | 13:51 |
| *** lazy_prince is now known as killer_prince | 13:55 | |
| Haomeng|2 | jbjohnso: ok | 13:57 |
| *** sabah has joined #openstack-ironic | 14:00 | |
| Haomeng|2 | jbjohnso: http://paste.openstack.org/show/90002/ | 14:03 |
| Haomeng|2 | jbjohnso: looks like did not run into else branch | 14:03 |
| Haomeng|2 | jbjohnso: brb | 14:04 |
| NobodyCam | good morning Ironic | 14:05 |
| Shrews | morning NobodyCam | 14:07 |
| linggao | good morning NobodyCam, Shrews, Haomeng|2 jbjohnso jroll, | 14:09 |
| linggao | I am back from the vacation. | 14:09 |
| Haomeng|2 | linggao: morning:) | 14:09 |
| Haomeng|2 | linggao: :) | 14:10 |
| linggao | Haomeng|2, hi | 14:10 |
| dtantsur | morning NobodyCam, Shrews, linggao, jroll and everyone else :) | 14:10 |
| Haomeng|2 | jbjohnso: one more question, can we support 'impitool sdr -v' with '-v' option to get sensor details | 14:10 |
| linggao | morning dtantsur | 14:11 |
| Haomeng|2 | jbjohnso: we have such requirement for our 'send sensor data to ceilometer' bp | 14:11 |
| Haomeng|2 | jbjohnso: I have worked out the impitool driver with command | 14:11 |
| Haomeng|2 | jbjohnso: 'impitool sdr -v' | 14:11 |
| Haomeng|2 | jbjohnso: but looks like we have no such 'sdr -v' function in our pyghmi sensor command | 14:12 |
| dtantsur | bug triager dashboard: http://ironic-bugs.divius.net/ any ideas what else to put there? | 14:12 |
| Haomeng|2 | linggao: any idea about my question here which is same with the mail sent to you:) | 14:12 |
| linggao | Haomeng|2, I have not gotten a chance to read your mail yet. Let me check now.... | 14:13 |
| Haomeng|2 | linggao: ok, no rush:) | 14:13 |
| Haomeng|2 | linggao: nice day | 14:13 |
| Haomeng|2 | linggao: I will go to sleep:) | 14:14 |
| Haomeng|2 | linggao: :) | 14:14 |
| jbjohnso | Haomeng|2, I'll probably have something in git review when you wake up | 14:14 |
| linggao | Haomeng|2, good night. I' ll respond to your mail. | 14:14 |
| jbjohnso | to try to see if you get sensor readings. | 14:14 |
| Haomeng|2 | linggao: thank you:) | 14:14 |
| Haomeng|2 | jbjohnso: ok | 14:14 |
| jbjohnso | Haomeng|2, and how many of sdr -vfields are actually being used? | 14:15 |
| Haomeng|2 | jbjohnso: add me into the path review, thank you | 14:15 |
| jbjohnso | Halacs, if you could enumerate which ones, I'll add them as attributes exposed in the reading oject | 14:15 |
| jbjohnso | Haomeng|2, I meant you... | 14:15 |
| jbjohnso | which parts of sdr -v | 14:15 |
| Haomeng|2 | jbjohnso: ceilometer take 'sensor reading' as the current value, and for other fields as the meta data | 14:15 |
| Haomeng|2 | jbjohnso: let me show you the result with '-v' | 14:15 |
| jbjohnso | does it do anything with the meta data? | 14:15 |
| jbjohnso | I see -v output | 14:15 |
| jbjohnso | but was wondering how much of it is really used. E.g. I assume they don't care about thresholds they could go and change | 14:16 |
| *** jgrimm has joined #openstack-ironic | 14:16 | |
| jbjohnso | I could see them caring about rendering a range for a sensor (nominal, upper critical, lower critical) | 14:16 |
| jbjohnso | but the rest of the data I'm wondering if they would bother doing anything or not | 14:17 |
| Haomeng|2 | jbjohnso: http://paste.openstack.org/show/85053/ | 14:17 |
| NobodyCam | morning Shrews linggao Haomeng|2 and jbjohnso | 14:17 |
| jbjohnso | hysteresis, message control, configurable thresholds | 14:17 |
| *** krtaylor has quit IRC | 14:17 | |
| Haomeng|2 | jbjohnso: I think we need, 'Sensor Reading', and 'Sensor ID' and 'Sensor Type' | 14:18 |
| Haomeng|2 | NobodyCam: morning | 14:18 |
| jbjohnso | Haomeng|2, that's already there | 14:18 |
| Haomeng|2 | jbjohnso: yes, the rest data ceilometer just store them into db | 14:19 |
| Haomeng|2 | jbjohnso: ok, thank you | 14:19 |
| jbjohnso | Haomeng|2, in your output | 14:19 |
| Haomeng|2 | jbjohnso: tired, will go to bed, nice day:) | 14:19 |
| jbjohnso | {'name': 'Planar 3.3V', 'imprecision': 0.007099999999999884, 'value': 3.2802000000000002, 'states': [], 'health': 0, 'units': 'V', 'type': 'Voltage'} | 14:19 |
| Haomeng|2 | jbjohnso: ? | 14:19 |
| jbjohnso | have a nice day | 14:19 |
| jbjohnso | 'type' is there | 14:19 |
| Haomeng|2 | jbjohnso: yes | 14:19 |
| Haomeng|2 | jbjohnso: yes | 14:19 |
| jbjohnso | sensor.name is there | 14:19 |
| jbjohnso | and reading, with the imprecision and units | 14:19 |
| Haomeng|2 | jbjohnso: ok | 14:19 |
| jbjohnso | anyway, have a good evening | 14:20 |
| jbjohnso | and let me know further attributes or if you want the numeric id too | 14:20 |
| Haomeng|2 | jbjohnso: I can mapping the current sensor command ouput to the same fields with '-v' option | 14:20 |
| Haomeng|2 | jbjohnso: ok, thank you | 14:20 |
| Haomeng|2 | jbjohnso: now we just get it works:) | 14:20 |
| Haomeng|2 | jbjohnso: ok, thank you | 14:20 |
| Haomeng|2 | jbjohnso: :) | 14:20 |
| jbjohnso | Haomeng|2, have a good evening and I'll try to put you on review for the non-linearizable sensor coping | 14:20 |
| Haomeng|2 | jbjohnso: thank you very much:) nice day:) | 14:21 |
| Haomeng|2 | jbjohnso: :) | 14:21 |
| *** rameshg87 has joined #openstack-ironic | 14:27 | |
| *** Mikhail_D_ltp has quit IRC | 14:28 | |
| *** rakesh_hs has quit IRC | 14:35 | |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic: Add DracDriver and its DracPower module https://review.openstack.org/104850 | 14:41 |
| Shrews | neat. http://specs.openstack.org/openstack/ironic-specs/ | 14:43 |
| *** jgrimm has quit IRC | 14:43 | |
| *** ryanpetrello has joined #openstack-ironic | 14:44 | |
| NobodyCam | :) | 14:50 |
| *** bvivek has quit IRC | 14:53 | |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Do get sensor reading factors for unrecognized linearizations https://review.openstack.org/111731 | 14:54 |
| *** ifarkas has quit IRC | 14:55 | |
| openstackgerrit | Vladyslav Drok proposed a change to openstack/ironic: Remove gettextutils _ injection https://review.openstack.org/110634 | 14:55 |
| dtantsur | NobodyCam, are we having a meeting today? | 14:56 |
| dtantsur | I suppose yes | 14:56 |
| NobodyCam | I think so | 14:56 |
| NobodyCam | I have a conf call atm | 14:57 |
| dtantsur | ok, see you then | 14:57 |
| *** dtantsur is now known as dtantsur|afk | 14:57 | |
| *** rakesh_hs has joined #openstack-ironic | 14:57 | |
| *** mdorman has joined #openstack-ironic | 14:59 | |
| *** SpamapS_ is now known as SpamapS | 15:01 | |
| *** SpamapS has quit IRC | 15:01 | |
| *** SpamapS has joined #openstack-ironic | 15:01 | |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Handle non-linear and unrecognized linearizations https://review.openstack.org/111731 | 15:02 |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Handle non-linear and unrecognized linearizations https://review.openstack.org/111731 | 15:04 |
| jbjohnso | JayF, lucas-holiday just let me know if/when you want to stop adding you as reviewers for my patches | 15:07 |
| *** rameshg87 has quit IRC | 15:23 | |
| *** jgrimm has joined #openstack-ironic | 15:25 | |
| openstackgerrit | Roman Prykhodchenko proposed a change to openstack/ironic: Add charset and engine settings to every table https://review.openstack.org/111402 | 15:29 |
| openstackgerrit | Roman Prykhodchenko proposed a change to openstack/ironic: Add charset and engine settings to every table https://review.openstack.org/111402 | 15:43 |
| *** eguz has joined #openstack-ironic | 15:45 | |
| *** Poornima has joined #openstack-ironic | 15:45 | |
| *** sabah has quit IRC | 15:50 | |
| *** rwsu has joined #openstack-ironic | 15:51 | |
| openstackgerrit | Roman Prykhodchenko proposed a change to openstack/ironic: Add charset and engine settings to every table https://review.openstack.org/111402 | 15:53 |
| openstackgerrit | Roman Prykhodchenko proposed a change to openstack/ironic: Add charset and engine settings to every table https://review.openstack.org/111402 | 15:58 |
| *** Mikhail_D_ltp has joined #openstack-ironic | 15:59 | |
| openstackgerrit | A change was merged to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/110860 | 16:00 |
| romcheg | Sorry for some spam :) | 16:00 |
| romcheg | I was playing with one of the changes | 16:01 |
| *** openstackgerrit has quit IRC | 16:01 | |
| *** openstackgerrit_ has joined #openstack-ironic | 16:03 | |
| *** openstackgerrit_ is now known as openstackgerrit | 16:03 | |
| NobodyCam | no need to be sorry :) | 16:09 |
| *** matty_dubs is now known as matty_dubs|lunch | 16:11 | |
| *** scubacuda has joined #openstack-ironic | 16:12 | |
| *** krtaylor has joined #openstack-ironic | 16:13 | |
| NobodyCam | call for agenda updates. anyone have anything they would like to add to the agenda | 16:15 |
| JayF | I will not be at the meeting, but I'd strongly suggest the sticking point on 97744 be figured out and noted in the meeting | 16:18 |
| romcheg | Morning guys! | 16:23 |
| romcheg | who tried to stop devstack and then run provisioning once again? | 16:24 |
| jroll | +1 JayF | 16:24 |
| jroll | romcheg: as in, ./unstack.sh && ./stack.sh? | 16:24 |
| romcheg | No, just reboot and ./rejoin-stack.sh | 16:24 |
| jroll | ah, I've never used that | 16:24 |
| romcheg | For me it all nodes hang in wait-callback now :( | 16:24 |
| jroll | hmm | 16:25 |
| jroll | are the "bare metal" VMs booting at all? | 16:25 |
| romcheg | Yes, I'm talking about ssh driver | 16:25 |
| romcheg | They get started | 16:25 |
| romcheg | but nothing in the logs | 16:25 |
| jroll | hmm | 16:25 |
| *** ellenh has joined #openstack-ironic | 16:26 | |
| jroll | I wonder if the tftp server is running | 16:26 |
| romcheg | I was going to check networking | 16:26 |
| NobodyCam | romcheg: tftp up? are the nodes getting ip via dhcp? | 16:26 |
| jroll | also, if the deploy ramdisk is still there | 16:26 |
| romcheg | but checking tftp sounds more reasonable | 16:26 |
| jroll | perhaps it was on a tmpfs | 16:26 |
| jroll | er no, the driver would grab it from glance again | 16:26 |
| *** bvivek has joined #openstack-ironic | 16:34 | |
| *** jistr has quit IRC | 16:40 | |
| romcheg | looks like the problem with bridges | 16:40 |
| romcheg | tftp is up but nodes don't boot | 16:41 |
| *** Mikhail_D_ltp has quit IRC | 16:41 | |
| *** pcrews has joined #openstack-ironic | 16:43 | |
| NobodyCam | ahh so its neutron that hasn't come back. | 16:45 |
| *** chuckC has quit IRC | 16:50 | |
| *** penick has joined #openstack-ironic | 16:56 | |
| victor_lowther | so, stupid question: What do I have to do in local.conf to test non-default ironic drivers in devstack? | 16:56 |
| jroll | victor_lowther: enabled_drivers= | 16:57 |
| *** penick has quit IRC | 16:57 | |
| NobodyCam | :) | 16:58 |
| *** Poornima has quit IRC | 16:58 | |
| victor_lowther | jroll; Yea, but where? Stuffing it under [[post-config|$IRONIC_CONF]] did not seem to work. | 16:59 |
| jroll | victor_lowther: in devstack? | 17:00 |
| victor_lowther | ya | 17:00 |
| victor_lowther | devstack's local.conf | 17:00 |
| jroll | ah | 17:00 |
| jroll | sec | 17:00 |
| jroll | victor_lowther: IRONIC_ENABLED_DRIVERS=fake,agent_ssh,agent_ipmitool | 17:00 |
| victor_lowther | Alrighty then. | 17:01 |
| *** penick has joined #openstack-ironic | 17:01 | |
| *** matty_dubs|lunch is now known as matty_dubs | 17:02 | |
| romcheg | NobodyCam: Do we have a meeting today? | 17:07 |
| *** pcrews has quit IRC | 17:07 | |
| NobodyCam | romcheg: si | 17:10 |
| *** rameshg87 has joined #openstack-ironic | 17:12 | |
| rameshg87 | jroll, hi | 17:15 |
| rameshg87 | jroll, need your reviews on ilo deploy drivers: https://review.openstack.org/#/c/97744/ and https://review.openstack.org/#/c/108445/ | 17:15 |
| jroll | heya rameshg87 | 17:15 |
| rameshg87 | jroll, please have a look | 17:16 |
| jroll | rameshg87: I'll get to it when I can | 17:16 |
| rameshg87 | jroll, thanks :-) | 17:16 |
| romcheg | NobodyCam: Grasias! | 17:16 |
| jroll | :) | 17:17 |
| JayF | rameshg87: I strongly suggest you take 97744 to the meeting | 17:17 |
| NobodyCam | brb | 17:17 |
| JayF | rameshg87: since there seems to be a misalignment between what we talked about here on Friday and what lifeless commented/voted | 17:17 |
| jroll | JayF: just added that to agenda | 17:18 |
| JoshNang | mrda-away: did you happen to find the PoC code for horizon? | 17:18 |
| jroll | NobodyCam: I added two discussion topics and a note about new specs cores to the agenda :) | 17:18 |
| rameshg87 | JayF, okay, i will take it to the meeting .. | 17:18 |
| *** pcrews has joined #openstack-ironic | 17:21 | |
| *** athomas has quit IRC | 17:24 | |
| *** shakamunyi has joined #openstack-ironic | 17:30 | |
| NobodyCam | jroll: :) | 17:30 |
| *** Haomeng has joined #openstack-ironic | 17:34 | |
| *** Haomeng|2 has quit IRC | 17:35 | |
| *** romcheg1 has joined #openstack-ironic | 17:38 | |
| rameshg87 | jroll, just had a couple of questions on the review for https://review.openstack.org/#/c/110560/2 | 17:42 |
| *** romcheg1 has quit IRC | 17:42 | |
| rameshg87 | jroll, regarding "I wonder if it would be good to also move the config options for `image_cache_ttl` and `image_cache_size` to the image_cache module." | 17:43 |
| jroll | rameshg87: I think we should do that | 17:43 |
| rameshg87 | jroll, image_cache_ttl and image_cache_size are specific to instances of ImageCache, right ? and the child classes are not defined in image_cache.py | 17:44 |
| jroll | I think they should be the same for all instances of ImageCache | 17:44 |
| rameshg87 | jroll, PXEImageCache() still lies on pxe.py | 17:44 |
| jroll | well, hmm | 17:44 |
| * jroll thinks | 17:44 | |
| rameshg87 | jroll, do you mean to assign them as default values for the parent ImageCache ? | 17:45 |
| jroll | rameshg87: you might be right | 17:46 |
| jroll | if they have different master_dir they should have independent sizes/tlls | 17:47 |
| jroll | ttls* | 17:47 |
| jroll | rameshg87: I think ignore that comment :) | 17:47 |
| rameshg87 | jroll, yeah okay | 17:47 |
| *** bvivek has quit IRC | 17:48 | |
| rameshg87 | jroll, and regarding the comment on tests for the same review 110560 | 17:56 |
| rameshg87 | jroll, i thought every portion of the code in image_cache.clean_up_caches() is covered well in the fetch_image testcase in test_pxe.py | 17:57 |
| rameshg87 | jroll, it has pretty good coverage and covers all cases. do you see anything that is not covered already ? | 17:58 |
| rameshg87 | jroll, it's just the way the tests were written already :-) | 17:58 |
| *** ndipanov has quit IRC | 17:59 | |
| jroll | rameshg87: it likely has good coverage, but I would like some dedicated tests as well | 18:00 |
| *** chuckC has joined #openstack-ironic | 18:00 | |
| rameshg87 | jroll, do you mean to add dedicated test for image_cache.clean_up_caches() ? | 18:00 |
| jroll | yes | 18:00 |
| rameshg87 | jroll, okay will do | 18:01 |
| jroll | thanks! | 18:01 |
| *** dtantsur|afk is now known as dtantsur | 18:02 | |
| Shrews | aaaarrgh. jenkins hates _all_ of my reviews today | 18:05 |
| Shrews | i think i've hit every issue on the rechecks page | 18:05 |
| * pcrews thought it was just him that jenkins disliked ;) | 18:09 | |
| * NobodyCam thinks Mr Jenkins dislikes everyone | 18:10 | |
| pcrews | http://issues.jenkins-ci.org/secure/attachmentzip/unzip/151847/24614%5B10%5D/word/media/image1.png | 18:12 |
| *** ellenh has quit IRC | 18:19 | |
| *** tatyana has joined #openstack-ironic | 18:21 | |
| Shrews | and here i thought pcrews was the devil, when it was jenkins this whole time | 18:21 |
| * pcrews muses that one of the devil's better tricks is misdirection, changes his name to pkeyser for a microsecond, and laughs maniacally | 18:24 | |
| NobodyCam | brb | 18:26 |
| *** rakesh_hs has quit IRC | 18:27 | |
| *** ellenh has joined #openstack-ironic | 18:33 | |
| * NobodyCam is very distracted by travel sickness :-p | 18:36 | |
| matty_dubs | All that honey-mustard popcorn? | 18:37 |
| jroll | lolol | 18:37 |
| NobodyCam | lol | 18:38 |
| jroll | damn you matty_dubs, now I want a pretzel | 18:38 |
| matty_dubs | O:-) | 18:38 |
| NobodyCam | lol | 18:38 |
| * jroll goes hunting for food | 18:38 | |
| * matty_dubs bad influence | 18:38 | |
| *** rloo has joined #openstack-ironic | 18:48 | |
| NobodyCam | ten minute beep | 18:49 |
| *** romcheg1 has joined #openstack-ironic | 18:53 | |
| *** romcheg1 has left #openstack-ironic | 18:54 | |
| *** igordcard has joined #openstack-ironic | 18:54 | |
| *** lucas-holiday is now known as lucasagomes | 18:55 | |
| devananda | o/ | 18:56 |
| devananda | hi all! Apparently I've developed pneumonia, thus my absense this morning | 18:57 |
| dtantsur | devananda, ouch... get better! | 18:58 |
| devananda | so I'm several days behind on emails and follow ups from the summit ... | 18:58 |
| lucasagomes | devananda, :( I was pretty bad as well | 18:58 |
| lucasagomes | I'm blaming the aircon | 18:58 |
| NobodyCam | oich | 18:58 |
| rloo | yikes, now I'm glad I didn't go to the meetup. Hope you both get better soon. | 18:58 |
| NobodyCam | ouch even | 18:58 |
| *** romcheg1 has joined #openstack-ironic | 18:59 | |
| NobodyCam | I hope thats not what I am getting | 18:59 |
| devananda | lucasagomes: ugh. I'm blaming RedHat. I got it in raleigh :p | 18:59 |
| lucasagomes | devananda, hah | 18:59 |
| lucasagomes | :( | 18:59 |
| devananda | well, I got a cold in raleigh. then I didn't rest, and it turned into this | 18:59 |
| NobodyCam | :( | 18:59 |
| devananda | anyway. I'll be semi-around for the meeting | 18:59 |
| lucasagomes | :( hope you get better soon | 18:59 |
| *** Mikhail_D_ltp has joined #openstack-ironic | 19:00 | |
| Shrews | ick. i may never travel again with all the diseases going around lately | 19:00 |
| devananda | Shrews: wise man | 19:01 |
| rloo | Shrews: except for the upcoming Paris summit, right? :-) | 19:01 |
| *** ellenh has quit IRC | 19:03 | |
| *** dwalleck has joined #openstack-ironic | 19:04 | |
| *** igordcard has quit IRC | 19:10 | |
| devananda | lucasagomes: new client tagged -- https://review.openstack.org/gitweb?p=openstack/python-ironicclient.git;a=tag;h=refs/tags/0.1.5 | 19:30 |
| lucasagomes | devananda, yay ta much! | 19:31 |
| *** eguz has quit IRC | 19:43 | |
| *** eghobo has joined #openstack-ironic | 19:48 | |
| *** eghobo has quit IRC | 19:59 | |
| *** linggao has quit IRC | 20:00 | |
| *** linggao has joined #openstack-ironic | 20:00 | |
| *** romcheg1 has quit IRC | 20:01 | |
| NobodyCam | awesome ty devananda | 20:01 |
| *** dwalleck has quit IRC | 20:01 | |
| *** wanyen has joined #openstack-ironic | 20:01 | |
| BadCub | feel better devananda! | 20:01 |
| *** ellenh has joined #openstack-ironic | 20:01 | |
| jroll | I'm not opposed to doing refactorings for this, but we need to decide the best way to do it, knowing what our longer-term goals are for this stuff | 20:02 |
| jroll | and it's unclear if this is the right direction | 20:02 |
| rloo | jroll: +1 | 20:02 |
| *** eghobo has joined #openstack-ironic | 20:03 | |
| jroll | I think I'd also like to see what the ilo patch ends up looking like, on top of this patch, I think that may help us decide what the right thing to do is | 20:03 |
| dtantsur | ok folks, I anyway can't add anything more to this discussion, so heading to bed :) | 20:03 |
| dtantsur | g'night | 20:03 |
| rloo | good meeting. ciao everyone, I'm off this week. See ya next Monday. | 20:03 |
| rloo | good night dtantsur! | 20:03 |
| jroll | have fun rloo :) | 20:03 |
| jroll | night dtantsur | 20:03 |
| rameshg87 | jroll, i can give a heads up of what ilo patch might look like | 20:03 |
| *** dtantsur is now known as dtantsur|afk | 20:04 | |
| jroll | rameshg87: that would be helpful :) | 20:04 |
| *** rloo has quit IRC | 20:04 | |
| * jroll bbiab | 20:05 | |
| lucasagomes | im back to my holiday :) | 20:05 |
| lucasagomes | have a great night everyone | 20:05 |
| *** lucasagomes is now known as lucas-afk | 20:05 | |
| lucas-afk | devananda, hope u get better soon! | 20:06 |
| devananda | lucas-afk: thanks! | 20:06 |
| wanyen | ramesh87> are u there? | 20:06 |
| devananda | lucas-afk: have a good holiday! | 20:06 |
| devananda | jroll: ++ | 20:06 |
| rameshg87 | jroll, it will give two classes like pxe,PXEDeploy() and pxe.VendorPassthru(), will strip off the pxe parts inbetween, will get in the ilo-vmedia specific code in | 20:06 |
| rameshg87 | wanyen, i am here | 20:06 |
| devananda | jroll: I think i can guess how they are reusing this code, bu tseeing it would help | 20:06 |
| rameshg87 | jroll, devananda, i can get you something more in a day or two, but since iscsi_deploy was the common part between the current pxe driver and proposed ilo-vmedia driver, we choose to go ahead this way | 20:08 |
| NobodyCam | brb | 20:08 |
| devananda | jroll: aiui, this is similar to the deploy_utils refactoring we did a while back | 20:08 |
| devananda | move common code into a util module so that it can be shared by >1 driver class | 20:08 |
| rameshg87 | jroll, ideally i hope contents of iscsi_deploy mature into the a implementation of a "deploy api" | 20:08 |
| rameshg87 | jroll, devananda, and the rest of the pxe code mature into a boot interface, AND then we can combine the two in the most efficient way so that boot and deploy mechanisms can be mix and matched easily | 20:09 |
| devananda | rameshg87: ++ | 20:10 |
| wanyen | ramesh87, the curent spec dos not call for this kind of refactoring so is it a longer term improvement for K? | 20:12 |
| rameshg87 | wanyen, yes, deploy/boot interface is a long term improvement | 20:13 |
| wanyen | ramesh87> ok. cuurent spec will provide good value but we can continue improving it for future releases | 20:14 |
| rameshg87 | devananda, jroll, and another concern was on the building of iso images - whether upto ironic to do it OR upto the operator. i guess only lifeless is the person disagreeing here. how do we handle that ? | 20:14 |
| devananda | rameshg87: does his most recent suggestion seem feasible to you? | 20:15 |
| devananda | rameshg87: my objection to your initial proposal for ironic to make images was that, IIUC, it would be writing to glance. | 20:15 |
| rameshg87 | devananda, yes, if you remember this was in the first proposal. | 20:16 |
| rameshg87 | devananda, instead lifeless proposes to put it in swift in ilo driver's container probably | 20:16 |
| devananda | rameshg87: exactly | 20:16 |
| devananda | rameshg87: https://review.openstack.org/#/c/97744/1/specs/ironic-ilo-virtualmedia-driver.rst lines 90 - 93 | 20:16 |
| devananda | rameshg87: that is what I objected to. Ironic should not modify glance images or metadata | 20:17 |
| jroll | devananda: I guess I just feel like the *pxe* code should be factored out, but I might be wrong | 20:17 |
| devananda | rameshg87: but if it is only uploading an object to swift, and that object can be re-used for subsequent boots of the same AMI, then I'm fine with it | 20:17 |
| jroll | or, not wrong, but that might make things harder short-term | 20:17 |
| devananda | jroll: harder how? | 20:17 |
| JayF | rameshg87: devananda" O | 20:18 |
| JayF | rameshg87: devananda: I'm OK with either approach, although I like the idea of thinking of 'image portability' as a feature we don't implement yet | 20:18 |
| rameshg87 | devananda, that seems feasible to me given that we have to change the contents of the first spec to now upload to glance, and instead put it on swift | 20:18 |
| rameshg87 | devananda, s/now/not/ | 20:19 |
| rameshg87 | devananda, the same machine image can be used to boot up machines - lifeless has suggested mechanism to creating a object name hashing kernel_id/ramdisk_id so that we have one boot_iso for each pair of kernel_id/ramdisk_id | 20:20 |
| jroll | devananda: *might* :) it's unclear to me without actually writing the code, I think | 20:20 |
| devananda | jroll: exactly. I agree that seeing the iLO driver code on top of that would help clarify it | 20:21 |
| jroll | right | 20:21 |
| *** linggao has quit IRC | 20:22 | |
| rameshg87 | devananda, jroll, would you like me to have the ilo driver code before the spec gets approved ? :-) | 20:22 |
| jroll | gah | 20:22 |
| devananda | jroll: but that's a lot to ask when the spec isn't approved yet :p | 20:22 |
| devananda | heh | 20:23 |
| jroll | yeah | 20:23 |
| jroll | details :) | 20:23 |
| devananda | JayF: I think it's less about portability between ironic drivers, and changing the requirements Nova places on users | 20:23 |
| JayF | devananda: wdym? | 20:24 |
| devananda | JayF: requiring a user, who is uploading their own image, to know to build a .iso image as well | 20:24 |
| wanyen | ramesh87, I learned that we can upload teh code before spec got approved. | 20:24 |
| devananda | JayF: when that's not anything Nova expects for any other driver today | 20:24 |
| devananda | JayF: is where I think lifeless' objection stems from | 20:25 |
| JayF | devananda: nova uses a kernel/ramdisk separately in other boot drivers? | 20:25 |
| JayF | devananda: or does pxe slurp the kernel/ramdisk out of an existing image as part of deploy? | 20:26 |
| *** Mikhail_D_ltp has quit IRC | 20:26 | |
| devananda | JayF: Nova with, say, libvirt can use ami/aki/ari trio just fine, and PXE driver also uses the same trio | 20:27 |
| JayF | Ah, okay | 20:28 |
| devananda | JayF: the objection here is that no existing nova driver uses an iso, so not only are we adding complexity for ironic's operators, we're exposing that in glance AND requiring users (who might want to make and upload their own bootable images) to understand it | 20:28 |
| JayF | okay, in the case I 100% agree with the objection and think we should abstract the booting detail (i.e. disk image vs pxe) from nova, which means Ironic has to generate the disk image | 20:29 |
| devananda | so | 20:29 |
| JayF | Is that the conclusion that everyone agrees with? or am I missing something? | 20:29 |
| devananda | I've also been paraphrasing a discusion i had with lifeless | 20:29 |
| devananda | without actually checking fo rmyself | 20:29 |
| devananda | i should not do that | 20:30 |
| devananda | http://docs.openstack.org/trunk/config-reference/content/iso-support.html | 20:30 |
| devananda | apparently nova alraedy supports ISOs just fine | 20:30 |
| JayF | Generally speaking, should an operator have to know when they're uploading a disk image what ironic driver is being used to provison it? | 20:30 |
| devananda | JayF: no | 20:31 |
| devananda | JayF: they may need to somehow limit what nodes that image can be used on, eg. via scheduler metadata | 20:31 |
| JayF | I guess the question then boils down to Operator complexity vs Ironic complexity | 20:32 |
| devananda | exactly | 20:32 |
| JayF | add complexity to Ironic (by building the image on the fly) or add complexity to the operator (you have to understand what images work with what drivers) | 20:32 |
| JayF | IMO we should take that hit on the Ironic side, but I have a sneaky suspicion there could be edge cases that'd make me regret that statement | 20:33 |
| JayF | lol | 20:33 |
| devananda | JayF: systems are better at handling this sort of complexity than humans, generally speaking | 20:33 |
| NobodyCam | JayF: would supporting image building like that affect uefi or trusted booting? | 20:34 |
| JayF | I tend to agree, but I'd be curious about where the line moves to | 20:34 |
| rameshg87 | devananda, jroll, this is how ilo driver will look like when reusing iscsi_deploy : https://gist.github.com/rameshg87/9764b2f70571402cd356 | 20:34 |
| rameshg87 | devananda, jroll, i just skipped the implementation of ilo things which you anyway won't be interested right now | 20:34 |
| devananda | JayF: in this case, for me, the line was at modifying anything in Glance -- I don't think Ironic should change existing image metadata | 20:34 |
| rameshg87 | devananda, but do you feel it's okay if we do it with swift instead as lifeless suggested | 20:35 |
| JayF | devananda: why doesn't us uploading the iso into our own swift container become essentially the same thing? | 20:35 |
| JayF | I want to understand why that is the distinction, because in the end we're still storing image data, just one way we do it via the openstack image service and other way we do it directly via swift... | 20:35 |
| jroll | rameshg87: thanks! maybe things are fine | 20:36 |
| devananda | JayF: one is a service modifying the artefacts that a user created, the other is a service storing artefacts for its own use | 20:36 |
| wanyen | NobodyCam, the 1st patch set of ilo spec handles both cases 1) iso image or 2) building iso out of aki and ari | 20:37 |
| lifeless | devananda: I'm not suggesting we change existing image metadatathough | 20:38 |
| jbjohnso | JayF, yeah, I don't know the best approach in cases like this... The code is dealing pretty deep in IPMI specification... Some test cases could be crafted to fake data from BMC, but I doubt that would clarify the underlying behavior... | 20:38 |
| devananda | lifeless: i know you're not | 20:38 |
| lifeless | ok | 20:38 |
| JayF | devananda: cool, that helps me understand why the line is there. I dig it. The only other question I'm left with is it OK to add a dep on swift for this driver? | 20:38 |
| devananda | JayF: it already has a dep on swift - so that's not new | 20:38 |
| JayF | ah, well that makes it easy then | 20:39 |
| devananda | :) | 20:39 |
| rameshg87 | JayF, it already has to upload the floppy image to swift to pass token and parameters to ramdisk | 20:39 |
| devananda | rameshg87: i think we have all agreed now | 20:39 |
| JayF | rameshg87: ah okay, yeah | 20:39 |
| rameshg87 | devananda, okay. so we build iso images and upload to swift. right ? | 20:39 |
| lifeless | devananda: aahh iso support is a thing now. sorry ;) | 20:39 |
| jbjohnso | btw, if a component handled injecting boot payload into service processors (rather than having service processor fetch) | 20:39 |
| JayF | I honestly would've been OK with either way :) I wanted to have the chat to understand why we picked one way over the other | 20:39 |
| devananda | rameshg87: if there is an ISO associated to the AMI, use that; if there is no ISO, check swifth for the predictably-named object; if that is not found, create it and upload to swift | 20:40 |
| lifeless | devananda: but! thats iso-as-root, no? | 20:40 |
| lifeless | devananda: not iso-as-boot | 20:40 |
| wanyen | ramesh87> Nobdaycam has a question about whetehr buiding iso one the fly might have imact on uefi boot and secure boot. We should take a look at that. | 20:40 |
| JayF | jbjohnso: I really think it would help to see the inputs and expected outputs in a unit-test style, even moreso if the other place to see what it's doing would be an IPMI spec :) | 20:40 |
| devananda | lifeless: ISO as the boot image for the instance | 20:40 |
| wanyen | s/one/on | 20:40 |
| lifeless | devananda: its moot; I'll drill in later to see what exactly landed | 20:41 |
| devananda | lifeless: https://wiki.openstack.org/wiki/BootFromISO#Nova_Changes | 20:41 |
| jbjohnso | JayF, well at least there is a lovely ipmi spec. A bit big, but at least it is more specific than some other specs... | 20:41 |
| lifeless | thanks | 20:41 |
| * devananda goes back to bed now | 20:42 | |
| *** penick has quit IRC | 20:42 | |
| jbjohnso | I might want to at the very least see if I can cross reference something in the spec to explain it all... | 20:42 |
| JayF | devananda: feel better :( | 20:42 |
| lifeless | ok so we should support that but as a separate thing for all drivers IMO :) | 20:42 |
| lifeless | devananda: ugh I'm still sick too | 20:42 |
| lifeless | devananda: it sucks | 20:42 |
| lifeless | ah I've caught up more fully with backscroll, cool | 20:43 |
| jbjohnso | the biggest question is how many people want to party with me in the muck of obscure binary data... | 20:43 |
| jbjohnso | so far people just think me perverse for fiddling in that neighborhood | 20:43 |
| JayF | jbjohnso: I personally wouldn't want to... but honestly even if I did it'd be intimidating to start without having tests to help know what's going on. That's usually my primary way of learning my way around a new codebase, seeing what the tests call and how it works, without that it's a little tough. | 20:44 |
| rameshg87 | lifeless, JayF, devananda, anyone around. just had one more question. | 20:53 |
| NobodyCam | shoot | 20:53 |
| rameshg87 | NobodyCam, finally devananda pointed this " if there is an ISO associated to the AMI, use that; if there is no ISO, check swifth for the predictably-named object; if that is not found, create it and upload to swift" . so ideally we would support both the things ?? | 20:54 |
| rameshg87 | NobodyCam, both the things as in - 1) user uploading the iso , 2) if user doesn't upload the iso, ironic creating the iso | 20:54 |
| NobodyCam | yes thats how I took that | 20:54 |
| rameshg87 | NobodyCam, okay, just wanted to confirm :-) | 20:55 |
| NobodyCam | :) | 20:55 |
| * NobodyCam steps afk... in an attemp to fight off the plague with rest | 20:58 | |
| *** matty_dubs is now known as matty_dubs|gone | 20:58 | |
| lifeless | NobodyCam: rameshg87: I don't think support the user uploading the ISO as a linked image no | 21:00 |
| lifeless | NobodyCam: rameshg87: because that would be new in Nova - nova iso support is 'the named image is an ISO' | 21:00 |
| *** ellenh has quit IRC | 21:01 | |
| lifeless | NobodyCam: rameshg87: I think we should support that too, of course. | 21:01 |
| lifeless | NobodyCam: rameshg87: but a future iteration | 21:01 |
| *** ellenh has joined #openstack-ironic | 21:01 | |
| rameshg87 | lifeless, so for now, are you suggesting to just leave ironic to create the ISOs ? | 21:02 |
| *** penick has joined #openstack-ironic | 21:04 | |
| *** jasondotstar has quit IRC | 21:08 | |
| lifeless | yes | 21:10 |
| lifeless | I can't see on the nova docs or the wiki page for isos a way to do iso-for-boot-and-separate-image-for-root-fs | 21:11 |
| openstackgerrit | Ramakrishnan G proposed a change to openstack/ironic-specs: iLO Virtual Media iSCSI Deploy Driver https://review.openstack.org/97744 | 21:23 |
| rameshg87 | lifeless, okay. i just left that part for now, may be we can add this later once nova supports this. | 21:24 |
| lifeless | rameshg87: yeah | 21:25 |
| openstackgerrit | Ramakrishnan G proposed a change to openstack/ironic-specs: iLO Virtual Media iSCSI Deploy Driver https://review.openstack.org/97744 | 21:27 |
| *** rameshg87 is now known as rameshg87_afk | 21:33 | |
| *** rameshg87_afk has quit IRC | 21:38 | |
| *** tatyana has quit IRC | 21:41 | |
| jroll | can anyone point me to our current tempest job and/or instructions on running it locally? | 21:44 |
| *** dhellmann is now known as dhellmann_ | 21:53 | |
| *** eghobo has quit IRC | 22:09 | |
| *** eghobo has joined #openstack-ironic | 22:10 | |
| wanyen | Jroll, I have a question regarding switch vlan config that IPA is using. Is it targeted for Juno? | 22:25 |
| JayF | I don't even think we have that up in a spec yet iirc | 22:25 |
| wanyen | JayF: it will be neutron spec right? | 22:27 |
| JoshNang | wanyen: JayF yeah, some of that code was in the original agent patch, but the latest hasn't been upstreamed yet (relies on a non-merged neutron plugin) | 22:27 |
| jroll | wanyen: someone else is working on a neutron thing to generically do what we want to do; that is targeted for juno | 22:28 |
| jroll | wanyen: ironic support will be K | 22:28 |
| *** lucas-afk has quit IRC | 22:29 | |
| wanyen | Jroll: that's good to know. Do you have the url link or pointer? | 22:29 |
| jroll | wanyen: here is the neutron spec https://blueprints.launchpad.net/neutron/+spec/neutron-external-attachment-points | 22:29 |
| wanyen | Jroll: thanks! | 22:29 |
| jroll | wanyen: here is the neutron code we're using today: https://github.com/rackerlabs/ironic-neutron-plugin | 22:29 |
| jroll | you're welcome :) | 22:30 |
| Haomeng | jbjohnso: hi, are you still around? | 22:33 |
| Haomeng | jbjohnso: I try with your patch - https://review.openstack.org/#/c/111731, still not working | 22:34 |
| Haomeng | jbjohnso: how can I help you to provid more debug info if you want? | 22:35 |
| *** radsy has joined #openstack-ironic | 22:43 | |
| *** radsy has joined #openstack-ironic | 22:43 | |
| *** jgrimm has quit IRC | 22:44 | |
| *** f13o_ has quit IRC | 23:04 | |
| Haomeng | jbjohnso: add comments with your patch, thanks | 23:06 |
| *** penick has quit IRC | 23:16 | |
| *** rwsu has quit IRC | 23:19 | |
| *** rwsu has joined #openstack-ironic | 23:40 | |
| jroll | this makes IPA work "out of the box" by the same standards that say pxe driver works out of the box with devstack: https://review.openstack.org/108457 | 23:53 |
| jroll | now someone help me figure out tempest :P | 23:53 |
| Shrews | jroll: whatcha need? | 23:54 |
| jroll | Shrews: where do I start with running tempest locally? | 23:54 |
| Shrews | jroll: devstack sets it up for you in /opt/stack/tempest | 23:54 |
| jroll | Shrews: don't feel like you have to help yet, I haven't dug too hard, you can tell me RTFM :) | 23:54 |
| Shrews | unless you've disabled it | 23:55 |
| jroll | ok | 23:55 |
| jroll | and then I just run run_tempest.sh? | 23:55 |
| jroll | or is there configuration etc I need to do? | 23:55 |
| Shrews | jroll: http://docs.openstack.org/developer/tempest/overview.html#quickstart | 23:56 |
| jroll | a ha | 23:56 |
| Shrews | jroll: so, a etc/tempest.conf should already exist after you start devstack | 23:56 |
| jroll | where do we store the config ironic typically uses? | 23:56 |
| Shrews | jroll: then just: testr init ; tester run <whatever.test> | 23:56 |
| jroll | ok | 23:56 |
| * jroll tries | 23:56 | |
| Shrews | s/tester/testr/ | 23:56 |
| Shrews | jroll: but... | 23:57 |
| Shrews | you'll need some stuff in your localrc for devstack first | 23:57 |
| jroll | yeah | 23:57 |
| Shrews | 1 sec | 23:58 |
| jroll | like, more stuff than I need to manually test ironic? | 23:58 |
| adam_g | jroll, only a certain subset of the tempest tests in trunk will pass against ironic. | 23:58 |
| Shrews | adam_g has a sample localrc somewhere... | 23:58 |
| jroll | adam_g: right... I'm assuming our gate job is in source control somewhere, and I can just grab things from that? | 23:58 |
| jroll | my end goal is making a check/gate job for IPA driver | 23:59 |
| jroll | as identical as possible to the existing job | 23:59 |
| adam_g | jroll, , https://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/jenkins_job_builder/config/devstack-gate.yaml#n542 that is what is running the tempest gate job that does the actauly ironic deployment now. note the fugly regex that makes testr target only tests we know to pass | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
