| naohirot | NobodyCam: and also Operator or deployer have to operate 2 set of Horizon Web GUI, one for virtual host such as KVM hypervisor, the other is for Ironic? | 00:00 |
|---|---|---|
| NobodyCam | naohirot: as of right there is no support for virtual and physical deploys | 00:00 |
| devananda | naohirot: it's possible to use different nova hypervisors in the same cloud by using host-aggregates, or cells | 00:02 |
| *** smoriya has joined #openstack-ironic | 00:02 | |
| devananda | however, I am not aware of any existing guides explaining how to do this with Ironic and libvirt | 00:02 |
| NobodyCam | lol I was just looking to see if there was any thing I could link on it | 00:02 |
| devananda | http://docs.openstack.org/openstack-ops/content/compute_nodes.html | 00:02 |
| devananda | Note | 00:02 |
| *** shakamunyi has joined #openstack-ironic | 00:02 | |
| devananda | It is also possible to run multiple hypervisors in a single deployment using host aggregates or cells. However, an individual compute node can run only a single hypervisor at a time. | 00:02 |
| naohirot | NobodyCam: In order for deployer to manage virtual and psychical, do they need to prepare 2 sets of OpenStack right now? | 00:03 |
| *** shakamunyi has quit IRC | 00:03 | |
| NobodyCam | they need to have setup cells or host aggregates | 00:04 |
| *** ryanpetrello has quit IRC | 00:04 | |
| NobodyCam | I have not gone thru and setup a cells install | 00:04 |
| NobodyCam | would be a great doc to have thou, I'd wager | 00:05 |
| naohirot | devananda: NobodyCam: I see, I'll take a look a the cells and host aggregation. | 00:05 |
| NobodyCam | :) | 00:06 |
| NobodyCam | naohirot: please let us know how it goes... | 00:06 |
| *** yuanying_ has quit IRC | 00:06 | |
| *** yuanying has joined #openstack-ironic | 00:07 | |
| *** david-lyle is now known as david-lyle_afk | 00:11 | |
| naohirot | NobodyCam: devananda: one thing I need to know is that Horizon Web GUI, is the Web GUI able to manage both virtual and physical if we use the cells and host aggregation? | 00:12 |
| *** andreykurilin_ has quit IRC | 00:12 | |
| devananda | naohirot: sorry, I do not know. you should ask the horizon developers that | 00:12 |
| *** andreykurilin_ has joined #openstack-ironic | 00:13 | |
| devananda | naohirot: that's not a great answer... and I'm happy to introduce you to them, but we haven't done much yet with horizon integration specifically for Ironic | 00:13 |
| devananda | naohirot: there is a Tuskar-UI plugin which has some integration between Horizon and Ironic. the developers of that hang out in #tripleo roomt | 00:14 |
| NobodyCam | I recall some very basic wire frame reviews but I am not at all sure if they actually landed | 00:14 |
| naohirot | devananda: I see | 00:15 |
| naohirot | NobodyCam: devananda: anyway I'll take a look at the cell and host aggregation first. thanks! | 00:15 |
| NobodyCam | :) happy to help | 00:15 |
| devananda | naohirot: actually, what I mean is, I don't believe there is currently support in Horizon for managing Ironic directly | 00:16 |
| devananda | naohirot: whethers Horizon may (or may not) support a multi-hypervisor Nova deployment, I can not say | 00:16 |
| NobodyCam | naohirot: https://blueprints.launchpad.net/horizon/+spec/node-management-ui | 00:17 |
| naohirot | devananda: If horizen cannnot manage multiple hyper-visors, I can easily imaging the case of hypervisors and ironic. This is very helpful information. | 00:19 |
| naohirot | s/imaging/imagine | 00:19 |
| *** Masahiro has joined #openstack-ironic | 00:19 | |
| *** jerryz has joined #openstack-ironic | 00:20 | |
| naohirot | devananda: NobodyCam: Anyway I understood that currently VM deploy and baremetal deploy haven't been integrated yet, we are on the way to the seamless integration, right? | 00:21 |
| devananda | naohirot: I do not understand what you mean. can you phrase the question differently? | 00:22 |
| *** yuanying has quit IRC | 00:23 | |
| naohirot | devananda: Okay, I'm thinking a use case, how can deployer to deploy a compute node where Ironic deploy a physical OS. | 00:23 |
| *** yuanying has joined #openstack-ironic | 00:23 | |
| naohirot | devananda: Is it possible right now? | 00:24 |
| naohirot | s/to deploy/deploy/ | 00:24 |
| NobodyCam | naohirot: with cells yes | 00:24 |
| *** Masahiro has quit IRC | 00:24 | |
| NobodyCam | but I'm not sure that horizon has full support for that. | 00:24 |
| devananda | naohirot: basically, you want Ironic to install OpenStack Nova on a server? | 00:24 |
| devananda | naohirot: or did I misunderstand? | 00:25 |
| *** igordcard has quit IRC | 00:25 | |
| NobodyCam | I was thinking naohirot was looking to do something like offer both vm and phy nodes to his customers | 00:25 |
| NobodyCam | bimw | 00:25 |
| NobodyCam | bibw * | 00:26 |
| devananda | NobodyCam: that's what I thought, too. not sure though | 00:26 |
| JayF | That definately works ... using cells :) | 00:26 |
| devananda | JayF: right. but then the question was, does Horizon support that, which I don't know | 00:27 |
| NobodyCam | JayF: do you know if horizon supports that setup atm | 00:27 |
| NobodyCam | lol | 00:27 |
| naohirot | devananda: I just wanted to know a concrete setup example of OpenStack, so any configuration is fine, I just wanted to know it's possible or not. | 00:27 |
| * NobodyCam point to JayF | 00:28 | |
| NobodyCam | *points | 00:28 |
| JayF | Given right now Ironic has no code in Horizon supporting it at all | 00:28 |
| JayF | I can answer that with a definitive no | 00:28 |
| JayF | lol | 00:28 |
| NobodyCam | naohirot: 00:26 | JayF > That definately works ... using cells :) | 00:28 |
| JayF | Combining Ironic + VMs in the same cloud :) | 00:28 |
| JayF | not using Horizon to manage them | 00:28 |
| * NobodyCam thinks he has never user it :( | 00:29 | |
| NobodyCam | gah | 00:29 |
| NobodyCam | used | 00:29 |
| * NobodyCam steps away | 00:29 | |
| naohirot | JayF: devananda: Yes, that's what I'd like to know | 00:30 |
| JayF | I think I've seen horizon running exactly once | 00:30 |
| naohirot | devananda: NobodyCam: JayF: I'll check the cell and the aggregation, thanks!! | 00:31 |
| *** igordcard has joined #openstack-ironic | 00:33 | |
| *** igordcard has quit IRC | 00:34 | |
| *** yuanying has quit IRC | 00:36 | |
| Haomeng | I run into this error "Node 65fe14d5-df35-4895-90ab-377a79c64375 is locked by host localhost.localdomain, please retry after the current operation is completed. (HTTP 409)", can we cancel the reservation by commands? | 00:42 |
| *** ryanpetrello has joined #openstack-ironic | 00:42 | |
| *** romcheg1 has quit IRC | 00:44 | |
| *** Masahiro has joined #openstack-ironic | 00:45 | |
| *** prontotest has joined #openstack-ironic | 00:49 | |
| *** prontotest has left #openstack-ironic | 00:49 | |
| NobodyCam | i think there is still a ope bug for that | 00:51 |
| NobodyCam | and morning Haomeng | 00:51 |
| Haomeng | NobodyCam: morning | 00:51 |
| Haomeng | NobodyCam: ok, let me check log and maybe we have gap for some logic for such state | 00:51 |
| NobodyCam | Haomeng: https://bugs.launchpad.net/ironic/+bug/1354147 | 00:52 |
| *** yuanying has joined #openstack-ironic | 00:52 | |
| Haomeng | NobodyCam: yes, same error behavior | 00:52 |
| NobodyCam | :( | 00:53 |
| Haomeng | NobodyCam: :) | 00:53 |
| jroll | so uh | 00:53 |
| jroll | naohirot: devananda, NobodyCam, why wouldn't horizon work with combined virt/metal using cells? | 00:53 |
| jroll | presumably those just show up as flavors | 00:54 |
| jroll | and the user clicks "boot an instance of virt||metal flavor | 00:54 |
| jroll | " | 00:54 |
| NobodyCam | jroll: have you used Horizon... I assume it would support.. But I can not say 100% it does | 00:54 |
| jroll | I haven't used it | 00:55 |
| jroll | but like | 00:55 |
| jroll | isn't it just a gui around novaclient? | 00:55 |
| jroll | for the purpose of booting instances? | 00:55 |
| naohirot | jroll: good evening | 00:56 |
| jroll | hi there | 00:56 |
| naohirot | jroll: everybody said that the cells and the host aggregation provide the way to handle both VM and bare metal. | 00:58 |
| jroll | right. they do. | 00:58 |
| naohirot | jroll: so I have to lean those first, :-) | 00:58 |
| jroll | we use cells in production, it totally works :) | 00:59 |
| jroll | Haomeng: by the way, that happens when a lock gets stuck, like if you restart a conductor while a node is locked | 00:59 |
| *** mikedillion has quit IRC | 01:00 | |
| *** mikedillion has joined #openstack-ironic | 01:01 | |
| naohirot | jroll: that's great, I thought that it would be severely inconvenient for deployer if we couldn't handle both. | 01:02 |
| jroll | naohirot: yep, we have it set up where ironic has its own cell, it's bee good to us for the most part | 01:03 |
| *** yuanying has quit IRC | 01:03 | |
| *** yuanying has joined #openstack-ironic | 01:04 | |
| naohirot | jroll: okay, I'll ask further question after learning what the cells and the host aggregation offer. Thanks! | 01:05 |
| *** yuanying has quit IRC | 01:05 | |
| jroll | cool, you're welcome :) | 01:05 |
| naohirot | jroll: :-) | 01:05 |
| * devananda wanders afk for the evening | 01:08 | |
| *** yuanying has joined #openstack-ironic | 01:10 | |
| *** rloo has quit IRC | 01:10 | |
| *** mikedillion has quit IRC | 01:14 | |
| *** ryanpetrello has quit IRC | 01:26 | |
| *** marcoemorais has quit IRC | 01:26 | |
| *** yuanying has quit IRC | 01:29 | |
| *** Masahiro has quit IRC | 01:31 | |
| * mrda is trying to work out how to add a field to the Ironic db, and handle migrations too | 01:31 | |
| *** chenglch has joined #openstack-ironic | 01:33 | |
| Shrews | mrda: http://docs.openstack.org/developer/ironic/cmds/ironic-dbsync.html#revision | 01:33 |
| *** Masahiro has joined #openstack-ironic | 01:34 | |
| mrda | thanks Shrews | 01:34 |
| *** spandhe has quit IRC | 01:34 | |
| *** Marga_ has quit IRC | 01:35 | |
| mrda | I was wondering how the revision id came about | 01:35 |
| mrda | Shrews: So at which point should I run ironic-dbsync? Once I've updtaed the models.py file, but before I try and start conductor? | 01:38 |
| Shrews | mrda: if you want to use the autogenerate stuff, yeah | 01:38 |
| mrda | "if I want"? | 01:39 |
| *** mitz- has quit IRC | 01:39 | |
| Shrews | change models.py, ironic-dbsync revision --autogenerate, ironic-dbsync upgrade (IIRC), then start conductor | 01:39 |
| mrda | Are the revision ids "special"? Or can I generate my own? (so long as they don't clash and refer to the last revision?) | 01:39 |
| Shrews | been a while since i did all that | 01:39 |
| mrda | ok, thanks Shrews | 01:39 |
| mrda | I'll have a play | 01:40 |
| Shrews | the revision id is autogenerated | 01:40 |
| Shrews | with the 'revision' command | 01:40 |
| Shrews | which then gives you a file you can use to 'upgrade' the db tables themselves | 01:40 |
| *** mitz_ has joined #openstack-ironic | 01:41 | |
| Shrews | i can't remember if 'revision' goes ahead and applies the changes. i don't think it does | 01:41 |
| Shrews | i think it just creates the update file | 01:41 |
| mrda | and it's the --message that gives the suffix for the migration script? i.e. 21b331f883ef_add_provision_updated_at.py would be ironic-dbsync revision --message add_provision_updated_at or similar? | 01:41 |
| Shrews | mrda: i think so | 01:42 |
| mrda | ta | 01:42 |
| Shrews | np | 01:42 |
| *** Marga_ has joined #openstack-ironic | 01:44 | |
| *** andreykurilin_ has quit IRC | 01:44 | |
| *** Marga_ has quit IRC | 01:48 | |
| Haomeng | jroll: ok, if we restart conductor again, if it will be released? | 01:51 |
| Haomeng | let me have test | 01:51 |
| Haomeng | jroll: after the conductor is restarted again, it still in reservation status, so should we catch the signal before conductor existing ? | 01:53 |
| *** shakamunyi has joined #openstack-ironic | 01:53 | |
| Haomeng | jroll: and try to wait or handle the nodes which are locked | 01:56 |
| *** yuanying has joined #openstack-ironic | 02:01 | |
| openstackgerrit | Merged openstack/ironic-specs: New Ironic provisioner state machine. https://review.openstack.org/133828 | 02:03 |
| *** yuanying has quit IRC | 02:10 | |
| *** harlowja has joined #openstack-ironic | 02:10 | |
| *** yuanying has joined #openstack-ironic | 02:10 | |
| *** Haomeng has quit IRC | 02:11 | |
| *** jb11211 has quit IRC | 02:13 | |
| *** r-daneel has quit IRC | 02:15 | |
| *** ChuckC has quit IRC | 02:18 | |
| *** Masahiro has quit IRC | 02:21 | |
| *** yuanying has quit IRC | 02:24 | |
| *** Masahiro has joined #openstack-ironic | 02:27 | |
| NobodyCam | w00 h00 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | 02:29 |
| *** ZhangYang has joined #openstack-ironic | 02:33 | |
| *** ZZhangYang has joined #openstack-ironic | 02:37 | |
| *** ZhangYang has quit IRC | 02:38 | |
| *** ZZhangYang has left #openstack-ironic | 02:39 | |
| *** ZZhangYang has joined #openstack-ironic | 02:39 | |
| *** ZZhangYang has quit IRC | 02:42 | |
| *** Haomeng has joined #openstack-ironic | 02:42 | |
| *** ZhangYang has joined #openstack-ironic | 02:43 | |
| *** yuanying has joined #openstack-ironic | 02:44 | |
| *** ramineni has joined #openstack-ironic | 02:44 | |
| *** yuanying has quit IRC | 02:46 | |
| *** yuanying has joined #openstack-ironic | 02:47 | |
| *** ryanpetrello has joined #openstack-ironic | 02:48 | |
| *** yuanying has quit IRC | 02:51 | |
| jroll | Haomeng: you need to manually unlock in the database | 02:52 |
| Haomeng | yes, I will:) | 02:52 |
| Haomeng | just concern for our user, how they can do | 02:53 |
| *** ChuckC has joined #openstack-ironic | 02:53 | |
| jroll | same thing :( | 02:53 |
| * jroll runs ironic in production and that is a pain point | 02:53 | |
| Haomeng | jroll: :) | 02:53 |
| jroll | good night Haomeng and everyone else :) | 02:54 |
| Haomeng | jroll: good night:) | 02:54 |
| *** ryanpetrello has quit IRC | 02:55 | |
| *** r-daneel has joined #openstack-ironic | 02:56 | |
| *** r-daneel has quit IRC | 03:00 | |
| *** jiangfei has left #openstack-ironic | 03:07 | |
| *** jiangfei has quit IRC | 03:07 | |
| *** jiangfei has joined #openstack-ironic | 03:07 | |
| *** ryanpetrello has joined #openstack-ironic | 03:10 | |
| *** harlowja is now known as harlowja_away | 03:12 | |
| *** ryanpetrello has quit IRC | 03:14 | |
| *** r-daneel has joined #openstack-ironic | 03:15 | |
| *** yuanying has joined #openstack-ironic | 03:16 | |
| *** kes has joined #openstack-ironic | 03:21 | |
| *** dlaube has quit IRC | 03:28 | |
| *** kes has quit IRC | 03:35 | |
| *** r-daneel has quit IRC | 03:50 | |
| *** kurtrao has quit IRC | 03:50 | |
| *** kurtrao has joined #openstack-ironic | 03:50 | |
| *** rameshg87 has joined #openstack-ironic | 03:58 | |
| *** kes has joined #openstack-ironic | 03:58 | |
| *** kes has quit IRC | 03:59 | |
| *** wanghua has joined #openstack-ironic | 04:01 | |
| *** kes has joined #openstack-ironic | 04:02 | |
| *** kes has quit IRC | 04:03 | |
| *** kurtrao has quit IRC | 04:13 | |
| *** kurtrao has joined #openstack-ironic | 04:14 | |
| *** jb11211 has joined #openstack-ironic | 04:21 | |
| rameshg87 | Haomeng, hello | 04:21 |
| *** jb11211 has quit IRC | 04:24 | |
| *** datajerk1 has joined #openstack-ironic | 04:30 | |
| *** datajerk has quit IRC | 04:32 | |
| *** achanda has joined #openstack-ironic | 04:35 | |
| *** dlaube has joined #openstack-ironic | 04:50 | |
| Haomeng | rameshg87: hi | 04:52 |
| rameshg87 | Haomeng, i had a questionn on IBM xCAT CI | 04:52 |
| rameshg87 | Haomeng, are you the contact for it ? :) | 04:52 |
| Haomeng | rameshg87: sure, i can help to find the contact | 04:53 |
| rameshg87 | Haomeng, i was checking the logs, but it seems to run a node with pxe_ssh driver | 04:53 |
| rameshg87 | Haomeng, i was expecting it is testing on real hardware | 04:53 |
| Haomeng | rameshg87: yes, it should be ipminative driver | 04:53 |
| Haomeng | rameshg87: let me check with xcat team | 04:53 |
| Haomeng | rameshg87: owned by xcat team | 04:54 |
| Haomeng | rameshg87: and response you once get the feedback | 04:54 |
| rameshg87 | Haomeng, okay | 04:54 |
| Haomeng | rameshg87: :) | 04:54 |
| rameshg87 | Haomeng, i just downloaded the log from it: http://sourceforge.net/projects/xcat/files/OpenStack/CI/139057/5/check/check-ironic-xcat-third-party/704/logs/ironic-conductor.log/download | 04:55 |
| Haomeng | ok | 04:55 |
| Haomeng | rameshg87: let me forward the log to xcat team | 04:55 |
| rameshg87 | Haomeng, when i checked the conductor log, i see a node registered with pxe_ssh driver and then deployed | 04:55 |
| rameshg87 | Haomeng, i couldn't see any node registered with pxe_ipmitool or pxe_ipminative driver | 04:56 |
| rameshg87 | Haomeng, hence my question | 04:56 |
| rameshg87 | Haomeng, may be you can just check the same point with xcat team | 04:56 |
| Haomeng | ok | 04:56 |
| *** achanda has quit IRC | 04:59 | |
| Haomeng | rameshg87: I get response from xcat, they are testing some cases with ssh_pxe driver temporary | 04:59 |
| rameshg87 | Haomeng, okay, so it's not real hardware right now, am i correct ? | 05:00 |
| Haomeng | rameshg87: and they will switch to pxe_ipminative driver in recently days | 05:00 |
| Haomeng | rameshg87: thaks for your information:) | 05:00 |
| rameshg87 | Haomeng, okay so they will switch to pxe_ipminative soon | 05:00 |
| rameshg87 | Haomeng, we are planning to start CI for hp hardware and we wanted some info on how many bare metals are you dedicating for CI ? | 05:01 |
| *** achanda has joined #openstack-ironic | 05:01 | |
| rameshg87 | Haomeng, may be we will get that info later on | 05:01 |
| Haomeng | rameshg87: let me forward these question to xcat | 05:03 |
| *** achanda has quit IRC | 05:05 | |
| chenglch | rameshg87: hi, now we have four case test on the real machine , like test__power_status test_set_boot_device_pxe | 05:05 |
| rameshg87 | chenglch, hi | 05:05 |
| rameshg87 | chenglch, okay | 05:05 |
| chenglch | we just change the nova boot to pxe_ssh temporary | 05:06 |
| rameshg87 | chenglch, okay | 05:06 |
| *** killer_prince is now known as lazy_prince | 05:06 | |
| rameshg87 | chenglch, just wanted to know how many machines you dedicate for xcat - so that we will have an fair idea how many bare metals are required for us too | 05:06 |
| rameshg87 | chenglch, we plan to run the deploy of our deploy drivers - iscsi_ilo which will take approximately same time as pxe_ilo driver | 05:07 |
| chenglch | now , we only have one baremetal machine | 05:07 |
| rameshg87 | chenglch, okay | 05:07 |
| rameshg87 | chenglch, i see xcat posts results in the same speed as that of other CI testing | 05:07 |
| rameshg87 | chenglch, so i guess xcat CI is comfortably running with one bare metal, right ? | 05:08 |
| rameshg87 | chenglch, i mean comfortably running deploy with one bare metal | 05:08 |
| chenglch | rameshg87: yes only one ,sometimes a little late response | 05:08 |
| rameshg87 | chenglch, okay | 05:08 |
| rameshg87 | chenglch, thanks a lot for the info | 05:08 |
| *** Marga_ has joined #openstack-ironic | 05:09 | |
| *** Marga_ has quit IRC | 05:13 | |
| *** Marga_ has joined #openstack-ironic | 05:14 | |
| *** pensu has joined #openstack-ironic | 05:16 | |
| *** pensu has quit IRC | 05:17 | |
| *** pensu has joined #openstack-ironic | 05:18 | |
| *** achanda has joined #openstack-ironic | 05:19 | |
| *** achanda has quit IRC | 05:26 | |
| *** harlowja_away is now known as harlowja | 05:28 | |
| *** harlowja is now known as harlowja_away | 05:29 | |
| *** jrist has quit IRC | 05:30 | |
| *** pradipta_away is now known as pradipta | 05:33 | |
| *** kurtrao has quit IRC | 05:33 | |
| *** kurtrao has joined #openstack-ironic | 05:33 | |
| *** Marga_ has quit IRC | 05:36 | |
| kurtrao | exit | 05:38 |
| *** kurtrao has quit IRC | 05:38 | |
| *** Marga_ has joined #openstack-ironic | 05:38 | |
| *** masco has joined #openstack-ironic | 05:38 | |
| *** rushiagr_away is now known as rushiagr | 05:40 | |
| *** jrist has joined #openstack-ironic | 05:42 | |
| *** achanda has joined #openstack-ironic | 05:52 | |
| *** achanda has quit IRC | 05:55 | |
| *** achanda has joined #openstack-ironic | 05:56 | |
| *** Marga_ has quit IRC | 05:56 | |
| *** Marga_ has joined #openstack-ironic | 05:56 | |
| *** jiangfei has quit IRC | 05:59 | |
| *** jiangfei has joined #openstack-ironic | 05:59 | |
| *** achanda has quit IRC | 06:00 | |
| *** jrist has quit IRC | 06:09 | |
| *** dlaube has quit IRC | 06:09 | |
| jiangfei | hi,Haomeng | 06:16 |
| Haomeng | jiangfei: hi | 06:17 |
| Haomeng | jiangfei: good afternoon | 06:17 |
| jiangfei | Haomeng: good afternoon. | 06:17 |
| jiangfei | Haomeng: today, i want to add plugin-driver in conductor tablse, so it can be checked by api, what about it | 06:18 |
| *** jrist has joined #openstack-ironic | 06:22 | |
| *** kes has joined #openstack-ironic | 06:24 | |
| *** kes has quit IRC | 06:24 | |
| *** kes has joined #openstack-ironic | 06:24 | |
| *** yuanying_ has joined #openstack-ironic | 06:24 | |
| *** kesmellon has joined #openstack-ironic | 06:24 | |
| *** kes has quit IRC | 06:24 | |
| Haomeng | jiangfei: it is auto-added during conductor starting | 06:24 |
| *** kesmellon has quit IRC | 06:24 | |
| Haomeng | jiangfei: if it is defined as python ext in entry_points | 06:25 |
| *** kes has joined #openstack-ironic | 06:25 | |
| Haomeng | jiangfei: setup.cnf define these extensions | 06:25 |
| Haomeng | jiangfei: setup.cfg entry_points sections | 06:25 |
| *** yuanying_ has quit IRC | 06:25 | |
| Haomeng | yjiangfei: you can add your new driver into ironic.drivers list | 06:26 |
| Haomeng | jiangfei: and run setup.py install again or change the python egginfo file directly, then restart conductor | 06:26 |
| Haomeng | jiangfei: monitor the log to see if your driver is loaded | 06:27 |
| jiangfei | Haomeng: i kown, but these drivers not in enable_dirviers, it doesn't work. | 06:27 |
| *** yuanying has quit IRC | 06:28 | |
| *** yuanying has joined #openstack-ironic | 06:28 | |
| Haomeng | jiangfei: yes, add it into ironic.conf enabled_drivers | 06:29 |
| jiangfei | Haomeng: so i want add plugin_drivers element in conductor tables, like this: mysql> select * from conductors; | 06:33 |
| jiangfei | +---------------------+---------------------+----+------------+-------------------+------------------------------------------ +--------+ | 06:33 |
| jiangfei | | created_at | updated_at | id | hostname | drivers | plugin_drivers | online | | 06:33 |
| jiangfei | +---------------------+---------------------+----+------------+-------------------+------------------------------------------ +--------+ | 06:33 |
| jiangfei | | 2014-11-28 06:28:56 | 2014-12-10 03:58:11 | 1 | test-host | ["fake_ipmitool"] | ["fake_ipmitool, pxe_ssh, pxe_ipmitool...."] | 0 | | 06:33 |
| jiangfei | | 2014-12-10 02:30:14 | 2014-12-10 03:58:11 | 2 | test-host1 | ["fake_ssh"] | ["fake_ssh","pxe_Iboot", "pxe_ipminative"] | 0 | | 06:33 |
| jiangfei | +---------------------+---------------------+----+------------+-------------------+------------------- +--------+ | 06:33 |
| Haomeng | jiangfei: you can paste it via paste.openstack.org to share | 06:33 |
| Haomeng | jiangfei: it is not added by manually | 06:34 |
| Haomeng | jiangfei: 1. make sure it is in setup.cfg entry_points sections | 06:34 |
| Haomeng | jiangfei: 2 make sure it is enabled by ironic.conf enabled_drivers | 06:34 |
| jiangfei | Haomeng: yes | 06:35 |
| Haomeng | jiangfei: then restart conductor and monitor the log to see if/how it is loaded | 06:35 |
| *** trown has quit IRC | 06:36 | |
| *** krtaylor has quit IRC | 06:36 | |
| *** trown has joined #openstack-ironic | 06:37 | |
| *** krtaylor has joined #openstack-ironic | 06:38 | |
| jiangfei | Haomeng: yes, i want to inserted these ironic.dirviers to database,not only ironic.conf enabled_drivers | 06:38 |
| Haomeng | jiangfei: the new driver is auto-loaded, dont insert into tables by manually, dont think conductor will pick them up fine | 06:39 |
| jiangfei | Haomeng: ii dont want to insert into tables by manually. | 06:45 |
| Haomeng | jiangfei: yes, confgurate new driver then it will be auto-loaded by conductor | 06:45 |
| *** kes has quit IRC | 06:46 | |
| *** kes has joined #openstack-ironic | 06:49 | |
| rameshg87 | jiangfei, after adding your new driver to setup.cfg, you can run "python setup.py egg_info". this will change ironic.egg-info/entry_points.txt | 06:52 |
| jiangfei | rameshg87: thks | 06:53 |
| kes | 用中文吧 | 06:54 |
| *** mrda is now known as mrda-away | 06:55 | |
| *** Haomeng has quit IRC | 07:05 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: Add AMT-PXE-Driver to deploy cloud on PC https://review.openstack.org/135184 | 07:06 |
| *** Haomeng has joined #openstack-ironic | 07:07 | |
| *** k4n0 has joined #openstack-ironic | 07:25 | |
| *** Marga_ has quit IRC | 07:41 | |
| *** ndipanov_gone is now known as ndipanov | 07:54 | |
| *** ndipanov has quit IRC | 07:55 | |
| *** dtantsur|afk is now known as dtantsur | 07:55 | |
| dtantsur | Morning Ironic | 07:55 |
| *** ndipanov has joined #openstack-ironic | 07:56 | |
| *** jerryz has quit IRC | 07:58 | |
| Haomeng | dtantsur: morning:) | 07:59 |
| naohirot | dtantsur: good morning :-) | 07:59 |
| dtantsur | Haomeng, naohirot o/ | 08:01 |
| *** jerryz has joined #openstack-ironic | 08:07 | |
| *** dividehex has quit IRC | 08:09 | |
| *** dividehex has joined #openstack-ironic | 08:10 | |
| *** jistr has joined #openstack-ironic | 08:17 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: In-band hardware properites introspection via ironic-discoverd https://review.openstack.org/135605 | 08:23 |
| *** jerryz_ has joined #openstack-ironic | 08:24 | |
| *** smoriya has quit IRC | 08:30 | |
| *** jcoufal has joined #openstack-ironic | 08:32 | |
| *** MattMan has quit IRC | 08:35 | |
| *** MattMan has joined #openstack-ironic | 08:36 | |
| *** erwan_taf has joined #openstack-ironic | 08:52 | |
| openstackgerrit | Merged stackforge/ironic-discoverd: Refactoring: split discoverd.py into 2 modules https://review.openstack.org/140102 | 09:05 |
| openstackgerrit | Merged stackforge/ironic-discoverd: Refactoring: cap complexity at 15 and coverage at 90 https://review.openstack.org/140136 | 09:05 |
| *** athomas has joined #openstack-ironic | 09:13 | |
| *** derekh has joined #openstack-ironic | 09:17 | |
| *** lsmola has quit IRC | 09:18 | |
| openstackgerrit | Merged stackforge/ironic-discoverd: Refactoring: split test.py and rewrite tests for process https://review.openstack.org/140339 | 09:22 |
| *** lsmola has joined #openstack-ironic | 09:22 | |
| openstackgerrit | Arata Notsu proposed openstack/ironic: Correct vmware ssh power manager https://review.openstack.org/139810 | 09:33 |
| *** kes has quit IRC | 09:41 | |
| pensu | Hey all, what does dhcp_provider=none in ironic.conf does? I mean, if I am not giving any DHCP, how am I supposed to boot a node? | 09:49 |
| openstackgerrit | malini-k-bhandaru proposed openstack/ironic-specs: Bare Metal Trust Using Intel TXT https://review.openstack.org/133902 | 09:51 |
| *** jerryz__ has joined #openstack-ironic | 09:54 | |
| *** jerryz__ has quit IRC | 09:54 | |
| *** jerryz_ has quit IRC | 09:54 | |
| *** subscope has joined #openstack-ironic | 09:56 | |
| *** pelix has joined #openstack-ironic | 10:07 | |
| *** Masahiro has quit IRC | 10:10 | |
| *** Masahiro has joined #openstack-ironic | 10:11 | |
| *** yuanying_ has joined #openstack-ironic | 10:15 | |
| Haomeng | pensu: the dhcp port is controled by neutron | 10:17 |
| *** jcoufal has quit IRC | 10:18 | |
| Haomeng | pensu: so depends on the net created from neutron | 10:19 |
| *** yuanying has quit IRC | 10:19 | |
| *** erwan_taf_ has joined #openstack-ironic | 10:20 | |
| pensu | Haomeng: Okay, what I am confused about it, if I want to use an external DHCP server, should I put dhcp_provider=none? Coz as I understand if it doesn't find anything, it's going to fall back to neutron as default... | 10:20 |
| Haomeng | pensu: yes, but I think we support neutron only | 10:23 |
| Haomeng | pensu: if put dhcp_provider=none, then mean will not call neutron api | 10:23 |
| rameshg87 | dtantsur, hi | 10:23 |
| dtantsur | rameshg87, o/ | 10:23 |
| rameshg87 | dtantsur, regarding comment on https://review.openstack.org/#/c/135899/6/specs/backlog/raid-interface.rst - "Maybe we need at least vendor?" | 10:23 |
| rameshg87 | dtantsur, doesn't the disk type serve the use-case | 10:24 |
| dtantsur | rameshg87, how? | 10:24 |
| rameshg87 | dtantsur, i don't know which disk is actually cheaper | 10:24 |
| rameshg87 | dtantsur, but typeA disk might be cheaper than typeB disk | 10:24 |
| rameshg87 | dtantsur, and operator might want to use those disks for a particular raid volume | 10:25 |
| rameshg87 | dtantsur, will the operator be so choosy in providing disks from vendorX specifically ? | 10:25 |
| dtantsur | rameshg87, I don't think type is enough to distinguish between e.g. hi-end and mid-end disks | 10:25 |
| dtantsur | that's the use case I was told about, I don't know the details myself :) | 10:26 |
| pensu | Haomeng: Then how would this work: https://review.openstack.org/#/c/112351/ | 10:26 |
| rameshg87 | dtantsur, i am just afraid the list keeps growing :( | 10:26 |
| pensu | Haomeng: I mean, either I can put neutron as DHCP provider or "none".....and as far as I can see only none can help me to use external DHCP... | 10:26 |
| dtantsur | yeah... I'm not sure how to solve it | 10:27 |
| pensu | Haomeng: would I still need neutron? | 10:27 |
| rameshg87 | dtantsur, ultimately we might keep providing enough information until we feel providing physical disks is better :) | 10:27 |
| Haomeng | pensu: the patch is for dhcp plugin framework | 10:28 |
| Haomeng | pensu: by default, we use neutron as dhcp provider | 10:28 |
| Haomeng | pensu: if we have more dhcp provider code, we can configure it as another provider | 10:29 |
| Haomeng | pensu: or set to none to use external dhcp which is out-of openstack I think | 10:29 |
| rameshg87 | dtantsur, i don't know - but adding vendor name there seems too much like "pets" to me | 10:30 |
| rameshg87 | dtantsur, i personally wouldn't like to have it there | 10:31 |
| dtantsur | rameshg87, then how to solve this use case? | 10:31 |
| *** naohirot has quit IRC | 10:31 | |
| pensu | Haomeng: Hmmm.....that's what I am trying to figure out, how do I use an out-of openstack DHCP.....let me give it shot....thanks.....:) | 10:31 |
| *** athomas has quit IRC | 10:32 | |
| *** Masahiro has quit IRC | 10:33 | |
| rameshg87 | dtantsur, i would like to know more about this use-case: hi-end and mid-end, what does it mean | 10:33 |
| dtantsur | I can find out and clarify (or let Imre do it :) | 10:33 |
| *** humble__ has joined #openstack-ironic | 10:34 | |
| rameshg87 | dtantsur, yeah some more information from you or Imre would be helpful | 10:34 |
| Haomeng | pensu: run dnsmasq to serv the dhcp request and provide pxe file server and bootfile-name | 10:34 |
| Haomeng | pensu: let me find some reference | 10:35 |
| Haomeng | pensu: https://www.debian-administration.org/article/478/Setting_up_a_server_for_PXE_network_booting | 10:36 |
| Haomeng | pensu: hope this can help you | 10:36 |
| *** wanghua has quit IRC | 10:36 | |
| Haomeng | pensu: and this one - http://blogging.dragon.org.uk/howto-setup-a-pxe-server-with-dnsmasq/ | 10:37 |
| *** athomas has joined #openstack-ironic | 10:39 | |
| pensu | Haomeng: hey, thanks a lot for that! The thing is in my scenario, I am trying to get past the bootfile (i.e pxelinux.0) as I am already running a bootloader, I want kernel and ramdisk to be sent directly.....so, I am planning to use external DHCP to do that.....not sure if external DHCP is the right way to do that! | 10:46 |
| Haomeng | pensu: the kernel and ramdisk is defined in pxeconfig folders | 10:48 |
| *** ramineni has quit IRC | 10:48 | |
| Haomeng | pensu: such as /tftpboot/pxelinux.cfg/default | 10:48 |
| pensu | Haomeng: yeah, that I was able figureo out, I just don't know how to go around pxelinux.0....the DHCP sends that first....and if I leave pxe_bootfile_name blank, it throws the error.... | 10:49 |
| Haomeng | pensu: http://www.syslinux.org/wiki/index.php/PXELINUX | 10:49 |
| Haomeng | pensu: pxelinux.0 is the bootloader which is defined in dhcp potions | 10:50 |
| Haomeng | pensu: s/potions/options | 10:50 |
| Haomeng | pensu: the dhcp options can be passed by feedfile - with the dnsmasq command line argument --dhcp-optsfile | 10:52 |
| pensu | Haomeng: yeah, I tried to remove that option from pxe_utils.py, and ironic.conf.....but it gave me error that the filed is empty... | 10:53 |
| pensu | Haomeng: oh, I haven't tried that.....will give it a shot.... | 10:54 |
| pensu | Haomeng: so, I just leave that filed blank? Or do I have to provide something else instead of pxelinux.0? | 10:54 |
| Haomeng | pensu: good luck, suggest to find a guide which try to configure the dnsmasq for pxe by manually | 10:54 |
| *** chenglch has quit IRC | 10:59 | |
| pensu | Haomeng: Thanks...:) | 11:00 |
| Haomeng | pensu: welcome | 11:00 |
| Haomeng | pensu: in face, it is the same result with dnsmasq process which is controled by neutron dhcp port, and with our manualy steps | 11:01 |
| Haomeng | pensu: to run dnsmasq with enough arguments to support pxe | 11:01 |
| sambetts | morning | 11:03 |
| *** humble__ has quit IRC | 11:06 | |
| *** humble__ has joined #openstack-ironic | 11:06 | |
| dtantsur | sambetts, morning | 11:06 |
| sambetts | morning dtantsur | 11:08 |
| *** rameshg87 has quit IRC | 11:12 | |
| *** pradipta is now known as pradipta_away | 11:22 | |
| *** lazy_prince is now known as killer_prince | 11:30 | |
| *** andreykurilin_ has joined #openstack-ironic | 11:34 | |
| *** Masahiro has joined #openstack-ironic | 11:34 | |
| *** alexpilotti has joined #openstack-ironic | 11:35 | |
| *** Masahiro has quit IRC | 11:38 | |
| *** jcoufal has joined #openstack-ironic | 11:46 | |
| openstackgerrit | Tan Lin proposed openstack/ironic-specs: Bare Metal Trust Using Intel TXT https://review.openstack.org/133902 | 11:47 |
| *** ifarkas has joined #openstack-ironic | 11:48 | |
| *** naohirot has joined #openstack-ironic | 11:54 | |
| *** pensu has quit IRC | 11:59 | |
| *** achanda has joined #openstack-ironic | 11:59 | |
| *** bradjones has joined #openstack-ironic | 12:03 | |
| *** david-lyle_afk is now known as david-lyle | 12:09 | |
| openstackgerrit | Dmitry Tantsur proposed stackforge/ironic-discoverd: Require manual power on if ipmi_setup_credentials is set in Node.extra https://review.openstack.org/140656 | 12:11 |
| *** ifarkas has quit IRC | 12:14 | |
| *** david-lyle is now known as david-lyle_afk | 12:17 | |
| *** david-lyle_afk is now known as david-lyle | 12:17 | |
| *** ifarkas has joined #openstack-ironic | 12:19 | |
| *** jistr has quit IRC | 12:20 | |
| *** david-lyle is now known as david-lyle_afk | 12:23 | |
| *** david-lyle_afk is now known as david-lyle | 12:23 | |
| *** romcheg1 has joined #openstack-ironic | 12:28 | |
| *** david-lyle is now known as david-lyle_afk | 12:31 | |
| *** david-lyle_afk is now known as david-lyle | 12:31 | |
| *** ryanpetrello has joined #openstack-ironic | 12:36 | |
| *** jistr has joined #openstack-ironic | 12:40 | |
| *** Masahiro has joined #openstack-ironic | 12:46 | |
| *** athomas has quit IRC | 12:50 | |
| *** Masahiro has quit IRC | 12:51 | |
| *** dprince has joined #openstack-ironic | 12:54 | |
| *** athomas has joined #openstack-ironic | 12:59 | |
| *** david-lyle is now known as david-lyle_afk | 13:00 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: Add AMT-PXE-Driver to deploy cloud on PC https://review.openstack.org/135184 | 13:06 |
| *** achanda has quit IRC | 13:17 | |
| *** andreykurilin_ has quit IRC | 13:27 | |
| *** masco has quit IRC | 13:27 | |
| Shrews | Good morning, Pixie Boots followers | 13:35 |
| *** Marga_ has joined #openstack-ironic | 13:42 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: Add AMT-PXE-Driver to deploy cloud on PC https://review.openstack.org/135184 | 13:43 |
| *** jerryz has quit IRC | 13:45 | |
| *** ndipanov has quit IRC | 13:50 | |
| dtantsur | Shrews, o/ :) | 13:53 |
| naohirot | Shrews: good morning | 13:55 |
| *** ndipanov has joined #openstack-ironic | 13:55 | |
| Shrews | hi dtantsur, naohirot :) | 13:55 |
| naohirot | Shrews: I have a question about python-ironicclient tox unit test | 13:56 |
| naohirot | Shrews: In my environment, ironicclient.tests.test_shell.ShellTestNoMox.test_node_list never pass. | 13:57 |
| naohirot | Shrews: I'm asking you this question because you kindly review my patch "py26" matter. | 13:58 |
| naohirot | s/review/reviewed/ | 13:58 |
| Shrews | uh, hrm. what's the output? | 13:59 |
| *** jcoufal_ has joined #openstack-ironic | 13:59 | |
| naohirot | just moment | 13:59 |
| *** yarker has quit IRC | 14:00 | |
| naohirot | Shrews: ConnectionRefused: Unable to establish connection to http://keystone.example.com:5000/v2.0/tokens | 14:01 |
| *** jcoufal has quit IRC | 14:01 | |
| naohirot | Shrews: I'm wondering if I need to start keystone server? | 14:02 |
| Shrews | naohirot: no. it shouldn't be trying to actually connect | 14:02 |
| naohirot | Shrews: but this is unit test. I believe that I don't have to. | 14:02 |
| *** yarker has joined #openstack-ironic | 14:03 | |
| Shrews | naohirot: i *think* httpretty should be simulating the connection. i wonder if that's installed | 14:04 |
| naohirot | Shrews: do you know why this unit test is named "ShellTestNoMox"? | 14:04 |
| Shrews | naohirot: no, i do not | 14:04 |
| naohirot | Shrews: I'm wondering if "NoMox" means "NoMock"? | 14:05 |
| Shrews | naohirot: not sure what could be happening. try rebuilding the tox env: tox -epy27 -r | 14:06 |
| soren | Well, it's not using mox. | 14:08 |
| soren | Maybe there used to be a set of tests that used mox and a set that didn't. | 14:08 |
| *** subscope has quit IRC | 14:08 | |
| naohirot | Shrews: Actually this problem is persisting in my environment since I submitted "py26" tox patch, and I recreated my environment several times from clean installation. | 14:09 |
| naohirot | soren: Hi | 14:09 |
| soren | https://github.com/openstack/python-heatclient/blob/master/heatclient/tests/test_shell.py#L312-314 | 14:10 |
| soren | Has a bit of an answer. | 14:11 |
| soren | (And clearly there's some cut/paste going on here) | 14:11 |
| naohirot | Shrews: is this tox unit test always executed at the Jenkins CI test? | 14:11 |
| *** rushiagr is now known as rushiagr_away | 14:11 | |
| * naohirot looking at the URL | 14:12 | |
| openstackgerrit | Dmitry Tantsur proposed stackforge/ironic-discoverd: Require manual power on if ipmi_setup_credentials is set in Node.extra https://review.openstack.org/140656 | 14:13 |
| Shrews | naohirot: http://logs.openstack.org/53/134253/2/check/gate-python-ironicclient-python27/573d7bc/console.html.gz#_2014-12-04_11_55_02_688 | 14:13 |
| Shrews | naohirot: something is broken in your local environment. that's difficult for me to debug :) | 14:14 |
| naohirot | Shrews: Okay, This case works at the gate, so it's my environment problem. | 14:15 |
| naohirot | soren: thanks, I found the comment. | 14:16 |
| naohirot | Shrews: soren: thanks, I'll check my environment thoroughly again. | 14:18 |
| *** subscope has joined #openstack-ironic | 14:21 | |
| *** mjturek has joined #openstack-ironic | 14:23 | |
| *** rloo has joined #openstack-ironic | 14:27 | |
| NobodyCam | good morning Ironic | 14:30 |
| * naohirot I see Mox is one of mock library, https://code.google.com/p/pymox/ | 14:30 | |
| dtantsur | NobodyCam, morning | 14:30 |
| NobodyCam | mrning dtantsur | 14:30 |
| NobodyCam | :) morning naohirot | 14:30 |
| *** igordcard has joined #openstack-ironic | 14:31 | |
| naohirot | NobodyCam: good morning :) | 14:31 |
| *** ryanpetrello_ has joined #openstack-ironic | 14:31 | |
| NobodyCam | :) | 14:31 |
| *** ryanpetrello has quit IRC | 14:31 | |
| *** mjturek has quit IRC | 14:31 | |
| *** ryanpetrello_ is now known as ryanpetrello | 14:31 | |
| *** mjturek has joined #openstack-ironic | 14:31 | |
| naohirot | NobodyCam: we overlapped in morning and in evening :) | 14:32 |
| NobodyCam | hehehe yes we did :) | 14:32 |
| naohirot | NobodyCam: yeah, so it's time to go to bed :) | 14:33 |
| NobodyCam | naohirot: I'll bet.. :) | 14:33 |
| NobodyCam | have a good night | 14:33 |
| naohirot | NobodyCam: dtantsur: Shrews: sore: good night, see you! | 14:34 |
| dtantsur | naohirot, g'night | 14:34 |
| naohirot | bye :) | 14:34 |
| *** naohirot has quit IRC | 14:34 | |
| *** Masahiro has joined #openstack-ironic | 14:35 | |
| NobodyCam | brb | 14:39 |
| *** Masahiro has quit IRC | 14:39 | |
| dtantsur | folks, is it desired behavior that Ironic restarts a node that shut down itself? Oo | 14:40 |
| dtantsur | it was VERY surprising | 14:44 |
| Shrews | dtantsur: probably the power sync thread, no? | 14:45 |
| dtantsur | yeah, it is | 14:45 |
| dtantsur | but I'm not sure if it's what we want. do we expect the users to never shutdown the machine? | 14:46 |
| Shrews | dtantsur: i'd expect they would shut it down via the API, not with a physical button press | 14:46 |
| dtantsur | Shrews, that sounds quite limiting to me... | 14:47 |
| dtantsur | and I of course mean proper shutdown, not physical button | 14:47 |
| Shrews | well, proper shutdown is via the API, no? | 14:47 |
| dtantsur | Shrews, from which point of view? from pov of OS it's sudo halt | 14:47 |
| dtantsur | Shrews, that causes problems with the ramdisks, like discoverd one | 14:48 |
| *** rushiagr_away is now known as rushiagr | 14:48 | |
| dtantsur | because it should shut down once it's finished completely | 14:48 |
| Shrews | dtantsur: ah, from that POV, I'm not sure what we'd expect then. good question | 14:48 |
| dtantsur | and now I have to create one more loop from ramdisk to discoverd just to do `sudo halt` :) | 14:49 |
| dtantsur | I would also expect some users to try halting their machine via CLI as well :) | 14:49 |
| dtantsur | devananda, once you're online, could you comment on ^^^ ? | 14:50 |
| *** ndipanov has quit IRC | 14:52 | |
| jroll | dtantsur: reason number two that we run without force_power_state or whatever | 14:53 |
| jroll | dtantsur: that way ironic will leave it alone, and then nova eventually notices and puts it in a STOPPED state | 14:54 |
| dtantsur | jroll, sorry, what do you mean? | 14:54 |
| *** romcheg1 has quit IRC | 14:54 | |
| dtantsur | ahm | 14:54 |
| dtantsur | oh damn, I'd expect behavior like force_power_state_during_sync=False | 14:55 |
| dtantsur | and the default is True... | 14:55 |
| jroll | right | 14:55 |
| jroll | I tried to make that default to false once | 14:55 |
| jroll | got shot down | 14:56 |
| dtantsur | hmm... changing option downstream and making it a requirement for discoverd sounds the easiest way to me | 14:56 |
| dtantsur | jroll, what was the reason? | 14:57 |
| jroll | dtantsur: I don't remember, trying to find the bug I filed | 14:57 |
| jroll | dtantsur: https://review.openstack.org/#/c/97361/ | 14:58 |
| *** romcheg1 has joined #openstack-ironic | 14:58 | |
| jroll | see lifeless' comment | 14:58 |
| *** Marga_ has quit IRC | 15:00 | |
| dtantsur | oh dear | 15:00 |
| dtantsur | that's where cloud-only case contradicts others... | 15:00 |
| jroll | yeah, I'm sure there was irc discussion here too | 15:02 |
| *** ndipanov has joined #openstack-ironic | 15:05 | |
| jroll | dtantsur: I think this came from us wanting long-running ramdisks, and so that's why people were averse to it | 15:05 |
| *** Marga_ has joined #openstack-ironic | 15:06 | |
| dtantsur | oh hmm... | 15:06 |
| dtantsur | jroll, thanks for explanation. All this complicates my life seriously... | 15:06 |
| jroll | dtantsur: just... set that option to false? | 15:06 |
| *** achanda has joined #openstack-ironic | 15:07 | |
| dtantsur | jroll, maybe. not sure if we can do it e.g. instack-wide, need to talk to guys... | 15:07 |
| jroll | mmm. | 15:08 |
| dtantsur | unfortunately it won't affect only discoverd :) | 15:08 |
| *** romcheg1 has quit IRC | 15:14 | |
| *** romcheg1 has joined #openstack-ironic | 15:16 | |
| NobodyCam | morning jroll | 15:17 |
| jroll | morning NobodyCam :) | 15:17 |
| NobodyCam | :) | 15:18 |
| *** Marga_ has quit IRC | 15:21 | |
| *** r-daneel has joined #openstack-ironic | 15:24 | |
| *** zz_jgrimm is now known as jgrimm | 15:29 | |
| *** jrist has quit IRC | 15:36 | |
| NobodyCam | brb | 15:39 |
| *** Jatin360 has joined #openstack-ironic | 15:43 | |
| *** achanda has quit IRC | 15:44 | |
| jroll | any mirantis folks around? is there a way to make fuel work with already provisioned servers? | 15:44 |
| jroll | like if I wanted to skip the pxe-boot-and-provision step | 15:44 |
| jroll | romcheg: ^ | 15:44 |
| romcheg | jroll: one sec | 15:44 |
| NobodyCam | jroll: thats a odd use case | 15:44 |
| romcheg | I will spawn someone, I'm still not that familiar with Fuel | 15:45 |
| jroll | NobodyCam: more curious than anything | 15:45 |
| NobodyCam | :) | 15:45 |
| NobodyCam | I have been asked just that question b4 .. btw | 15:45 |
| *** holser has joined #openstack-ironic | 15:46 | |
| holser | Hi guys | 15:46 |
| * romcheg spawnes holser | 15:46 | |
| jroll | hiya holser | 15:46 |
| jroll | is there a way to make fuel work with already provisioned servers? | 15:46 |
| jroll | like if I wanted to skip the pxe-boot-and-provision step | 15:46 |
| NobodyCam | morning romcheg and holser | 15:46 |
| romcheg | Morning NobodyCam, and everyone else! | 15:47 |
| holser | actually, we have tiny software in provision image | 15:47 |
| * holser that notifies when ready, when we know that provisioned server is ready, we can use it for deployments. | 15:48 | |
| jroll | holser: what I'm thinking is, I have a bunch of servers running say, ubuntu. then I want to run mirantis cloud on them. without pxe booting | 15:48 |
| jroll | that tiny software is fuel-agent, right? | 15:48 |
| holser | jroll, out of the box - no. | 15:49 |
| holser | yes, fuel-agent | 15:49 |
| romcheg | it's nailgun-agent | 15:49 |
| holser | nailgun-agent* | 15:49 |
| jroll | could I just run that on my machines and it would provision the rest of the openstack software? | 15:49 |
| jroll | ah, ok | 15:49 |
| romcheg | fuel-agent is a completely different story, related to image-based provisioning | 15:49 |
| jroll | right | 15:49 |
| holser | if you have agent and communication layer configured (mcollective) then all nodes will appear so you’ll be able to create environment and deploy it | 15:50 |
| holser | but that functionality is not out of the box | 15:51 |
| jroll | neat | 15:51 |
| holser | well, we have image-based provisioning you may create image based on it | 15:51 |
| *** romcheg1 has quit IRC | 15:51 | |
| *** jrist has joined #openstack-ironic | 15:53 | |
| NobodyCam | dtantsur: question: I'm looking at 135605 (discoverd spec) | 15:54 |
| dtantsur | 1 moment, I'm on call | 15:55 |
| jroll | cool, thanks for the info holser :) | 15:55 |
| *** achanda has joined #openstack-ironic | 15:55 | |
| holser | jroll, you are welcome | 15:55 |
| *** romcheg1 has joined #openstack-ironic | 15:58 | |
| *** zsmithnyc has joined #openstack-ironic | 15:58 | |
| romcheg | jroll: if you have more questions you can always jump in #fuel-dev as well | 15:58 |
| jroll | romcheg: right on, thanks | 15:59 |
| *** achanda has quit IRC | 16:02 | |
| dtantsur | NobodyCam, I'm back | 16:03 |
| NobodyCam | I'll add as a comment on the review. but wanted to ask how you came up with port 5050 | 16:04 |
| *** asr9k has joined #openstack-ironic | 16:04 | |
| dtantsur | NobodyCam, out of random. does it conflicts with something? | 16:04 |
| NobodyCam | https://isc.sans.edu/port.html?port=5050 | 16:04 |
| jroll | yahoo messenger, msn messenger, unity game engine | 16:05 |
| jroll | I don't think there will be conflicts here ;) | 16:05 |
| NobodyCam | :) I run my chat clients in the cloud | 16:06 |
| dtantsur | NobodyCam, port 5000 is also occupied by some malware, I don't think keystone cares :D | 16:06 |
| NobodyCam | :-p | 16:06 |
| jroll | lol | 16:06 |
| *** rameshg87 has joined #openstack-ironic | 16:06 | |
| *** Jatin360 has quit IRC | 16:06 | |
| *** anderbubble has joined #openstack-ironic | 16:07 | |
| rameshg87 | NobodyCam, hi | 16:07 |
| rameshg87 | NobodyCam, just saw comments on https://review.openstack.org/#/c/100951/51/specs/kilo/ironic-node-properties-discovery.rst | 16:07 |
| NobodyCam | hi rameshg87 | 16:07 |
| * NobodyCam looks | 16:07 | |
| rameshg87 | NobodyCam, it should have been MANAGED right instead of ENROLL | 16:07 |
| *** theunraveler has joined #openstack-ironic | 16:07 | |
| NobodyCam | oh you are correct | 16:07 |
| rameshg87 | NobodyCam, as per https://github.com/openstack/ironic-specs/blob/master/specs/kilo/new-ironic-state-machine.rst | 16:07 |
| rameshg87 | NobodyCam, okay | 16:07 |
| *** cuihaozhi has quit IRC | 16:08 | |
| rameshg87 | NobodyCam, Nisha is on vacation | 16:08 |
| NobodyCam | :( yep thats what I get for reviewing before coffee | 16:08 |
| rameshg87 | NobodyCam, since all are nits, i will just address them quickly | 16:08 |
| dtantsur | NobodyCam, anyway, you can change the port, if you run MSN on your undercloud machine :) | 16:08 |
| NobodyCam | want me to put up a comment to that effect rameshg87 | 16:08 |
| rameshg87 | NobodyCam, i will do, np | 16:08 |
| NobodyCam | dtantsur: :) I'm good with that option :) | 16:09 |
| NobodyCam | rameshg87: Thank you :) | 16:09 |
| *** dtantsur is now known as dtantsur|brb | 16:10 | |
| *** jistr has quit IRC | 16:10 | |
| *** dlaube has joined #openstack-ironic | 16:11 | |
| *** theunraveler has quit IRC | 16:12 | |
| *** achanda_ has joined #openstack-ironic | 16:13 | |
| *** dlpartain has joined #openstack-ironic | 16:13 | |
| *** masco has joined #openstack-ironic | 16:16 | |
| openstackgerrit | Ramakrishnan G proposed openstack/ironic-specs: Discover node properties using new CLI node-inspect https://review.openstack.org/100951 | 16:16 |
| rameshg87 | NobodyCam, dtantsur|brb, ^^ | 16:17 |
| NobodyCam | Thank you rameshg87 I will take a look in just a few minutes | 16:17 |
| rameshg87 | NobodyCam, thanks | 16:17 |
| *** jb11211 has joined #openstack-ironic | 16:19 | |
| *** Marga_ has joined #openstack-ironic | 16:20 | |
| *** erwan_taf has quit IRC | 16:21 | |
| *** erwan_taf_ has quit IRC | 16:22 | |
| *** Masahiro has joined #openstack-ironic | 16:24 | |
| *** ryanpetrello has quit IRC | 16:27 | |
| *** ryanpetrello has joined #openstack-ironic | 16:28 | |
| *** Masahiro has quit IRC | 16:28 | |
| *** baker_ has joined #openstack-ironic | 16:29 | |
| *** mikedillion has joined #openstack-ironic | 16:29 | |
| *** romcheg1 has quit IRC | 16:31 | |
| *** baker__ has joined #openstack-ironic | 16:33 | |
| *** baker_ has quit IRC | 16:33 | |
| *** subscope has quit IRC | 16:35 | |
| NobodyCam | rameshg87: +2'd | 16:35 |
| *** Marga_ has quit IRC | 16:37 | |
| *** Marga_ has joined #openstack-ironic | 16:37 | |
| dlaube | g'morning | 16:40 |
| NobodyCam | morning dlaube :) | 16:40 |
| rameshg87 | NobodyCam, thanks :) | 16:42 |
| jroll | heya dlaube :) | 16:42 |
| *** athomas has quit IRC | 16:42 | |
| *** mikedillion has quit IRC | 16:44 | |
| *** shakamunyi has quit IRC | 16:45 | |
| *** victor_lowther has quit IRC | 16:45 | |
| *** romcheg1 has joined #openstack-ironic | 16:46 | |
| dlaube | jroll: you ready for this big storm heading our way? | 16:46 |
| *** romcheg1 has quit IRC | 16:47 | |
| jroll | dlaube: I grew up in michigan where winter storms were 2 feet of snow, this is nothing :) | 16:47 |
| dlaube | nice! I'm originally from upstate NY where we have seen the same | 16:48 |
| zsmithnyc | dlaube: isn't jake in Michigan too? | 16:48 |
| dlaube | I only worry about flooding out here | 16:48 |
| jroll | haha | 16:49 |
| dlaube | zsmithnyc: minnesota I thought | 16:49 |
| dlaube | i could be wrong | 16:49 |
| jroll | I'm in the deep east bay, I don't think I'll have many problems | 16:49 |
| jroll | flooding wise | 16:49 |
| dlaube | jroll: good deal, east bay will probably be high and dry | 16:50 |
| jroll | not quite dry :P | 16:50 |
| dlaube | heh | 16:50 |
| *** subscope has joined #openstack-ironic | 16:51 | |
| *** rameshg87 has quit IRC | 16:51 | |
| NobodyCam | brb ... make'n mo coffee | 16:51 |
| *** dlpartain has quit IRC | 16:52 | |
| *** romcheg1 has joined #openstack-ironic | 16:57 | |
| *** achanda_ has quit IRC | 16:58 | |
| *** igordcard has quit IRC | 17:00 | |
| *** achanda has joined #openstack-ironic | 17:02 | |
| *** david-lyle_afk is now known as david-lyle | 17:06 | |
| *** baker__ has quit IRC | 17:08 | |
| zsmithnyc | jroll: heard you've been schooling mr dlaube a bit. thanks for the help ;-) | 17:08 |
| jroll | ha! | 17:09 |
| jroll | not a problem :) | 17:09 |
| *** Marga_ has quit IRC | 17:12 | |
| *** k4n0 has quit IRC | 17:12 | |
| devananda | morning, all | 17:21 |
| NobodyCam | good morning devananda | 17:21 |
| Shrews | now that's odd. why would ironic-dbsync not create tables at the most recent revision? hrm, a mystery.... | 17:22 |
| NobodyCam | Shrews: because it doesn't like you? | 17:23 |
| Shrews | NobodyCam: that's likely true | 17:23 |
| NobodyCam | :-p | 17:23 |
| Shrews | perhaps it's because i haven't had lunch yet. i should remedy that right now | 17:23 |
| Shrews | biab | 17:24 |
| *** dtantsur|brb is now known as dtantsur | 17:27 | |
| *** marcoemorais has joined #openstack-ironic | 17:28 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/140779 | 17:30 |
| *** victor_lowther has joined #openstack-ironic | 17:41 | |
| NobodyCam | gah why cann't I compaire patch versions to one another.. I could last night... | 17:43 |
| *** Marga_ has joined #openstack-ironic | 17:43 | |
| *** rloo has quit IRC | 17:44 | |
| NobodyCam | oh and just a side comment: Thanks to everyone for all the work and effort on the new state machine !!!! | 17:44 |
| jroll | ^^^^^^^^ | 17:44 |
| *** pensu has joined #openstack-ironic | 17:46 | |
| JayF | did it land? | 17:46 |
| NobodyCam | yes last night | 17:46 |
| JayF | woohoo | 17:47 |
| JayF | ty victor_lowther | 17:47 |
| JayF | among many others | 17:47 |
| *** afaranha has quit IRC | 17:47 | |
| *** Marga_ has quit IRC | 17:49 | |
| *** shakamunyi has joined #openstack-ironic | 17:52 | |
| *** Marga_ has joined #openstack-ironic | 17:52 | |
| *** andreykurilin_ has joined #openstack-ironic | 17:53 | |
| *** masco has quit IRC | 17:54 | |
| *** shakamunyi has quit IRC | 17:55 | |
| victor_lowther | :) | 17:56 |
| NobodyCam | victor_lowther: Thank you :) | 17:57 |
| Shrews | ok, so 'ironic-dbsync upgrade' on an empty schema does NOT create the latest table versions, but 'ironic-dbsync create_schema' does | 18:02 |
| Shrews | so that's fun | 18:02 |
| *** derekh has quit IRC | 18:02 | |
| Shrews | anyone know alembic really well? | 18:02 |
| jroll | Shrews: what version does "upgrade" end up at? | 18:03 |
| NobodyCam | quick poll.. what are folks thoughs on Ironic creating images | 18:03 |
| jroll | noooooo | 18:03 |
| Shrews | jroll: 3cb628139ea4 | 18:03 |
| jroll | why? | 18:03 |
| NobodyCam | https://review.openstack.org/#/c/134022 | 18:03 |
| NobodyCam | I should have said deploy images | 18:04 |
| JayF | I think it's very reasonable to build ISOs out of existing, operator supplied kernel/ramdisk pairings | 18:04 |
| JayF | as the fact the transport mechanism for the image is an ISO vs PXE is a hardware abstraction that Ironic should hide | 18:04 |
| jroll | oh, yeah | 18:05 |
| jroll | I see that as packaging images | 18:05 |
| jroll | not building | 18:05 |
| jroll | Shrews: ... weird | 18:05 |
| NobodyCam | and we kinda do the floppy thing now for the ilo/drac's type stuff | 18:05 |
| Shrews | jroll: and when the conductor tries to start, you end up getting the same error as this person was asking about: http://lists.openstack.org/pipermail/openstack/2014-December/010698.html | 18:06 |
| *** shakamunyi has joined #openstack-ironic | 18:06 | |
| jroll | Shrews: I have a thought | 18:06 |
| jroll | maybe not | 18:07 |
| jroll | Shrews: btw, I think you can also get that error when a driver fails to load | 18:07 |
| devananda | NobodyCam: deploy kernel/ramdisk? or the instance image? | 18:11 |
| *** Masahiro has joined #openstack-ironic | 18:13 | |
| openstackgerrit | Merged openstack/ironic-specs: Set blueprint link for drivers capabilities spec https://review.openstack.org/140267 | 18:13 |
| openstackgerrit | Merged openstack/ironic-specs: fix the filename for hardware-capabilities.rst https://review.openstack.org/140252 | 18:14 |
| JayF | d'oh ^ I didn't see that fix but ty to whoever fixed that | 18:14 |
| JayF | I knew better than to name it the wrong thing | 18:14 |
| *** jlk has left #openstack-ironic | 18:14 | |
| *** Masahiro has quit IRC | 18:17 | |
| *** dlfryar has joined #openstack-ironic | 18:20 | |
| *** anderbubble has quit IRC | 18:21 | |
| *** anderbubble has joined #openstack-ironic | 18:23 | |
| NobodyCam | devananda: deploy image | 18:26 |
| *** MattMan has quit IRC | 18:26 | |
| NobodyCam | oh JayF was just looking over your comments on 134022.. Re: glance and uplading of images see: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ilo/deploy.py#L133-L135 | 18:27 |
| devananda | NobodyCam: what is "deploy image" ? | 18:27 |
| devananda | NobodyCam: I think I'm missing some obvious context ... | 18:28 |
| NobodyCam | the deploy ISO for vmedia based deploys | 18:28 |
| *** ifarkas has quit IRC | 18:28 | |
| NobodyCam | devananda: question was from spec 134022 | 18:28 |
| NobodyCam | :) | 18:28 |
| * NobodyCam hands devananda some more coffee too | 18:28 | |
| dtantsur | devananda, thanks for reviewing the in-band introspection spec, I have some clarification questions inline. | 18:30 |
| *** Marga_ has quit IRC | 18:30 | |
| *** Marga_ has joined #openstack-ironic | 18:31 | |
| *** harlowja_away is now known as harlowja | 18:31 | |
| *** Marga_ has quit IRC | 18:32 | |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-specs: Implement Cleaning States https://review.openstack.org/102685 | 18:32 |
| *** Marga_ has joined #openstack-ironic | 18:32 | |
| devananda | dtantsur: lemme see. I've also just commented on Nisha's spec which yours depends on | 18:35 |
| devananda | 100951 | 18:35 |
| dtantsur | cool, will have a look too | 18:36 |
| *** Marga_ has quit IRC | 18:37 | |
| *** Nisha has joined #openstack-ironic | 18:39 | |
| *** pelix has quit IRC | 18:40 | |
| dtantsur | NobodyCam, rloo, I do think that --dryrun for inspection worth a separate spec... | 18:44 |
| NobodyCam | really? | 18:45 |
| *** Hefeweizen has joined #openstack-ironic | 18:46 | |
| NobodyCam | I was thinking it was just a skip the actual properties update type change | 18:46 |
| dtantsur | NobodyCam, even the question "where to store this dryrun information" looks complicated to me... | 18:46 |
| NobodyCam | you see it as larger then that | 18:46 |
| *** humble__ has quit IRC | 18:46 | |
| dtantsur | suggesting to start using 'extra' for temporary dropping this information? | 18:46 |
| NobodyCam | hummm | 18:47 |
| *** humble__ has joined #openstack-ironic | 18:49 | |
| devananda | dtantsur: replied | 18:49 |
| devananda | I don't like discover --dryrun | 18:49 |
| dtantsur | thanks | 18:49 |
| devananda | we dont have provision --dryrun ... | 18:50 |
| devananda | or set-power-state --dryrun | 18:50 |
| devananda | what makes this interface need that, when others didn't? | 18:50 |
| NobodyCam | other interfaces have a validate. like with power i can validate it will work.. with discover I have no way to "test" what its going to do | 18:51 |
| NobodyCam | but I do see the point your making here | 18:52 |
| devananda | oh, i need to draft a spec for enroll->validate->managed | 18:52 |
| devananda | which is different from the validate() that each driver interface implements -- those don't actually touch the node | 18:53 |
| *** pensu has quit IRC | 18:53 | |
| devananda | they only check whether the driver *believes* it has the information it needs. not whether the information is *correct* | 18:53 |
| NobodyCam | doesn't power do a test poke... Imbw in that thought... | 18:53 |
| devananda | nope | 18:54 |
| devananda | we removed that a while back | 18:54 |
| NobodyCam | ahh | 18:54 |
| NobodyCam | :-p | 18:54 |
| *** rloo has joined #openstack-ironic | 18:54 | |
| * NobodyCam updates brain index | 18:54 | |
| devananda | https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ipmitool.py#L604 | 18:54 |
| *** Marga_ has joined #openstack-ironic | 18:56 | |
| *** Marga_ has quit IRC | 18:56 | |
| NobodyCam | so I do agree that this interface should not be different then the others | 18:56 |
| devananda | https://bugs.launchpad.net/ironic/+bug/1314954 fix actually added that comment | 18:56 |
| devananda | https://bugs.launchpad.net/ironic/+bug/1314961 is Wishlisted | 18:57 |
| devananda | that'll be solved by imlementing enroll->validating->managed | 18:57 |
| *** Marga_ has joined #openstack-ironic | 19:01 | |
| *** marcoemorais has quit IRC | 19:02 | |
| *** marcoemorais has joined #openstack-ironic | 19:03 | |
| *** marcoemorais has quit IRC | 19:03 | |
| *** marcoemorais has joined #openstack-ironic | 19:03 | |
| *** jiangfei has quit IRC | 19:05 | |
| *** jistr has joined #openstack-ironic | 19:05 | |
| *** jiangfei has joined #openstack-ironic | 19:06 | |
| *** andreykurilin_ has quit IRC | 19:09 | |
| *** dlfryar has quit IRC | 19:11 | |
| devananda | NobodyCam: in the fsm.py patch, is there a reason you added **watchers only to a few states? | 19:14 |
| devananda | oh -- never mind. it's on the _states_, not _transitions_. ignore me :) | 19:15 |
| NobodyCam | :-p | 19:16 |
| devananda | NobodyCam: going to fix up a few more things in it, in response to rloo's comments | 19:18 |
| devananda | I'm hoping to have some time today to implement the next step | 19:18 |
| *** pawel-palucki has quit IRC | 19:18 | |
| devananda | are you working on it, by chance? | 19:18 |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic-specs: Add a new driver for Fuel Agent https://review.openstack.org/138115 | 19:23 |
| NobodyCam | the next step or the next review.. /me is not doing any coding for fsm atm | 19:25 |
| devananda | k k | 19:26 |
| *** baker has joined #openstack-ironic | 19:27 | |
| *** jistr has quit IRC | 19:30 | |
| *** erwan_taf has joined #openstack-ironic | 19:31 | |
| *** dlfryar has joined #openstack-ironic | 19:31 | |
| *** erwan_taf_ has joined #openstack-ironic | 19:33 | |
| *** shakamunyi has quit IRC | 19:33 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add fsm.py module https://review.openstack.org/139215 | 19:34 |
| *** shakamunyi has joined #openstack-ironic | 19:34 | |
| NobodyCam | Thank you for the new cli release devananda :) | 19:34 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Begin using the state machine for node deploy/teardown https://review.openstack.org/139216 | 19:35 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Begin using the state machine for node deploy/teardown https://review.openstack.org/139216 | 19:36 |
| openstackgerrit | Dmitry Tantsur proposed stackforge/ironic-discoverd: Support updating IPMI credentials from within ramdisk https://review.openstack.org/140814 | 19:37 |
| *** jiangfei has quit IRC | 19:40 | |
| dtantsur | g'night folks | 19:40 |
| *** dtantsur is now known as dtantsur|afk | 19:40 | |
| NobodyCam | have a good night dtantsur|afk | 19:40 |
| *** marcoemorais has quit IRC | 19:40 | |
| *** jiangfei has joined #openstack-ironic | 19:41 | |
| *** marcoemorais has joined #openstack-ironic | 19:41 | |
| *** marcoemorais has quit IRC | 19:42 | |
| *** marcoemorais has joined #openstack-ironic | 19:43 | |
| *** jcoufal_ has quit IRC | 19:49 | |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic-specs: Add a new driver for Fuel Agent https://review.openstack.org/138115 | 19:56 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Refactor async helper methods in conductor/manager.py https://review.openstack.org/139217 | 19:57 |
| *** mrda-away is now known as mrda | 19:57 | |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-specs: Implement Zapping States https://review.openstack.org/140826 | 19:58 |
| mrda | Morning Ironic! | 19:58 |
| NobodyCam | morning mrda | 19:58 |
| mrda | hey NobodyCam | 19:58 |
| devananda | g'morning, mrda! | 19:59 |
| mrda | hi devananda | 19:59 |
| jroll | \o mrda | 20:00 |
| mrda | hiya jroll | 20:00 |
| *** Masahiro has joined #openstack-ironic | 20:01 | |
| *** dlfryar has quit IRC | 20:02 | |
| NobodyCam | devananda: looking at the add fsm.py patch... you see any reason to add something like: https://github.com/openstack/taskflow/blob/master/taskflow/tests/unit/test_types.py | 20:04 |
| devananda | NobodyCam: something like that, yes | 20:05 |
| devananda | there aren't any unit tests in my patch series for fsm.py, which is less than ideal | 20:05 |
| NobodyCam | :) | 20:06 |
| devananda | so ya, class FSMTest would make sense to port | 20:06 |
| *** Masahiro has quit IRC | 20:06 | |
| devananda | I can do that, unless someone else wants to | 20:06 |
| NobodyCam | I can attempt to do that | 20:07 |
| devananda | k k | 20:07 |
| *** ryanpetrello has quit IRC | 20:11 | |
| *** ryanpetrello_ has joined #openstack-ironic | 20:11 | |
| *** ryanpetrello_ is now known as ryanpetrello | 20:12 | |
| *** dlfryar has joined #openstack-ironic | 20:12 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Do not log configdrive contents at INFO. https://review.openstack.org/140830 | 20:14 |
| JayF | jroll: ^ fix that bug I just found | 20:14 |
| jroll | JayF: s/at INFO.// maybe? | 20:14 |
| JayF | meh? | 20:15 |
| jroll | makes it sound like you're changing the log level | 20:15 |
| jroll | if you don't read the rest | 20:15 |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Do not log configdrive contents ever. https://review.openstack.org/140830 | 20:15 |
| jroll | <3 | 20:15 |
| jroll | I was going to troll you for a bug number | 20:16 |
| jroll | turns out you had one, great work | 20:16 |
| JayF | of course I did | 20:16 |
| JayF | I'd enforce that on you :) | 20:16 |
| *** Nisha has quit IRC | 20:18 | |
| NobodyCam | brb | 20:22 |
| *** jcoufal has joined #openstack-ironic | 20:23 | |
| devananda | aweeks: looking into how to do callbacks from fsm.py right now | 20:23 |
| devananda | aweeks: it's not possible right now, because the callbacks need to happen in the worker pool -- not in the same thread | 20:24 |
| lifeless | dtantsur|afk: jroll: ironic powering off something running ironic managed code is different to that case IMO - no? | 20:28 |
| jroll | lifeless: at the risk of getting into a conversation that I don't want to have right now... different to which case? | 20:30 |
| lifeless | jroll: so the review I commented on was changing the default of force-power-state | 20:31 |
| lifeless | jroll: there are two broad cases AIUI - there's the case where you want Ironic to be like a generic IPMI proxy, e.g. its not managing power unless you say 'please change X', which folk with non-cloud environments may want. | 20:32 |
| lifeless | jroll: and there is the case where Ironic is managing the machines - doing deployments and so forth, so all the concerns I raised in that comment apply. | 20:33 |
| JayF | Why does that use case belong in OpenStack? Caring about pets != cloudy | 20:33 |
| lifeless | JayF: I'd be delighted to remove it entirely | 20:33 |
| lifeless | jroll: discoverd, assuming I got the context right, is about doing hardware interrogation to figure out what is inside the box | 20:34 |
| JayF | lifeless: actually, that use case could fit in a machine in the new MANAGED state | 20:34 |
| lifeless | jroll: and thus thats either part of the Ironic state machine | 20:34 |
| * JayF exits the conversation before the gravitational pull gets too strong | 20:35 | |
| lifeless | jroll: or if its being down by something a layer up, they should put it in maintenance mode | 20:35 |
| jroll | lifeless: what we're talking about is what happens when something running on the machine calls 'halt' | 20:35 |
| jroll | lifeless: for example, a user that doesn't know they should call nova stop instead | 20:35 |
| *** dlfryar has quit IRC | 20:35 | |
| lifeless | jroll: ok. So if the tenant turns it off, it should stay off ? | 20:36 |
| jroll | yes | 20:36 |
| lifeless | jroll: I'd be ok with that; false *on* is what is a big issue for me. | 20:36 |
| jroll | it's the tenant's machine | 20:36 |
| jroll | sure | 20:36 |
| lifeless | jroll: you won't get pushback from me on this | 20:37 |
| jroll | lifeless: I can get behind that, though I tend to think that's a rare case :) | 20:37 |
| lifeless | jroll: the prior proposal's issues were all around false on. | 20:37 |
| jroll | right | 20:37 |
| Shrews | jroll: trying to reproduce that ironic-dbsync problem in a non-dockerized environment... can't do it. more weirdness :( | 20:37 |
| jroll | lifeless: because we get a little bit hacky over here | 20:37 |
| jroll | Shrews: stale pyc files maybe somehow? | 20:37 |
| Shrews | jroll: possible | 20:38 |
| Shrews | likely, even | 20:38 |
| *** dlfryar has joined #openstack-ironic | 20:41 | |
| NobodyCam | brb | 20:43 |
| *** rushiagr is now known as rushiagr_away | 20:43 | |
| Shrews | JayF: lol @ bug 1401244 | 20:44 |
| *** shakamunyi has quit IRC | 20:53 | |
| Shrews | dhellmann: Any idea on when the next release of oslo.messaging will occur? Specifically looking for this fix: https://review.openstack.org/139982 | 20:53 |
| Shrews | i suppose i could just install it from git to silence the warnings :/ | 20:54 |
| *** Haomeng|2 has joined #openstack-ironic | 20:54 | |
| alexm__ | Hi guys, how can I make my ssh keypair be copied over my metal node that I provisioned with Ironic? It seems like cloud-init cannot do this, since there is no metadata service on 169.254 | 20:54 |
| *** Haomeng has quit IRC | 20:54 | |
| devananda | alexm__: either run the nova metadata service, so that that works ... or bake it into the image you are deploying | 20:55 |
| *** shakamunyi has joined #openstack-ironic | 20:56 | |
| jroll | or use configdrive! (disclaimer it doesn't quite work yet) | 20:57 |
| alexm__ | devananda: the metadata service is running on the controller node with Ironic/nova, but the metal node cannot route to it. Must I add some extra properties to the neutron subnet? | 20:57 |
| devananda | alexm__: quite possibly. it depends on your network environment | 20:58 |
| jroll | you almost certainly need to do something on the networking side of things | 20:58 |
| alexm__ | baking the service on the image, or baking the ssh pub keypair for that matter, would do the trick. However, it’s not really friendly to ask clients to do this all the time, what’s the strategy for this? | 20:59 |
| devananda | alexm__: proper networking so that cloud-init can reach the metadata service :) | 20:59 |
| alexm__ | devananda: must be a common issue, is everyone doing this by hand? Thought Ironic would take care of that ;) | 21:00 |
| devananda | alexm__: ironic doesn't proscribe the network environment it is run in | 21:00 |
| devananda | so, yes, everyone has to set up their networks | 21:00 |
| devananda | and people choose different ways to do that | 21:01 |
| alexm__ | alright I’ll see how we can deploy that, thanks! | 21:02 |
| *** erwan_taf_ has quit IRC | 21:02 | |
| *** erwan_taf has quit IRC | 21:02 | |
| JayF | Shrews: strange the things you see when tailing logs during a deployment :) | 21:03 |
| *** achanda has quit IRC | 21:15 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Update dev quick-start for devstack https://review.openstack.org/140124 | 21:15 |
| JayF | I just had a pep8 job fail in an interesting way; a traceback in pep8 itself. https://gist.github.com/jayofdoom/dc8df0c64a8f594fd5ae | 21:26 |
| NobodyCam | brb | 21:28 |
| jroll | I feel like I've seen this before | 21:28 |
| JayF | this was in the gate | 21:32 |
| JayF | for my one line change from earlier | 21:32 |
| *** jcoufal has quit IRC | 21:43 | |
| adam_g | JayF, link to the logs? | 21:46 |
| adam_g | i feel like we saw that in the ironic tree when PYTHONHASHSEED=random, but maybe that was sphinx | 21:47 |
| *** Masahiro has joined #openstack-ironic | 21:50 | |
| *** enterprisedc has quit IRC | 21:52 | |
| *** Masahiro has quit IRC | 21:55 | |
| JayF | adam_g: it's for https://review.openstack.org/140830 | 21:58 |
| JayF | adam_g: ipa | 21:58 |
| *** ryanpetrello has quit IRC | 22:04 | |
| *** dprince has quit IRC | 22:04 | |
| adam_g | JayF, heh its an H803 error | 22:05 |
| adam_g | JayF, see here https://bugs.launchpad.net/horizon/+bug/1229342 | 22:05 |
| adam_g | removing the . from the commit message fixes it locally | 22:05 |
| jroll | lol | 22:05 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-python-agent: Do not log configdrive contents ever https://review.openstack.org/140830 | 22:06 |
| jroll | thanks adam_g | 22:06 |
| *** shakamunyi has quit IRC | 22:08 | |
| adam_g | does the agent ramdisk get its networking from DHCP on boot or does it derive it from somewhere else? the pxe ramdisk is currently getting it from kernel parameters appended by syslinux | 22:09 |
| jroll | it DHCPs afaik | 22:09 |
| *** shakamunyi has joined #openstack-ironic | 22:10 | |
| *** zigo has quit IRC | 22:11 | |
| *** zigo has joined #openstack-ironic | 22:12 | |
| devananda | NobodyCam: I think 'wait call-back', on error, goes to 'deploy failed' | 22:15 |
| *** r-daneel has quit IRC | 22:16 | |
| devananda | there's a path which isn't covered by unit tests ... just found ^ | 22:16 |
| devananda | via conductor.utils.cleanup_after_timeout | 22:16 |
| devananda | should that go to deployfail or error? | 22:17 |
| *** dlfryar has quit IRC | 22:20 | |
| NobodyCam | oh | 22:25 |
| *** andreykurilin_ has joined #openstack-ironic | 22:25 | |
| NobodyCam | I'm thinking deployfail | 22:25 |
| NobodyCam | as we're not sure ware the error is.. conductor,node, network | 22:26 |
| devananda | right | 22:26 |
| *** dlfryar has joined #openstack-ironic | 22:26 | |
| devananda | http://git.openstack.org/cgit/openstack/ironic/tree/ironic/conductor/utils.py#n131 | 22:26 |
| devananda | current code goes to DEPLOYFAIL | 22:27 |
| NobodyCam | yep | 22:27 |
| devananda | i'm probably missing something, but this seems like it should have failed unit tests | 22:28 |
| devananda | https://review.openstack.org/#/c/139216/5/ironic/conductor/utils.py | 22:28 |
| devananda | because there is no 'fail' transition from DEPLOYWAIT in https://review.openstack.org/#/c/139215/5/ironic/common/states.py | 22:30 |
| *** yjiang5_away is now known as yjiang5 | 22:32 | |
| NobodyCam | yep we've got deleteing and deploying | 22:32 |
| NobodyCam | let me add | 22:32 |
| devananda | sec. I have it in the patch I'm working on | 22:33 |
| NobodyCam | ok | 22:33 |
| devananda | I'm experimenting with moving all the callback stuff into task_manager | 22:33 |
| devananda | seems to work well for (event) -> STATING -> STATED -> NEWSTATE | 22:34 |
| devananda | not as well for STATING -> (event) -> STATEFAIL | 22:34 |
| devananda | because if the callback from event hits NoFreeConductorWorker, it doesn't run, but the node is being left in STATEFAIL | 22:35 |
| devananda | without actually transitioning | 22:35 |
| devananda | ** without the driver being called to perform the transition | 22:35 |
| *** aszeszo has quit IRC | 22:39 | |
| *** aszeszo has joined #openstack-ironic | 22:39 | |
| NobodyCam | oh | 22:48 |
| *** baker has quit IRC | 22:48 | |
| NobodyCam | devananda: fyi - http://paste.openstack.org/show/F7pNvjPmbxSlyMMi3ekv | 22:50 |
| NobodyCam | needs clean up... | 22:50 |
| devananda | NobodyCam: needs significant refactoring. that's not threadsafe | 22:53 |
| devananda | using module globals to track test progress is going to break when run in parallel | 22:53 |
| NobodyCam | ack | 22:53 |
| devananda | harlowja: looks like taskflow isn't running testr with --parallel ? | 22:54 |
| harlowja | devananda hmmm, not sure | 22:54 |
| devananda | https://github.com/openstack/taskflow/blob/master/tox.ini | 22:54 |
| NobodyCam | that is the initial stab | 22:55 |
| harlowja | devananda hmmm, could be | 22:55 |
| harlowja | that might make it go quicker, ha | 22:55 |
| harlowja | is anyone running it in parallel? | 22:56 |
| devananda | hmm | 22:56 |
| harlowja | is ironic :-/ | 22:57 |
| devananda | harlowja: my bad. I thought the link NobodyCam sent me was from taskflow | 22:57 |
| harlowja | kk | 22:57 |
| devananda | harlowja: i'm pretty sure we are... but it's not in our tox.ini either, so now I've just opened a can of worms :p | 22:57 |
| harlowja | :-P | 22:57 |
| harlowja | maybe by default parallel? | 22:57 |
| harlowja | nope, not that, ha | 22:59 |
| harlowja | odd | 22:59 |
| harlowja | well thats weird, lol | 22:59 |
| devananda | it is in the gate env, tho | 22:59 |
| harlowja | ya, mysterious | 23:01 |
| NobodyCam | devananda: oh no that was my initial port of the tests | 23:03 |
| harlowja | NobodyCam u need to make a neater example FSM vs the jump up down one, lol | 23:06 |
| NobodyCam | :) | 23:06 |
| openstackgerrit | Merged openstack/ironic-python-agent: Do not log configdrive contents ever https://review.openstack.org/140830 | 23:06 |
| NobodyCam | humm: https://github.com/openstack/ironic/blob/master/tox.ini#L15 | 23:07 |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-python-agent: Add command metrics to IPA API https://review.openstack.org/119981 | 23:07 |
| *** dlfryar has quit IRC | 23:08 | |
| adam_g | jroll, are you familiar with elilo_efi_pxe_config.template at all? you were listed as a co-author | 23:26 |
| NobodyCam | oh the testr option is --no-parallel | 23:26 |
| NobodyCam | til | 23:26 |
| jroll | adam_g: not really, that's a uefi thing iirc | 23:26 |
| jroll | adam_g: I just cleaned up a bunch of stuff on that patch | 23:26 |
| adam_g | jroll, ah, ok | 23:26 |
| adam_g | jroll, do you know the original authors irc nick? | 23:27 |
| jroll | link to the patch? | 23:27 |
| jroll | what are you trying to figure out? | 23:27 |
| *** ChuckC has quit IRC | 23:28 | |
| adam_g | trying to figure out the '%H::on' is doing here, and why it might be needed. ip=%I:{{pxe_options.tftp_server}}:%G:%M:%H::on" | 23:29 |
| *** ryanpetrello has joined #openstack-ironic | 23:29 | |
| adam_g | i'd like to try to remove the use of passing net info to ramdisks via ip=* | 23:29 |
| adam_g | patch at https://review.openstack.org/#/c/114357/ | 23:30 |
| *** anderbubble has quit IRC | 23:30 | |
| *** r-daneel has joined #openstack-ironic | 23:33 | |
| *** r-daneel has quit IRC | 23:38 | |
| *** Masahiro has joined #openstack-ironic | 23:39 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Enable async callbacks from task.process_event() https://review.openstack.org/140868 | 23:40 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert check_deploy_timeout to use process_event https://review.openstack.org/140869 | 23:40 |
| devananda | NobodyCam: ^ | 23:42 |
| devananda | aweeks: example of the callbacks in action, so far as I have yet gotten them to work, is above | 23:42 |
| devananda | harlowja: probably not how you envisioned the on_event stuff working -- but a lot of our transitions happen in a different thread than the one which receives the event ... soooo ... ^ | 23:43 |
| *** Masahiro has quit IRC | 23:43 | |
| harlowja | interesting | 23:44 |
| devananda | https://review.openstack.org/#/c/140868/1/ironic/conductor/task_manager.py | 23:45 |
| harlowja | ya, thats starting to look like the reaction stuff sorta :-P | 23:45 |
| devananda | harlowja: it's tightly coupled to the task_manager context though | 23:46 |
| harlowja | ya, i wonder if that can be de-coupled | 23:46 |
| harlowja | if that matters | 23:46 |
| harlowja | depends on how much u want to invert things | 23:46 |
| devananda | this is mostly getting triggered via __exit__ | 23:47 |
| harlowja | kk, spawn_after gets triggered there? | 23:48 |
| harlowja | ya, looks like it | 23:49 |
| *** r-daneel has joined #openstack-ironic | 23:50 | |
| *** ryanpetrello has quit IRC | 23:50 | |
| harlowja | this though might be ok, its similar in concept to a reaction that i had, in that the reaction u have spawns a callback, sets up an error handler, does some other stuff... | 23:50 |
| harlowja | gets triggered whenever __exit__ happens | 23:50 |
| harlowja | that then causes state machine to do something else, react to something new, and repeat | 23:51 |
| jroll | adam_g: ah, that person's nick is faizan, they are sometimes on irc | 23:51 |
| adam_g | jroll, thanks | 23:52 |
| devananda | harlowja: very close, ya. except that handlers don't return a pointer to the next handler | 23:53 |
| devananda | harlowja: or s/handler/reaction/ | 23:53 |
| harlowja | ya | 23:53 |
| harlowja | which i think is ok, to much inverting/refactoring sometimes to do that | 23:53 |
| devananda | harlowja: oh. another thing. I may want to instantiate the FSM within the API service, just to test the possibility of a state transition (maybe...) | 23:56 |
| harlowja | should be doable ya? | 23:56 |
| devananda | harlowja: for me to use reactions, I need to reference ironic.conductor.manager.$METHOD | 23:56 |
| harlowja | i think cinder wants something similar, actually there's is so that they can dump back the states to horizon | 23:56 |
| devananda | which isn't do-able within the API service | 23:56 |
| harlowja | sure, or u have to refactor it so that its not ironic.conductor.manager.$METHOD but is in some ironic.machine.conductor.reactions.$METHOD or something, which has all the default callbacks, the error handlers, the non node specific stufff (aka, the $METHODs are like functools.partial objects, not fully ready to be activated) | 23:58 |
| devananda | yup. possible | 23:58 |
| harlowja | then u can make it in API, just don't use it (since u won't be able to fill in the partially made functions/methods with real 'node' objects) | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!