| *** absubram has joined #openstack-ironic | 00:14 | |
| *** ociuhandu has joined #openstack-ironic | 00:16 | |
| *** diga has joined #openstack-ironic | 00:33 | |
| *** jcoufal has joined #openstack-ironic | 00:35 | |
| *** gyee has quit IRC | 00:40 | |
| *** jcoufal has quit IRC | 00:48 | |
| *** rloo has quit IRC | 00:56 | |
| *** ociuhandu has quit IRC | 01:02 | |
| *** ricolin has joined #openstack-ironic | 01:03 | |
| *** cp has joined #openstack-ironic | 01:24 | |
| openstackgerrit | Nguyen Van Trung proposed openstack/ironic master: Support firmware upgrade for irmc driver https://review.opendev.org/599992 | 01:40 |
|---|---|---|
| openstackgerrit | Merged openstack/ironic stable/rocky: CI: remove quotation marks from TEMPEST_PLUGINS variable https://review.opendev.org/676517 | 02:11 |
| *** gregoryo has joined #openstack-ironic | 02:12 | |
| *** absubram has quit IRC | 02:12 | |
| *** verma-varsha has joined #openstack-ironic | 02:22 | |
| *** verma-varsha1 has joined #openstack-ironic | 02:40 | |
| *** verma-varsha has quit IRC | 02:40 | |
| *** verma-varsha1 is now known as verma-varsha | 02:40 | |
| *** verma-varsha has quit IRC | 02:55 | |
| openstackgerrit | Merged openstack/ironic stable/stein: RAID creation fails with 'ilo5' RAID interface https://review.opendev.org/674449 | 03:04 |
| *** verma-varsha has joined #openstack-ironic | 03:16 | |
| *** altlogbot_2 has quit IRC | 03:16 | |
| *** altlogbot_1 has joined #openstack-ironic | 03:17 | |
| openstackgerrit | Merged openstack/ironic master: Add Redfish Virtual Media Boot support https://review.opendev.org/638453 | 03:30 |
| openstackgerrit | Nguyen Van Trung proposed openstack/ironic master: Support firmware upgrade for irmc driver https://review.opendev.org/599992 | 03:38 |
| openstackgerrit | Merged openstack/ironic master: Add set_boot_device hook in `redfish` boot interface https://review.opendev.org/672123 | 03:55 |
| openstackgerrit | Merged openstack/ironic master: Add `filename` parameter to Redfish virtual media boot URL https://review.opendev.org/671054 | 03:59 |
| *** gkadam has joined #openstack-ironic | 04:07 | |
| *** gkadam has quit IRC | 04:08 | |
| *** verma-varsha has quit IRC | 04:21 | |
| *** ociuhandu has joined #openstack-ironic | 04:30 | |
| *** ociuhandu has quit IRC | 04:35 | |
| *** rh-jelabarre has quit IRC | 04:54 | |
| *** kaifeng has joined #openstack-ironic | 05:25 | |
| *** mgoddard has quit IRC | 05:54 | |
| *** mgoddard has joined #openstack-ironic | 06:01 | |
| etingof | hurray \o/ | 06:06 |
| *** verma-varsha has joined #openstack-ironic | 06:23 | |
| arne_wiebalck | Good morning, ironic! | 06:26 |
| kaifeng | moring etingof arne_wiebalck o/ | 06:34 |
| arne_wiebalck | gm kaifeng o/ | 06:34 |
| *** belmoreira has joined #openstack-ironic | 06:55 | |
| *** ricolin_ has joined #openstack-ironic | 07:03 | |
| *** ricolin has quit IRC | 07:05 | |
| *** jtomasek has joined #openstack-ironic | 07:16 | |
| openstackgerrit | Varsha Verma proposed openstack/sushy-tools master: Add Drive resource support https://review.opendev.org/675085 | 07:16 |
| *** zhangfei has joined #openstack-ironic | 07:18 | |
| *** jawad_axd has joined #openstack-ironic | 07:26 | |
| *** tssurya has joined #openstack-ironic | 07:32 | |
| *** gregoryo has quit IRC | 07:32 | |
| *** ricolin_ is now known as ricolin | 07:33 | |
| *** verma-varsha has quit IRC | 07:56 | |
| *** lucasagomes has joined #openstack-ironic | 07:57 | |
| *** verma-varsha has joined #openstack-ironic | 08:10 | |
| *** e0ne has joined #openstack-ironic | 08:15 | |
| *** alexmcleod has joined #openstack-ironic | 08:40 | |
| *** derekh has joined #openstack-ironic | 08:40 | |
| *** ricolin has quit IRC | 09:07 | |
| *** jawad_axd has quit IRC | 09:29 | |
| *** verma-varsha has quit IRC | 09:49 | |
| *** hamzy_ has joined #openstack-ironic | 09:51 | |
| *** hamzy has quit IRC | 09:54 | |
| *** belmoreira has quit IRC | 10:03 | |
| openstackgerrit | Varsha Verma proposed openstack/sushy-tools master: Add Drive resource support https://review.opendev.org/675085 | 10:08 |
| *** mgoddard_ has joined #openstack-ironic | 10:14 | |
| *** Lucas_Gray has joined #openstack-ironic | 10:20 | |
| TheJulia | good morning | 10:54 |
| arne_wiebalck | gm TheJulia o/ | 11:00 |
| TheJulia | arne_wiebalck: regading the two patches addressing the same basic issue in IPA. I'm thinking your patch may be better in that the bunch of conditionals kind of makes me uncomfortable when sorting through diks | 11:00 |
| TheJulia | disks | 11:00 |
| * arne_wiebalck needed to stop listening to music and closed the office door to understand to logic | 11:01 | |
| arne_wiebalck | the * | 11:01 |
| TheJulia | hmm... today should be a music day | 11:03 |
| arne_wiebalck | TheJulia: the block device listing is already complicated, the extended if does not make it much worse I think | 11:03 |
| TheJulia | Well... true | 11:04 |
| *** cdent has joined #openstack-ironic | 11:04 | |
| arne_wiebalck | TheJulia: It was actually the place I wanted to add things as well | 11:04 |
| arne_wiebalck | TheJulia: I just couldn't figure out how | 11:04 |
| TheJulia | hmm | 11:04 |
| arne_wiebalck | TheJulia: testing of this takes time | 11:04 |
| arne_wiebalck | TheJulia: with all the cases | 11:05 |
| TheJulia | Indeed, and for that very reason more conditionals just makes me think your solution is a little better | 11:05 |
| cdent | arne_wiebalck: I've come back with another thought about something to tryto speed up the nova-compute that is managing a bunch of ironic nodes: | 11:05 |
| TheJulia | Good morning cdent | 11:05 |
| cdent | allow the nova-compute to talk to the database | 11:05 |
| cdent | morning TheJulia | 11:05 |
| cdent | (instead of the conductor) | 11:05 |
| cdent | rpc overhead is huge | 11:05 |
| arne_wiebalck | TheJulia: how about we wait for Raphael to comment? I just would like to make sure that the cases that were solved already (like a broken RAID-0) remain solved | 11:06 |
| arne_wiebalck | Hey cdent o/ | 11:06 |
| TheJulia | cdent: we don't rpc our api read requests, they come directly out of the DB, but that API interaction overhead is also huge | 11:06 |
| TheJulia | arne_wiebalck: absolutely | 11:07 |
| cdent | TheJulia: I mean in the nova-compute itself | 11:07 |
| cdent | for its db needs | 11:07 |
| cdent | since the reason the nova-compute isn't allowed to access the db is because of vm breakouts, that's not really a concern in the "managing ironic nodes" scenario | 11:08 |
| jroll | cdent: that is a good idea and way too obvious | 11:08 |
| TheJulia | cdent: I got that :) | 11:08 |
| cdent | jroll: comes out of https://review.opendev.org/#/c/676672/ | 11:08 |
| patchbot | patch 676672 - nova - Add a "Caveats" section to the eventlet profiling ... - 1 patch set | 11:08 |
| jroll | nice | 11:08 |
| cdent | jroll: I was doing that profiling for the vmware scenario and (where the vms are off in another cluster) and thought "instead of trying to optimize this code, maybe we shoudn't run it" | 11:09 |
| TheJulia | cdent: This database access idea is growing on me..... | 11:09 |
| *** belmoreira has joined #openstack-ironic | 11:09 | |
| arne_wiebalck | belmoreira ^^^ | 11:09 |
| cdent | where "this code" is "the rpc stuff" | 11:09 |
| cdent | object serialization may be a factor | 11:09 |
| jroll | if _runs_vms_locally(): cmd_comon.block_db_access('nova-compute') | 11:10 |
| cdent | however it is hard to be super sure, because that rpc code is also something the screws up the profiling data | 11:10 |
| TheJulia | There is no pagination overhead.... list returns on the database cursor | 11:10 |
| jroll | TheJulia: I believe cdent is referring to nova-compute accessing the nova db, not the ironic db | 11:11 |
| TheJulia | We would have to split out our object model unless ironic was installed and not used... that is if the object/db overhead is even acceptable | 11:11 |
| TheJulia | oh | 11:11 |
| *** belmoreira has quit IRC | 11:11 | |
| * TheJulia wonders if why not both? | 11:11 | |
| cdent | TheJulia: yeah, jroll is right about what i meant, but it "both" is a very good question | 11:11 |
| cdent | an IronicDirect (in the vein of PlacementDirect [1]) might be an interesting experiment: | 11:12 |
| TheJulia | I mean... we can improve ironic's api to allow a user to say "forget pagination, just give me data now"... which might help other interactions | 11:12 |
| cdent | [1] https://opendev.org/openstack/placement/src/branch/master/placement/direct.py | 11:12 |
| jroll | IronicDirect == nova-baremetal with separate worker processes :D | 11:13 |
| cdent | jroll: yes that would be tidier | 11:13 |
| *** belmoreira has joined #openstack-ironic | 11:14 | |
| TheJulia | cdent: there is a legitimate reason keep rest api for interaction with ironic.... at least in developing scenarios where users want/need multitenancy in ironic. I guess that makes me think the best thing is to keep that surface and interaction relatively consistent and just start addressing the data return performance that folks like jroll see. | 11:15 |
| * cdent nods | 11:16 | |
| cdent | multi-pronged attack | 11:16 |
| jroll | I really do think the killer performance boost will come from ironic managing its resources in placement. | 11:17 |
| TheJulia | FWIW, I'm actively pushing someone to get someone else to take some interns code and get it upstream to make multitenant ironic. It was done as a proof of concept this summer, but they didn't want to touch the underlying ironic data model (which is actually a necessary change for performance) | 11:17 |
| jroll | I 100% believe we'll continue to see problems in this area until we do that | 11:17 |
| jroll | (alternate solutions include separating the resource tracker loop into a separate process) | 11:18 |
| TheJulia | jroll: I don't think we will ever get rid of all of the 'problems', there will always be something out there | 11:18 |
| jroll | TheJulia: sorry, the performance problems in the resource tracker when running ironic, specifically | 11:18 |
| TheJulia | jroll: so in this, are you thinking ironic reporting to placement directly? | 11:18 |
| jroll | yes | 11:19 |
| TheJulia | I think that too is a worthwhile thing | 11:19 |
| arne_wiebalck | Ironic managin resources in placement has probably higher chances of being accepted than allowing for direct db access, no? | 11:19 |
| arne_wiebalck | managing | 11:19 |
| TheJulia | arne_wiebalck: I would suspect so.... | 11:19 |
| jroll | goal being to make the RT a noop (though it might still be needed to manage the compute_nodes table? is that ever going away?) | 11:19 |
| *** zhangfei has quit IRC | 11:23 | |
| *** ijw has joined #openstack-ironic | 11:27 | |
| *** ijw has quit IRC | 11:31 | |
| *** kaifeng has quit IRC | 11:42 | |
| *** Lucas_Gray has quit IRC | 11:45 | |
| *** rh-jelabarre has joined #openstack-ironic | 12:11 | |
| openstackgerrit | Varsha Verma proposed openstack/sushy-tools master: Add Storage and Storage Controllers resource support https://review.opendev.org/674339 | 12:30 |
| *** ricolin has joined #openstack-ironic | 12:37 | |
| openstackgerrit | Varsha Verma proposed openstack/sushy-tools master: Add Drive resource support https://review.opendev.org/675085 | 12:39 |
| *** rloo has joined #openstack-ironic | 12:47 | |
| *** verma-varsha has joined #openstack-ironic | 12:53 | |
| *** tssurya has quit IRC | 13:03 | |
| *** tssurya has joined #openstack-ironic | 13:04 | |
| openstackgerrit | Julia Kreger proposed openstack/python-ironicclient master: Remove the ironic command https://review.opendev.org/676515 | 13:29 |
| *** verma-varsha has quit IRC | 13:38 | |
| *** priteau has joined #openstack-ironic | 13:38 | |
| *** belmoreira has quit IRC | 13:45 | |
| *** belmoreira has joined #openstack-ironic | 13:50 | |
| *** verma-varsha has joined #openstack-ironic | 14:19 | |
| *** diga has quit IRC | 14:20 | |
| *** mkarpiarz has joined #openstack-ironic | 14:28 | |
| *** cdearborn has joined #openstack-ironic | 14:37 | |
| *** gyee has joined #openstack-ironic | 14:57 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Change [deploy]/default_boot_option to local https://review.opendev.org/492114 | 14:59 |
| *** kaiokmo has quit IRC | 15:06 | |
| verma-varsha | etingof: Why are we checking hasattr() here (https://github.com/openstack/sushy-tools/blob/master/sushy_tools/emulator/resources/indicators/staticdriver.py#L39). Wouldn't it always have the make_permanent attribute? | 15:12 |
| etingof | verma-varsha, o/ I think that's for the sake of unit-testing | 15:13 |
| verma-varsha | etingof: Also, is it possible to use tuple as a key for the PersistantDict? | 15:15 |
| etingof | I think the idea is that in some cases _indicators can be just 'dict' | 15:15 |
| etingof | yes, it's normal dict so anything immutable can serve as a key | 15:15 |
| verma-varsha | Thanks! I am thinking of using a tuple as a key for Volumes just as I did for Drives. | 15:16 |
| etingof | since you are on that, you must have the best judgement! | 15:24 |
| *** belmoreira has quit IRC | 15:26 | |
| *** belmoreira has joined #openstack-ironic | 15:28 | |
| *** belmoreira has quit IRC | 15:38 | |
| *** sthussey has joined #openstack-ironic | 15:38 | |
| cdearborn | good morning ironic | 15:41 |
| *** verma-varsha has quit IRC | 15:41 | |
| cdearborn | i'm trying to get a devstack environment working | 15:41 |
| cdearborn | and getting a failure that the sushy library is not installed | 15:41 |
| cdearborn | any thoughts on how to make devstack install it? | 15:42 |
| TheJulia | cdent: hmm | 15:45 |
| TheJulia | err | 15:45 |
| TheJulia | cdearborn: hmm | 15:45 |
| cdent | :) | 15:45 |
| * etingof thinks that fishy stuff gets pulled conditionally - when redfish hw type is being used | 15:46 | |
| *** efried has quit IRC | 15:46 | |
| TheJulia | yeah, it is based on what hardware type is set | 15:46 |
| * etingof would first configure devstack (local.conf), then run the whole thing | 15:46 | |
| etingof | so devstack would know what components should be installed | 15:47 |
| TheJulia | well, it is all about is_*_deployed | 15:47 |
| TheJulia | is_drac_enabled triggers python-dracclient being installed | 15:47 |
| etingof | does dracclient invoke sushy? | 15:48 |
| TheJulia | i guess not... | 15:48 |
| etingof | yeah... | 15:48 |
| cdearborn | hrm - i have IRONIC_ENABLED_HARDWARE_TYPE=idrac,redfish,ipmi | 15:49 |
| etingof | how about IRONIC_IS_HARDWARE? | 15:50 |
| cdearborn | set to True | 15:50 |
| etingof | hmm | 15:50 |
| * etingof would look into bash output when ironic devstack plugin is running | 15:52 | |
| TheJulia | so we need to revise the logic around 691 in the devstack plugin | 15:52 |
| cdearborn | urgh - found it | 15:52 |
| TheJulia | specifically add is_redfish_enabled | 15:52 |
| cdearborn | so, i have IRONIC_DEPLOY_DRIVER=ipmi | 15:53 |
| *** cdent has quit IRC | 15:53 | |
| cdearborn | i thought that setting was specifically for the VMs that are created for emulating baremetal? | 15:53 |
| TheJulia | no, it is the default hardware type to use for enrollment | 15:53 |
| cdearborn | ah - thx | 15:54 |
| etingof | the other interesting thing is here -- https://github.com/openstack/ironic/blob/master/devstack/lib/ironic#L1039 | 15:54 |
| cdearborn | soooo - i was trying to set up a devstack where i could use either redfish or idrac hardware type | 15:54 |
| *** verma-varsha has joined #openstack-ironic | 15:55 | |
| cdearborn | but, looking at the code, it appears that it will only install the libraries for whatever you have IRONIC_DEPLOY_DRIVER set to | 15:55 |
| cdearborn | is that right? | 15:57 |
| etingof | should we rather use IRONIC_ENABLED_HARDWARE_TYPES in is_deployed_by*? | 15:58 |
| *** lucasagomes has quit IRC | 15:58 | |
| etingof | or may be better to call setup_sushy() based on redfish presence in IRONIC_ENABLED_HARDWARE_TYPES, not in IRONIC_DEPLOY_DRIVER? | 16:00 |
| cdearborn | i like that idea, but this would have to change... https://github.com/openstack/ironic/blob/master/devstack/lib/ironic#L2150 | 16:00 |
| *** dsneddon has joined #openstack-ironic | 16:02 | |
| cdearborn | well, at any rate, i can setup devstack for 1 driver, and if i need to switch, stand up a new devstack vm or redeploy devstack with the other driver | 16:02 |
| TheJulia | it is an easy enough thing to fix | 16:02 |
| TheJulia | I can change it right now and you can base on top of the patch | 16:02 |
| etingof | perhaps the idea is that you have to respin devstack every time you change IRONIC_DEPLOY_DRIVER so it will reinstall driver requirements | 16:06 |
| *** verma-varsha has quit IRC | 16:07 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Install sushy if redfish is a hardware type https://review.opendev.org/676732 | 16:08 |
| TheJulia | But for doing something fancy, it might make sense not to really | 16:10 |
| etingof | "fishy" is the new convention | 16:11 |
| cdearborn | TheJulia, etingof: thx! | 16:11 |
| TheJulia | so fishy things can be fishy? | 16:12 |
| TheJulia | :) | 16:12 |
| arne_wiebalck | TheJulia: I've left reviews on 3 of the 4 open IPA s/w RAID related patches now. Have to leave now, will try to have a look at the last one tmrw. | 16:15 |
| arne_wiebalck | TheJulia: reviews/commens | 16:15 |
| TheJulia | arne_wiebalck: thanks! :) | 16:15 |
| arne_wiebalck | bye, everyone o/ | 16:15 |
| *** ricolin has quit IRC | 16:17 | |
| *** tssurya has quit IRC | 16:18 | |
| *** verma-varsha has joined #openstack-ironic | 16:21 | |
| *** absubram has joined #openstack-ironic | 16:26 | |
| *** e0ne has quit IRC | 16:28 | |
| *** absubram_ has joined #openstack-ironic | 16:28 | |
| *** absubram has quit IRC | 16:31 | |
| *** absubram_ is now known as absubram | 16:31 | |
| *** alexmcleod has quit IRC | 16:33 | |
| *** absubram has quit IRC | 16:37 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Change [deploy]/default_boot_option to local https://review.opendev.org/492114 | 16:39 |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: update some ci jobs https://review.opendev.org/676734 | 16:39 |
| *** mgoddard_ has quit IRC | 17:04 | |
| *** derekh has quit IRC | 17:05 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic master: Support firmware upgrade for irmc driver https://review.opendev.org/599992 | 17:16 |
| openstackgerrit | Nisha Brahmankar proposed openstack/ironic-python-agent master: Adds CI job that uses DIB build images https://review.opendev.org/676239 | 17:21 |
| *** verma-varsha has quit IRC | 17:27 | |
| *** priteau has quit IRC | 17:39 | |
| *** e0ne has joined #openstack-ironic | 17:52 | |
| *** verma-varsha has joined #openstack-ironic | 17:54 | |
| *** absubram has joined #openstack-ironic | 18:13 | |
| *** e0ne has quit IRC | 18:36 | |
| *** ociuhandu has joined #openstack-ironic | 18:56 | |
| *** ociuhandu has quit IRC | 18:58 | |
| *** verma-varsha has quit IRC | 19:12 | |
| *** gregwork has quit IRC | 19:16 | |
| *** gregwork has joined #openstack-ironic | 19:17 | |
| *** verma-varsha has joined #openstack-ironic | 19:26 | |
| *** vesper11 has quit IRC | 19:30 | |
| TheJulia | looks like we're getting random timeouts on standalone jobs | 19:34 |
| TheJulia | And that is causing an abnormally high failure count | 19:35 |
| TheJulia | I'm going to take a look after I have some dinner | 19:35 |
| *** jcoufal has joined #openstack-ironic | 19:41 | |
| *** verma-varsha has quit IRC | 19:49 | |
| *** jcoufal has quit IRC | 19:49 | |
| *** dougsz has joined #openstack-ironic | 20:10 | |
| *** dougsz has quit IRC | 20:18 | |
| *** jtomasek has quit IRC | 20:27 | |
| *** dougsz has joined #openstack-ironic | 20:33 | |
| *** cdearborn has quit IRC | 20:37 | |
| *** jcoufal has joined #openstack-ironic | 20:43 | |
| *** dougsz has quit IRC | 20:48 | |
| *** jcoufal has quit IRC | 20:49 | |
| *** vesper11 has joined #openstack-ironic | 20:51 | |
| *** bdodd has quit IRC | 20:56 | |
| openstackgerrit | Merged openstack/ironic master: Do not tear down node upon cleaning failure https://review.opendev.org/650942 | 21:14 |
| openstackgerrit | Steve Baker proposed openstack/metalsmith master: Allow reserve_node to backfill from existing node https://review.opendev.org/676525 | 21:43 |
| openstackgerrit | raphael.glon proposed openstack/ironic-python-agent master: Softraid: partitioning fixes https://review.opendev.org/674819 | 21:48 |
| openstackgerrit | raphael.glon proposed openstack/ironic-python-agent master: Software raid: default partition table type change https://review.opendev.org/674803 | 22:01 |
| *** trident has quit IRC | 22:03 | |
| *** ijw has joined #openstack-ironic | 22:05 | |
| *** trident has joined #openstack-ironic | 22:10 | |
| *** ijw has quit IRC | 23:29 | |
| *** absubram has quit IRC | 23:33 | |
| *** sthussey has quit IRC | 23:34 | |
| openstackgerrit | Digambar proposed openstack/ironic stable/rocky: DRAC: Fix OOB introspection to use pxe_enabled flag in idrac driver https://review.opendev.org/648360 | 23:39 |
| *** igordc has joined #openstack-ironic | 23:46 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!