| clarkb | I'm not seeing a source fs option for mkfs.xfs which may be why | 00:00 |
|---|---|---|
| clarkb | this is likely why they went the route of loopback mounting | 00:01 |
| clarkb | it is ultimately more flexible but does cost you root | 00:01 |
| tobiash | Probably | 00:04 |
| *** rlandy has quit IRC | 00:46 | |
| *** harlowja has quit IRC | 00:46 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Late bind projects https://review.openstack.org/553618 | 01:20 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP: Make config objects freezable https://review.openstack.org/562816 | 01:20 |
| *** threestrands has joined #zuul | 02:59 | |
| *** swest has joined #zuul | 05:07 | |
| *** elyezer has quit IRC | 06:29 | |
| *** hashar has joined #zuul | 06:31 | |
| openstackgerrit | Markus Hosch proposed openstack-infra/zuul master: Store build start/end time in UTC https://review.openstack.org/562273 | 06:49 |
| *** threestrands has quit IRC | 07:00 | |
| openstackgerrit | Benedikt Löffler proposed openstack-infra/zuul master: Add role information to task in zuul_json callback https://review.openstack.org/559998 | 07:10 |
| openstackgerrit | Benedikt Löffler proposed openstack-infra/zuul master: Add role information to task in zuul_json callback https://review.openstack.org/559998 | 07:34 |
| *** ssbarnea_ has joined #zuul | 08:00 | |
| *** jpena|off is now known as jpena | 08:00 | |
| *** electrofelix has joined #zuul | 08:33 | |
| *** hashar has quit IRC | 09:19 | |
| *** bhavik1 has joined #zuul | 09:39 | |
| *** rcarrillocruz has joined #zuul | 09:40 | |
| *** bhavik1 has quit IRC | 09:47 | |
| *** gundalow has joined #zuul | 10:04 | |
| *** xinliang has quit IRC | 10:42 | |
| *** xinliang has joined #zuul | 10:42 | |
| *** xinliang has quit IRC | 10:42 | |
| *** xinliang has joined #zuul | 10:42 | |
| *** pwhalen has quit IRC | 10:43 | |
| *** pwhalen has joined #zuul | 10:44 | |
| *** pwhalen has joined #zuul | 10:44 | |
| *** ssbarnea_ has quit IRC | 11:00 | |
| *** ssbarnea_ has joined #zuul | 11:01 | |
| *** ssbarnea_ has quit IRC | 11:11 | |
| *** ssbarnea_ has joined #zuul | 11:14 | |
| *** rlandy has joined #zuul | 12:32 | |
| *** elyezer has joined #zuul | 12:40 | |
| *** JasonCL has quit IRC | 13:35 | |
| tobiash | corvus: I have a use case I currently don't know how to solve this: a project wants a buildset which first builds a docker image (if some files changed) and after that several jobs that use that image or the current latest image | 13:40 |
| tobiash | using job dependencies doesn't work because the child jobs won't run if the first is filtered out | 13:40 |
| tobiash | do you think it would be an option to mark dependencies to non-existing jobs in the current buildset satisfied? | 13:42 |
| tobiash | I think that could make that optional dependency work | 13:42 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul master: WIP: Update to Ansible 2.5 https://review.openstack.org/562668 | 13:47 |
| corvus | tobiash: you could always run the build job and have it short-circuit if nothing has changed | 13:47 |
| tobiash | corvus: yes, that was my first thought, but that still would allocate an unneeded node, push repos there and then notice that nothing has to be done | 13:48 |
| Shrews | this is getting fun: https://review.openstack.org/#/c/562668/5/zuul/ansible/callback/zuul_json.py | 13:48 |
| tobiash | so that is not an optimal solution | 13:48 |
| tobiash | Shrews: oh, a try-catch chain | 13:49 |
| corvus | tobiash: true. i think it'd have to be an option then, that you declare on the dependency. because sometimes you may want that behavior, or sometimes you may want the behavior where nothing runs. | 13:49 |
| tobiash | corvus: as dependencies is currently just a list, add a weak-dependencies? | 13:50 |
| *** JasonCL has joined #zuul | 13:50 | |
| tobiash | corvus: two variants: https://etherpad.openstack.org/p/ORaD48jE56 | 13:52 |
| tobiash | I think I'd favor the first one | 13:53 |
| *** dims has quit IRC | 13:58 | |
| corvus | tobiash: yeah, we've done the string-or-dict thing before. i prefer it to option 2. there's also option 3. | 14:00 |
| *** mnaser has quit IRC | 14:00 | |
| *** mnaser has joined #zuul | 14:01 | |
| tobiash | corvus: option 2 would be fine for me either | 14:01 |
| tobiash | that's probably the simplest for the user | 14:02 |
| *** dims has joined #zuul | 14:04 | |
| Shrews | w00t. got 2.5 working. well, 2.5.1 | 14:13 |
| Shrews | i need to set that as the min in requirements.txt i guess since we didn't actually test 2.5.0 | 14:14 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul master: Update to Ansible 2.5 https://review.openstack.org/562668 | 14:16 |
| Shrews | unWIP'd | 14:17 |
| corvus | reminder: today would be a great today to jot down some notes in the weekly update etherpad: https://etherpad.openstack.org/p/zuul-update-email | 14:21 |
| *** acozine1 has joined #zuul | 14:31 | |
| *** nguyenhai has quit IRC | 14:33 | |
| *** JasonCL has quit IRC | 14:43 | |
| *** JasonCL has joined #zuul | 14:45 | |
| tobiash | Shrews: yay, thanks for doing my task for this week :) | 14:45 |
| corvus | we should say tobiash plans to cheer Shrews on about upgrading ansible :) | 14:49 |
| Shrews | that's what i do. hide behind the scenes making others look good :) | 14:51 |
| Shrews | this is how mordred became so successful | 14:51 |
| Shrews | :D | 14:52 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add regex support to project stanzas https://review.openstack.org/535713 | 15:06 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add regex support to project stanzas https://review.openstack.org/535713 | 15:08 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add regex support to project stanzas https://review.openstack.org/535713 | 15:30 |
| tobiash | corvus: that conflicted with your config refactor stack so I rebased it onto your stack ^ | 15:32 |
| *** electrofelix has quit IRC | 15:33 | |
| openstackgerrit | Merged openstack-infra/zuul master: Update encrypt_secret to be API aware https://review.openstack.org/561681 | 15:40 |
| *** rbergeron has quit IRC | 15:49 | |
| *** dims has quit IRC | 16:06 | |
| *** dims has joined #zuul | 16:11 | |
| SpamapS | has anybody made a role for pulling all the journals on a box with journalctl? | 16:30 |
| pabelanger | SpamapS: not yet, I've just restored to in-tree playbook: http://git.openstack.org/cgit/openstack/windmill/tree/tests/collect-logs.yaml But would happily use a role if there was one | 16:32 |
| pabelanger | resorted* | 16:33 |
| clarkb | SpamapS: pabelanger devstack does something like that | 16:38 |
| clarkb | maybe that role should be made more generic | 16:39 |
| clarkb | We do human readable output for openstack services and "syslog" but also dump the full journal in an export format then xz them (as it does far better than gzip on the data) | 16:39 |
| pabelanger | clarkb: yah, making it more generic would be helpful | 16:40 |
| clarkb | SpamapS: are you looking for the export format or human readable or both? | 16:42 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Only emit parent-change-enqueued if needed https://review.openstack.org/563194 | 16:50 |
| *** rlandy is now known as rlandy|brb | 16:52 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Limit search scope of getChangesDependingOn to tenant https://review.openstack.org/563196 | 17:09 |
| *** jpena is now known as jpena|off | 17:09 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Only emit parent-change-enqueued if needed https://review.openstack.org/563194 | 17:10 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Limit search scope of getChangesDependingOn to tenant https://review.openstack.org/563196 | 17:10 |
| corvus | clarkb, tobiash: copy.deepcopy doesn't work on a mappingproxy applied to an ordereddict :( | 17:12 |
| clarkb | corvus: it isn't a recursive copy? | 17:12 |
| clarkb | I guess that makes sense? unfortunate though | 17:12 |
| corvus | yeah, i think it's trying to reconstruct a mappingproxy | 17:13 |
| corvus | oh, actually, it may not even work on a mappingproxy with a regular dict | 17:13 |
| corvus | yep, this fails: copy.deepcopy(types.MappingProxyType({})) | 17:14 |
| *** rlandy|brb is now known as rlandy | 17:14 | |
| tobiash | corvus: do we still have to do deep copies with your refactor? | 17:14 |
| corvus | that puts a crimp in my plans | 17:14 |
| corvus | tobiash: i'm not sure -- the main thing i'm worried about is when we create frozen jobs from variants. we copy a bunch of the variant attributes onto the job. it's possible some or all of those don't need to be deepcopies anymore. | 17:15 |
| corvus | (indeed, "# TODO(jeblair): determine if deepcopy is required" it already may not be required :) | 17:16 |
| corvus | i guess i will be doing that todo right now :) | 17:18 |
| tobiash | :) | 17:18 |
| tobiash | and yay finally I have at least one item on my list from this week | 17:19 |
| corvus | wow i found an undocument/unused job attribute :| | 17:20 |
| corvus | okay, that section is safe to remove deepcopy. i think i also need to check variables. | 17:23 |
| corvus | this is going to take some time, but i think it's going to be worth doing (as it should clean up a bunch of unecessary copying) | 17:23 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP: Make config objects freezable https://review.openstack.org/562816 | 17:24 |
| corvus | that's just a checkpoint, i have to afk for a bit | 17:24 |
| AJaeger | corvus, tobiash, tristanC, regarding https://review.openstack.org/548082 - should we handle a single command - or allow multiple ones as tristanC suggests? | 17:25 |
| tobiash | I think if we do that job I'd slightly prefer a command list as that's more flexible for the easy use cases | 17:27 |
| tobiash | but a downside of this is that it encourages overuse | 17:27 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: DNM: test tox-cover https://review.openstack.org/563199 | 17:33 |
| AJaeger | let me try that in a separate change... | 17:39 |
| tobiash | AJaeger: commented on 559348 | 17:42 |
| AJaeger | tobiash: thanks | 17:48 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Add simple run-test-commands job https://review.openstack.org/563204 | 17:52 |
| AJaeger | tristanC: that's the alternative ^ | 17:53 |
| fungi | i'm not sure what the reason is for the +2 dog-piling on 535543 but i'm going to break the cycle by approving it | 18:05 |
| fungi | or does tristanC need to acknowledge the patchset? | 18:05 |
| fungi | since i guess he wasn't the uploader for ps17 | 18:06 |
| SpamapS | clarkb: I have roles that pull individual journals in human readable format for convenience and debugging. | 18:06 |
| SpamapS | clarkb: I wouldn't know what to do with the binary versions. :-P | 18:07 |
| SpamapS | Would be really cool if I could do the binary versions and then html-prettied versions. | 18:07 |
| fungi | i think the devstack docs mention how to make use of the exported binary format | 18:08 |
| fungi | arguably convenient since you get things like inherent filtering support | 18:09 |
| fungi | but yes, i always end up having to reread instructions for that | 18:09 |
| SpamapS | Yeah if I knew the tools I'd probably like that. Though usually I just want to poke around in the browser and find a string or two | 18:09 |
| SpamapS | like "what sploded?" | 18:09 |
| fungi | same here | 18:13 |
| clarkb | the killer feature is having everything interleaved but being able to filter out what you don't want | 18:23 |
| clarkb | rather than multiple disparate human readable files | 18:23 |
| *** hashar has joined #zuul | 18:25 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Add simple run-test-commands job https://review.openstack.org/563204 | 18:35 |
| fungi | https://review.openstack.org/563217 is a straw man for the sb project group reorganizing i mentioned on tuesday | 18:36 |
| fungi | just took a stab at guessing which repos we want in which group(s) | 18:37 |
| AJaeger | tristanC: could you check https://review.openstack.org/#/c/563204 , please? That's an alternative | 18:46 |
| * AJaeger signes off and wishes a great weekend | 18:46 | |
| fungi | have a great weekend yourself, AJaeger! | 18:46 |
| AJaeger | thanks! | 18:46 |
| *** harlowja has joined #zuul | 19:02 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: WIP: Test parent-change-enqueued with github https://review.openstack.org/563242 | 19:11 |
| *** acozine1 has quit IRC | 19:11 | |
| *** JasonCL has quit IRC | 19:13 | |
| *** elyezer has quit IRC | 19:14 | |
| SpamapS | what I really want is ARA but for journals | 19:18 |
| SpamapS | so like "show me only errors" | 19:18 |
| openstackgerrit | melissaml proposed openstack-infra/zuul master: Trivial: Update pypi url to new url https://review.openstack.org/563248 | 19:19 |
| *** elyezer has joined #zuul | 19:26 | |
| *** JasonCL has joined #zuul | 20:33 | |
| *** hashar has quit IRC | 20:39 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Test parent-change-enqueued with github https://review.openstack.org/563242 | 20:47 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Extend github testing using app auth https://review.openstack.org/563297 | 20:47 |
| *** ssbarnea_ has quit IRC | 21:04 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Support getting the github client via installation id https://review.openstack.org/563311 | 21:06 |
| corvus | SpamapS: maybe os- | 21:20 |
| corvus | gr | 21:21 |
| corvus | SpamapS: maybe os-loganalyze can be expanded to support journal files, if it doesn't already? | 21:21 |
| SpamapS | corvus: yeah that would be sweet | 21:21 |
| corvus | clarkb: ^ | 21:21 |
| corvus | my feeling is that it's not *very* openstack specific, so could probably be generalized without too much trouble | 21:22 |
| clarkb | it came up when we added those files to the log collection | 21:22 |
| corvus | i thought that sounded familiar :) | 21:22 |
| clarkb | the format is quite good for machines. My only concern with a direct rendering is I'm unsure of how stable the format is | 21:23 |
| clarkb | can also use journalctl and similar tools to deal with it if not very stable | 21:23 |
| corvus | i'm going to bow out of this conversation though, cause i don't have anything nice to say about binary logfiles | 21:23 |
| clarkb | oh the export format isn't binary | 21:25 |
| clarkb | at least I don't think it is | 21:25 |
| SpamapS | I'm pretty sure journalctl is the only safe way. | 21:35 |
| SpamapS | unless it has an underlying library. | 21:35 |
| *** rlandy has quit IRC | 21:53 | |
| SpamapS | need second set of eyes | 22:32 |
| SpamapS | http://paste.openstack.org/show/719669/ | 22:32 |
| SpamapS | what's wrong with that? | 22:32 |
| SpamapS | http://paste.openstack.org/show/719670/ <-- error I'm getting | 22:33 |
| clarkb | corvus: https://review.openstack.org/#/c/562273/3 straightforward timestamp + database + timezone issue if you want to take a look (I didn't approve because I think eyeballs on potentially end user behavior alterations even if minor deserve eyeballs) | 22:35 |
| clarkb | SpamapS: my reading of that is zuul expects a single string dependencies value? | 22:37 |
| *** JasonCL has quit IRC | 22:37 | |
| clarkb | rather than a list type | 22:37 |
| clarkb | the docs definitely say it should be a list though | 22:38 |
| *** JasonCL has joined #zuul | 22:38 | |
| *** JasonCL has joined #zuul | 22:39 | |
| clarkb | SpamapS: http://paste.openstack.org/show/719671/ is an example from the tests, maybe try converting your jobs: list to be like that rather than name: foo\ndependencies: | 22:40 |
| clarkb | SpamapS: but keep it as a list | 22:40 |
| openstackgerrit | Merged openstack-infra/zuul master: Remove extra words from gerrit setup docs https://review.openstack.org/561672 | 22:40 |
| openstackgerrit | Merged openstack-infra/zuul master: Update test instructions https://review.openstack.org/562289 | 22:40 |
| clarkb | ya looks like its tested as both a list and a string, I think it may just be the yaml structure off earlier | 22:41 |
| SpamapS | ohhhhhhh | 22:41 |
| SpamapS | ok | 22:41 |
| SpamapS | it's the name: thing | 22:41 |
| SpamapS | should be deploy-to-prod: | 22:41 |
| clarkb | ya | 22:41 |
| SpamapS | http://paste.openstack.org/show/719672/ | 22:45 |
| SpamapS | fixed it | 22:45 |
| *** JasonCL has quit IRC | 22:48 | |
| clarkb | fungi: https://review.openstack.org/#/c/554352/6 looks ready except for tobiash's comment, do you want to do a quick update on that? | 22:55 |
| * fungi catches back up after a very fine blackened sheepshead fillet | 23:09 | |
| fungi | clarkb: corvus had mentioned he was planning to add his ssh key there to fix that todo. my concern is keeping some reminder for ourselves that we should have a minimum of 3 contacts listed there | 23:10 |
| clarkb | ah | 23:11 |
| corvus | clarkb: re timezone -- yeah, i think merging 562273 is the only thing we can do. folks with servers in utc (like openstack) will be unchanged, but folks with localized timezones will have a gregorian event. | 23:11 |
| clarkb | corvus: it will soon pass fortunately | 23:12 |
| corvus | clarkb: yeah, though that reminds me that expiration is still unimplemented :) | 23:12 |
| clarkb | I'm currently trying to understand zuu's symlinking of roles so I can approve this tobiash bugfix | 23:12 |
| clarkb | corvus: tobiash in https://review.openstack.org/#/c/560554/2//COMMIT_MSG the error message seems to indicate to me that it did use the name ansible and that didn't get truncated to '' | 23:15 |
| corvus | clarkb: i believe there's supposed to be an "ansible" at the endo f the role_0 thing | 23:16 |
| clarkb | corvus: oh, the arrow points in the wrong direction | 23:17 |
| clarkb | that is not confusing at all python | 23:17 |
| corvus | clarkb: nope. i feel it's par for the course with symlinks. | 23:17 |
| clarkb | I'm going to test that really quickly locally but ya that makes sense if that is what it is | 23:17 |
| clarkb | that is totally what it is | 23:18 |
| corvus | clarkb: yeah, do test, but iirc, that's supposed to be creating a directory (role_0) with one item in it (ansible) which is a symlink to a repo. then role_0 gets added to the roles path, so if you do "roles: ansible" it finds that in the path | 23:18 |
| clarkb | imo that arrow should point the other way since it is creating a link role_0 to target ansible | 23:19 |
| clarkb | not target ansible to link role_0 | 23:19 |
| corvus | agreed | 23:20 |
| corvus | that set me back like 2 hours when debugging this originally | 23:21 |
| corvus | error messages 101: it's better to say nothing than say the opposite of the truth | 23:22 |
| corvus | clarkb: one thing about mappingproxy: isinstance(types.MappingProxyType({}), dict) == False | 23:24 |
| clarkb | hrm | 23:26 |
| clarkb | I guess from a is it a duck standpoint that is true | 23:26 |
| clarkb | (it has a reduced api) | 23:26 |
| clarkb | and can't quack | 23:26 |
| corvus | yeah. but i only care if, um, it has feathers. | 23:26 |
| fungi | peking duck has no feathers | 23:27 |
| corvus | so, it makes things like traversal difficult (where you do "if list; elif dict; else ...") | 23:27 |
| corvus | fungi: but lots of fat, and so do pigs... | 23:27 |
| fungi | mmm, duck/porkfat | 23:27 |
| corvus | (at least, it's *supposed* to have lots of fat) | 23:27 |
| openstackgerrit | Merged openstack-infra/zuul master: Store build start/end time in UTC https://review.openstack.org/562273 | 23:27 |
| clarkb | hrm I kinda awnt duck now | 23:28 |
| clarkb | maybe that is what I can grill this weekend | 23:28 |
| clarkb | or smoke | 23:28 |
| fungi | smoked duck might turn out really tasty | 23:29 |
| corvus | ++ | 23:31 |
| openstackgerrit | Merged openstack-infra/zuul master: Fix implicit role for repos named ansible https://review.openstack.org/560554 | 23:34 |
| corvus | okay, i think i've ended up getting all the deepcopies out of the model (except on the unparsedtenantconfig, but that's fine for now, and it'll go away when we fully reuse config objects. i think) | 23:34 |
| corvus | i also found a solution to actually make mappingproxy deep-copyable, but i'd prefer not to use it | 23:34 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP: Make config objects freezable https://review.openstack.org/562816 | 23:35 |
| clarkb | corvus: tobiash I've approved https://review.openstack.org/#/c/559745/3 but also noted a potential cleanup there | 23:39 |
| clarkb | corvus: tristanC for the mqtt connection stuff, do we really need to explicitly set the tls versions? isn't that something clients typically do based on what the server supports? | 23:41 |
| clarkb | feel similar to the ipv6 + statsd thing. Is this a knob that we really need or can it be determined at runtime | 23:41 |
| clarkb | (and we'll haev to update zuul everytime tls makes a new version) | 23:42 |
| corvus | clarkb: yeah, i like the extra caution around the release since it's in an exception handler. it's a bit more future proof. | 23:43 |
| corvus | clarkb: maybe a q for mtreinish? | 23:44 |
| clarkb | "By default (if the python version supports it) the highest TLS version is detected. If unavailable, TLS v1 is used." says the docs | 23:46 |
| clarkb | so I guess the issue will be if the server supports older TLS than the client | 23:47 |
| corvus | clarkb: sounds like it doesn't have a fallback process? | 23:49 |
| corvus | ie, a negotiation | 23:49 |
| clarkb | https://docs.python.org/3/library/ssl.html#ssl.PROTOCOL_TLS is the negotiation process I think | 23:49 |
| clarkb | which it falls back on if tlsv1.2 fails | 23:50 |
| clarkb | oh wait hrm | 23:50 |
| clarkb | I think it would be worth testing using ^ | 23:50 |
| clarkb | oh taht is new in python3.6? | 23:51 |
| openstackgerrit | Merged openstack-infra/zuul master: Fix missing semaphore release on zk error https://review.openstack.org/559745 | 23:51 |
| clarkb | no it says that but its in 3.5 docs too | 23:52 |
| clarkb | I think we may want to use that and then exclude ssl | 23:52 |
| clarkb | https://github.com/eclipse/paho.mqtt.python/blob/master/src/paho/mqtt/client.py#L684-L690 | 23:53 |
| corvus | clarkb: it's also in the 2.7 docs | 23:54 |
| clarkb | my reading of that is we maybe don't need to worry about tls versions since we require python3.5 anyways which should be new enough to support that | 23:54 |
| corvus | https://docs.python.org/2.7/library/ssl.html#ssl.PROTOCOL_TLS | 23:54 |
| corvus | something's really weird with that | 23:54 |
| corvus | clarkb: i agree this is worth holding the change for | 23:55 |
| clarkb | the difference between python3.6 and python3.5 is 3.6 won't use sslv3 or v2 at all | 23:56 |
| clarkb | but 3.5 will if tls isn't supported | 23:56 |
| clarkb | sorry 3.5 will only use sslv3 if not tls, it will not use sslv2 | 23:56 |
| clarkb | I think what we want to do is rely on the PROTOCOL_TLS constant and then maybe do a pull request to paho mqtt to disable SSLv3 at https://github.com/eclipse/paho.mqtt.python/blob/master/src/paho/mqtt/client.py#L690 using https://docs.python.org/3.5/library/ssl.html#protocol-versions or even https://docs.python.org/3.5/library/ssl.html#ssl.create_default_context | 23:58 |
| clarkb | (I'm not super concerned about us accidentally running an mqtt without tls support though | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!