=== gary_poster is now known as gary_poster|away | ||
rogpeppe | mornin' all | 08:18 |
---|---|---|
dimitern | morning | 08:26 |
TheMue | morning | 08:35 |
jam | afternoon all :) | 08:45 |
mramm | morning all | 10:17 |
jam | morning mramm | 10:23 |
dimitern | jam, we'll probably skip the standup, due to the other call | 10:36 |
jam | standup: https://plus.google.com/hangouts/_/7acpielmpa29pcug39874tclck | 10:44 |
jam | we're using a different hangout today | 10:44 |
jam | mgz: mramm, TheMue: ^^ | 10:45 |
jam | natefinch: ^^ | 10:45 |
TheMue | sry, phone call | 10:46 |
mgz | er, copying hangout link | 10:48 |
* rogpeppe goes to grab a bowl of cereal | 11:18 | |
jam | so somehow I ran "juju debug-hooks" and exited in a way that running it again tells me "duplicate session" | 11:18 |
jam | is there a way to get *into* that session? | 11:18 |
axw__ | jam: you should be able to "tmux attach-session -t <unit-name>" | 11:27 |
axw__ | sounds like a nasty bug tho | 11:27 |
jam | axw__: thanks, I couldn't find attach-session in the command list | 11:27 |
jam | axw__: if you use ^A D to exit the session with a disconnect rather than stopping it | 11:33 |
jam | it obviously leaves the session running | 11:33 |
jam | which then means the process is a bit hung there | 11:33 |
jam | axw__: is there a "create a new session or attach to an existing one" ? | 11:33 |
jam | I suppose "tmux attach-session -s $NAME || tmux new-session -s $NAME" ? | 11:34 |
axw__ | jam: not sure. I think the script did that at one point tho | 11:35 |
axw__ | so you could check history | 11:35 |
* axw__ is still cooking or would look himself | 11:35 | |
jam | anyway, not what I'm working on *right* now :) | 11:35 |
jam | axw__: how do you cook and irc at the same time? | 11:35 |
axw__ | intermittently | 11:35 |
natefinch | that's what fire extinguishers are for | 11:36 |
axw__ | night all | 11:50 |
jam | natefinch: https://codereview.appspot.com/33870043/ | 12:34 |
jam | first one is just a straight rollback | 12:34 |
jam | natefinch: then https://codereview.appspot.com/33890043 which bumps the revno back to 1.16.4 | 12:35 |
jam | natefinch: and finally: https://codereview.appspot.com/33880043/ which merges the 1.16.2+update | 12:35 |
jam | fwereade: as on call reviewer today, it is my duty to poke you that https://code.launchpad.net/~fwereade/juju-core/prepare-leave-scope/+merge/181065 and https://code.launchpad.net/~fwereade/juju-core/provider-skeleton/+merge/189638 have both been languishing in "needs tweaks" for quite a while. I don't expect you to do anything with them today, but refresh them slightly in your mental TODO list. | 12:57 |
jam | rogpeppe: domas seems to feel he has addressed your requests: https://code.launchpad.net/~tasdomas/juju-core/charm-store-auth/+merge/194130 | 12:58 |
jam | rogpeppe: you also have an LGTM on https://code.launchpad.net/~rogpeppe/juju-core/463-state-ensure-availability/+merge/196177 | 12:58 |
rogpeppe | jam: i'm sure he has. i'll take a look anway | 12:59 |
rogpeppe | anyway | 12:59 |
jam | rogpeppe: well, that was why I was poking yo u:) | 12:59 |
jam | I guess I'm not actually OCR until tomorrow, I'm going to go away now :) | 13:00 |
rogpeppe | jam: thanks | 13:00 |
=== gary_poster|away is now known as gary_poster | ||
jam | natefinch: ping, thinking about "juju unset" and 1.16 compatibility | 13:02 |
natefinch | jam: what about it? | 13:03 |
jam | natefinch: d have "juju unset" in 1.16, right? | 13:03 |
jam | did we have | 13:03 |
rogpeppe | jam: i was really hoping for a glimmer of feedback from fwereade on the addmachine refactoring, but i guess i'll just submit anyway | 13:04 |
jam | well, I guess we did in 1.16.3 | 13:04 |
jam | rogpeppe: well, if you want it, ask for it | 13:04 |
rogpeppe | jam: i did, but i don't want to badger | 13:04 |
jam | I just saw we had gotten to an LGTM no particular pressuer | 13:04 |
jam | rogpeppe: well, as I'm OCR tomorrow, you can wait until I poke you then :) | 13:04 |
* rogpeppe is embarrassed to see he was OCR yesterday and didn't even notice | 13:05 | |
jam | natefinch: anyway, I was trying to figure out how to be compatible when the API doesn't exist, (if we needed to), but if it was there in 1.16.3 clearly we *can* be compat | 13:05 |
jam | rogpeppe: well, we're all a bit distracted | 13:05 |
natefinch | jam: you let me know when you need me to be more than a rubber duck in this conversation ;) | 13:06 |
jam | natefinch: so, something that I'm pretty sure was you. NewServiceSetForClient | 13:06 |
jam | natefinch: is it reasonable in your opinion | 13:06 |
jam | to just fall back to ServiceSet with a warning? | 13:06 |
natefinch | jam: yes | 13:06 |
jam | k | 13:06 |
jam | natefinch: the other option is that we *could* fall back to direct DB | 13:08 |
natefinch | jam: if we just say "hey, we have to fall back to the old version, so this will actually unset those values" it seems reasonable. People lived without being able to set empty strings for a long time. | 13:08 |
jam | which is what 1.16 did, IIRC | 13:08 |
natefinch | jam: up to you. I think maintaining the same functionality would be good, but it depends on how much work that is | 13:10 |
jam | natefinch: yeah, checking now | 13:10 |
natefinch | jam: brb, gotta bring the baby to her mother | 13:10 |
jam | natefinch: np | 13:11 |
natefinch | jam: not sure if you had more for me | 13:38 |
jam | natefinch: not immediately | 13:38 |
jam | I'm off for the day | 13:39 |
jam | fwereade: when you're ready, the 1.16 branch has been pivoted, so we can just land new stuff there that we want in 1.16.4. | 13:49 |
jam | wallyworld_: ^^ | 13:50 |
fwereade | jam, sweet | 13:50 |
jam | fwereade: I also have: lp:~jameinel/juju-core/preparation-for-1.16.5 which will restore your patches to 1.16 with proper ancestry once we're ready for it again | 13:50 |
jam | fwereade: the one patch we need to discuss is sinzui and the "naked -" bug fix | 13:51 |
jam | which essentially is just bumping to a newer version of goyaml, but that was in the changes that aren't in 1.16.3 | 13:51 |
fwereade | jam, ah ok | 13:51 |
sinzui | are you gentlemen discussing backing out the terminate machine changes to 1.16? | 13:53 |
jam | sinzui: already been done | 13:54 |
sinzui | oh goody | 13:54 |
fwereade | sinzui, yes -- we're hoping to rebuild a sane version of history in which we get (1) a 1.16.4 with safe-backup-related changes -- probably to deliver as a hotfix to begin with -- and (2) readd the original1.16.4 bits, with compatibility, as 1.16.5, once we're past all that | 13:55 |
jam | sinzui: so the only one left is bug #1227952 which got reverted along with the rest | 13:55 |
_mup_ | Bug #1227952: juju get give a "panic: index out of range" error <regression> <goyaml:Fix Committed by dave-cheney> <juju-core:Fix Committed by dave-cheney> <juju-core 1.16:Fix Committed by sinzui> <https://launchpad.net/bugs/1227952> | 13:55 |
jam | if it is important and low-risk for NEC | 13:55 |
jam | then we can just land the goyaml bump to dependencies.tsv | 13:56 |
* jam is going to go spend time with my family now | 13:56 | |
fwereade | jam, I am inclined not to include fixes that weren't part of 1.16.3 on general principles, but OTOH I can't come up with a distinct reason to leave that one out | 13:56 |
fwereade | jam, enjoy | 13:56 |
sinzui | jam, fwereade: I can test these dep changes with 1.16 to close this bug in a 1.16 release : https://bugs.launchpad.net/gomaasapi/+bug/1239558 | 13:57 |
_mup_ | Bug #1239558: --upload-tools failure preventing bootstrap completing <golang> <sync-tools> <upload-tools> <Go MAAS API Library:Fix Committed by wallyworld> <Go OpenStack Exchange:Fix Committed by wallyworld> <Go Windows Azure Client Library:Fix Committed by wallyworld> <juju-core:Fix Committed by wallyworld> <https://launchpad.net/bugs/1239558> | 13:57 |
sinzui | ^ actually, I did test those deps and know the suite is fine. I can ask CI to walk though its tests | 13:58 |
dimitern | rvba, ping | 13:59 |
fwereade | sinzui, so is that affecting 1.16.3? | 13:59 |
dimitern | fwereade, ping | 14:00 |
rvba | dimitern: hi | 14:00 |
fwereade | dimitern, pong | 14:00 |
dimitern | rvba, hey, so we're trying to help the cts guys here how to find the agent_name value to pass to maascli node aquire | 14:00 |
sinzui | fwereade, I believe the issue is does. I rejected the bug for 1.16.2 because we didn't have a release policy for those projects | 14:00 |
dimitern | rvba, if you can join the hangout (sent a link as a direct message) | 14:02 |
rvba | dimitern: well, the agent_name is generated by Juju and stored in the config file IIRC | 14:03 |
dimitern | rvba, do you remember the key name in the .jenv file? is it "agent_name" or something else? | 14:03 |
rvba | dimitern: it's 'environment-uuid', see provider/maas/config.go | 14:04 |
dimitern | rvba, ok, thanks | 14:04 |
rvba | np | 14:04 |
rvba | dimitern: err, it's been renamed to 'maas-agent-name' | 14:05 |
rvba | (I forgot to update my copy of juju-core) | 14:05 |
dimitern | rvba, ah, ok | 14:05 |
=== dstroppa_ is now known as dstroppa | ||
natefinch | dimitern: sorry, my connection is crappy today for some reason. Ping me if you need me, but it sounds like he's just having MaaS issues | 15:21 |
dimitern | natefinch, sure, np | 15:22 |
rogpeppe | fwereade: https://codereview.appspot.com/34050043/ | 15:39 |
fwereade | rogpeppe, LGTM, just minor comments | 15:48 |
rogpeppe | fwereade: thanks | 15:48 |
rogpeppe | fwereade: if you have a moment at some point, i'd still like to have your views on this, even though I already have a LGTM. https://code.launchpad.net/~rogpeppe/juju-core/463-state-ensure-availability/+merge/196177 | 15:58 |
rogpeppe | fwereade: also, when you say "apply the safe mode patch against 1.16", which 1.16 version should i do that against. when are we going to pivot? | 15:59 |
fwereade | rogpeppe, jam has backed out the original 1.16.4 changes | 15:59 |
rogpeppe | fwereade: ah, cool | 15:59 |
fwereade | rogpeppe, we can land directly on 1.16 | 15:59 |
fwereade | rogpeppe, thanks for the reminder, I will try to finish that branch off this half-hour | 16:02 |
mgz | so, the lp:juju-core/1.16 branch is where stuff for this needs landing now, right? | 16:07 |
natefinch | ahh... reboot the router, and wifi is so much better... | 16:08 |
sinzui | Hi lads there is confusion about some bug tags. I am going to rename openstack to openstack-provider. What should I rename ssh-provider too? | 16:13 |
fwereade | rogpeppe, bunch of trivials, but basically awesome, LGTM, TYVM | 16:14 |
fwereade | mgz, yeah | 16:14 |
rogpeppe | fwereade: thanks | 16:14 |
sinzui | fwereade, what are we calling the null/ssh provider now? | 16:20 |
fwereade | sinzui, "manual" | 16:27 |
sinzui | thank you | 16:27 |
mgz | okay, the juju-update-bootstrap command is on the 1.16 branch, and I've updated the instructions to either/or | 16:50 |
dimitern | mgz, are you sure StateInfo() will give you the correct public address, and not localhost? | 16:50 |
mgz | in this context? yeah, pretty sure | 16:51 |
dimitern | mgz, I was having issues like that in the API - the server knows only localhost for both state/api addresses | 16:51 |
mgz | the alternative is to open an api connection and get the address through that | 16:51 |
dimitern | mgz, ok then | 16:51 |
mgz | which is newer/better code | 16:51 |
dimitern | mgz, if you tested it gets the correct address | 16:51 |
dimitern | mgz, i just remembered this slight issue i had with it before | 16:52 |
mgz | I got it printing the right thing in all my testing | 16:52 |
mgz | I've not played with their maas setup however | 16:52 |
mgz | it should be pretty obvious if it's doing the wrong thing though, everything would break | 16:52 |
dimitern | yeah, true | 16:53 |
dimitern | natefinch, ping | 16:53 |
mgz | hm, the fallback instrctions here have the script with $ADDR but no actual setting of that | 16:53 |
mgz | oh, I guess it does, it's not exported... which is fine | 16:53 |
natefinch | dimitern: here | 16:59 |
dimitern | natefinch, will you be around for support on the call? | 17:00 |
natefinch | dimitern: yep | 17:00 |
dimitern | natefinch, cheers | 17:00 |
jam | sinzui: we really need to give you a "hovey-bot@canonical.com" account, so when you run your script I can filter out stuff I should really read (we have a regression) from stuff I can just ignore | 17:27 |
sinzui | jam, I think we want a juju-qa-bot | 17:28 |
jam | sinzui: that would be true if you didn't do it on LP before Juju, so *clearly* we need a hovey-bot first :) | 17:28 |
=== BradCrittenden is now known as bac | ||
davecheney | sinzui: ping | 21:30 |
sinzui | Hi davecheney | 21:30 |
davecheney | sinzui: may i have 10 minutes of your time for a hangout call ? | 21:31 |
sinzui | yes | 21:31 |
davecheney | let me see if I can computer and make this work | 21:31 |
davecheney | sinzui: https://plus.google.com/hangouts/_/7acpjtoo0k2vn3aaa34vra9k2g?authuser=1&hl=en | 21:33 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!