[06:26] <dimitern> morning all
[06:27] <dimitern> (just in case it's not just me today :)
[06:27] <natefinch> dimitern: morning.  I'm here, but only just for a minute or two before bed
[06:28] <dimitern> natefinch, hey, btw do you need another review on that ensure availability API branch?
[06:28] <natefinch> dimitern: yes please.  Just pushed a bunch of changes.  There's a few comments from Andrew that I just can't get to before I fall over tonight, but mostly I think it's fine.
[06:29] <natefinch> http://reviews.vapour.ws/r/1299/
[06:29] <dimitern> natefinch, ok, np - looking
[06:29] <natefinch> way past bedtime for me.  Good night
[06:30] <dimitern> good night natefinch !
[06:35] <dimitern> jam, I guess you're not here today
[06:42] <jam> dimitern: I'm going to be a bit late for our 1:1, I have to take my son to a friends house. I'll ping you when I get back.
[06:43] <dimitern> jam, sure, np
[09:18]  * dimitern steps out for ~30m
[09:22] <perrito666> morning
[09:33] <jam> dimitern: sorry I forgot to ping you back. Do you have any pressing items to discuss? How are you feeling about prep for next week?
[09:54]  * dimitern is back
[09:54] <dimitern> jam, hey
[09:55] <dimitern> jam, perhaps we can have a quick chat now if you're available?
[09:55] <dimitern> morning perrito666
[09:59] <dimitern> perrito666, aren't you off today?
[10:05] <perrito666> dimitern: nope, I have been on vacations for two weeks
[10:05] <perrito666> but everyone else seems to be off
[10:06] <dimitern> perrito666, yeah :)
[14:33] <natefinch> wwitzel3: standup?
[14:55] <natefinch> dimitern: about utils_RandomPassword ... I intentionally use an unusual format w/ the underscore, since it's replacing utils.RandomPassword... I find it makes it easy to tell what function is being replaced by this variable when you're just looking at the code.  otherwise I could just call it randomPassword()... but I like having an easy way to tell what it's replacing without having to go to definition on it.
[14:56] <dimitern> natefinch, well what's wrong with utilsRandomPassword ?:)
[14:57] <natefinch> dimitern: it's still not as obvious that the first part is part of the package name... the underscore really separates it out, and it still looks a lot like the original namespaced form: var utils_RandomPassword = utils.RandomPassword
[14:58] <natefinch> dimitern: but if people really don't like it, it's not a huge deal.
[14:59] <dimitern> natefinch, I still rather not have underscores in code, but it that's the only thing remaining and others are ok with it, I'll accept that
[15:00] <dimitern> s/but it/but if/
[15:00] <natefinch> dimitern: heh... there's a ton of other stuff remaining :)
[15:23] <natefinch> dimitern:  I'm not sure what you meant when you said there was inconsistent indenting with the txn.Ops.  I just copied the other three functions that were formatted exactly like mine (as far as I can tell).
[15:31] <dimitern> natefinch, in a call, will get back to you
[15:43] <dimitern> natefinch, ah, so I though my comment there made it clear,
[15:44] <natefinch> dimitern: not to me, at least :)
[15:45] <dimitern> natefinch, I think you either use []txn.Op{{\nC:XYZ,\n,...,\n}, {\n,C:UVT,\n...,\n}} or []txn.Op{\n{\nC:XYZ,\n...,\n},\n{\nC:UVT,\n...,\n}\n,}\n
[15:46] <dimitern> natefinch, does it make more sense?
[15:47] <natefinch> dimitern: I think I'm doing that first one
[15:48] <natefinch> or are you asking me to un-indent the stuff inside?
[15:48] <dimitern> natefinch, sorry, let me have a look at the review again
[15:50] <dimitern> natefinch, ugh, I've seen it now - it's fine, so please ignore my comment and I'll drop the issue
[15:50] <natefinch> ok :)  good, I thought I was going crazy or something, which is entirely possible running on 4 hours sleep :)
[15:50] <dimitern> :)
[15:52]  * dimitern reached EOD
[15:59] <cherylj> ericsnow: you around?
[15:59] <ericsnow> cherylj: yep
[16:00] <cherylj> ericsnow: I'm working on bug 1439447 that we talked about last Friday and I need to get some advice...  natefinch had an idea that we could specify a new config option to indicate whether or not to use the proxy when contacting the bootstrap node
[16:00] <mup> Bug #1439447: tools download in cloud-init should not go through http[s]_proxy <cloud-installer> <landscape> <juju-core:In Progress by cherylj> <juju-core 1.23:In Progress by cherylj> <https://launchpad.net/bugs/1439447>
[16:01] <cherylj> ericsnow: and I've been digging through things for a while and I'm wondering if this flag should be part of proxy.Settings since it does relate to how we use the proxy...
[16:02] <cherylj> proxy.Settings is part of juju/utils
[16:02] <cherylj> and it would break backwards compatibility, but we branched utils because of the license changes I made anyway...
[16:03] <ericsnow> cherylj: I thought about that too
[16:03] <ericsnow> cherylj: no_proxy already covers this sort of
[16:04] <ericsnow> cherylj: if it supported IP ranges it might be good enough
[16:04] <cherylj> ericsnow: yeah, and another thought I had was to just add the state server IPs to the no_proxy list when this new flag is set
[16:04] <ericsnow> cherylj: yep
[16:04] <cherylj> ericsnow: then if we add new state servers, or remove them, then we need to prune the list
[16:05] <cherylj> ericsnow: and if the flag is unset, do we remove them?  What if the user had manually added them and we just don't know?
[16:05] <ericsnow> cherylj: proxy.Settings could also grow a new field like "ProxyPrivateAddresses"
[16:06] <cherylj> ericsnow: yeah, that was my thought
[16:06] <ericsnow> with that setting we wouldn't need to add addresses to no_proxy
[16:06] <cherylj> ericsnow: are there other places (other than this curl command) that we would need to change to check this?  I'm not sure if there are other places communication is happening with the state servers through the proxy
[16:08] <ericsnow> cherylj: looks like several packages use proxy.Settings: utils/apt/apt.go and a few places in juju core
[16:11] <cherylj> ericsnow: okay, I'll take a look at those places
[16:12] <cherylj> ericsnow: so you think adding the new field to proxy.Settings sounds good?  I think it will be the clearest way to add this config option.
[16:17] <ericsnow> cherylj: ideally the proxy.Settings methods would handle that new field properly by adding the appropriate IP address
[16:17] <ericsnow> cherylj: however, without IP range support that probably isn't doable
[16:18] <cherylj> cherylj: I had also thought that in environs/config/config.go where we return a new proxy.Settings, we could check there if this flag was set, and add the IP addresses to the no_proxy list before returning
[16:19] <ericsnow> cherylj: definitely
[16:19] <cherylj> I'm not sure if that list is used in any permanent way
[16:19] <cherylj> or if ProxySettings() is called every time we need the proxy info
[16:20] <ericsnow> cherylj: it would be nice, though, if we had an implicit solution; that way every use of proxy.Settings doesn't need to handle the new field
[16:21] <cherylj> yeah
[16:28] <ericsnow> cherylj: for now we could just add the IP address of each state server on a private IP address to Settings.NoProxy
[16:29] <ericsnow> cherylj: or prepend "HTTP_PROXY= HTTPS_PROXY= " to that one curl command if the IP in the curl command is private
[16:30] <ericsnow> cherylj: in the interest of solving the 1.23 bug
[16:30] <ericsnow> cherylj: then open a bug against 1.24 for the better solution
[16:31] <cherylj> ericsnow: yeah, I wasn't sure what the timeframe looked like for this bug.
[16:32] <ericsnow> cherylj: we could probably argue for deferring it to 1.23.1, but it would be nice to have a fix for 1.23.0. :)
[16:32] <cherylj> ericsnow: it's easy to determine if we're not the bootstrap node, and we could prepend the "HTTP_PROXY= ..." in that case.  Do I need to check if the IP is also private?
[16:34] <ericsnow> cherylj: it depends on whether or not cloudinit for a new HA instance pulls from another state server
[16:36] <ericsnow> cherylj: I'm guessing we can simply key off of w.mcfg.Bootstrap
[16:36] <cherylj> ericsnow: that's what I was thinking
[16:36] <ericsnow> cherylj: we would need to check whether or not that is set for HA instances
[16:41] <ericsnow> natefinch: do you know if MachineConfig.Bootstrap (as used by cloudinit) is set for HA instances?
[16:42] <natefinch> ericsnow: I do not.   I'd have to go look it up
[16:42] <ericsnow> natefinch: no worries
[16:44] <natefinch> gotta run for a bit, back in an hour-ish.
[16:50] <cherylj> ericsnow: I don't think that it is set.  Just did a little test with ensure-availability and it wasn't...
[16:50] <cherylj> I gotta take a quick lunch break.  brb
[16:50] <ericsnow> cherylj: k
[18:11] <perrito666> brb
[18:12] <natefinch> >     >     >> >>> On Apr 5, 2015, at 11:28 AM, Someone with an annoying plaintext email program replied....
[18:13] <wwitzel3> lol
[18:52] <cherylj> ericsnow: do you know how can I get the script output like what's included in that bug to see the actual commands being run?
[18:53] <ericsnow> cherylj: look for cloudinit.log on the new instance
[18:54] <ericsnow> cherylj: I don't recall exactly where the cloudinit script is written
[18:54] <cherylj> ok
[19:20]  * perrito666 had to go replace his fron't door lock after he had to dismantle the door to get out of the house
[19:21] <perrito666> mondays...
[19:22] <mup> Bug #1433577 changed: Vivid unit tests need to pass <ci> <test-failure> <vivid> <juju-core:Fix Released by ericsnowcurrently> <juju-core 1.23:Fix Released by ericsnowcurrently> <https://launchpad.net/bugs/1433577>
[19:22] <mup> Bug #1435860 changed: certSuite.TestNewDefaultServer failure <ci> <regression> <test-failure> <unit-tests> <juju-core:Fix Released by ericsnowcurrently> <juju-core 1.22:Fix Released by ericsnowcurrently> <juju-core 1.23:Fix Released by ericsnowcurrently> <https://launchpad.net/bugs/1435860>
[19:22] <mup> Bug #1435974 changed: Copyright information is not available for some files <juju-core:Fix Released by cherylj> <juju-core 1.22:Fix Released by cherylj> <juju-core 1.23:Fix Released by cherylj> <https://launchpad.net/bugs/1435974>
[19:22] <mup> Bug #1437040 changed: unit test failure: TestNewDefaultServer.N40_github_com_juju_juju_cert_test.certSuite <ci> <unit-tests> <juju-core:Fix Released by ericsnowcurrently>
[19:22] <mup> <juju-core 1.22:Fix Released by ericsnowcurrently> <juju-core 1.23:Fix Released by ericsnowcurrently> <https://launchpad.net/bugs/1437040>
[20:44] <natefinch> thumper: got a minute?
[20:44] <thumper> natefinch: otp with alexisb
[20:44] <natefinch> thumper: k
[21:40] <mup> Bug #1439880 changed: Container's interfaces are all on private networks instead of host's eth0 network <lxc> <maas-provider> <network> <oil> <juju-core:Triaged by dimitern> <https://launchpad.net/bugs/1439880>
[21:47] <thumper> natefinch: ok, here now
[21:47] <thumper> natefinch: do I need coffee?
[21:51]  * thumper goe to get breakfast
[23:05] <axw_> katco: Internet just dropped out, I guess because of the weather. Just FYI, in case I don't turn up at the standup
[23:16] <katco> axw: k ty
[23:19] <katco> axw: since it's just us, we can probably just coordinate over irc. i have a few questions for you, but they can wait until later (after my dinner)
[23:22] <katco> axw: i'll just catch up with you over irc after dinner
[23:29] <axw> katco: :/  sorry
[23:30] <axw> katco: we can try again if you like. otherwise may be best to email me
[23:31] <thumper> axw, katco: seems like your team is as small as mine this week :)
[23:32] <axw> cosy
[23:37] <thumper> axw: makes for quick standup calls :)
[23:38]  * thumper relocates for better design thinking space
[23:59] <katco> thumper: lol yep
[23:59] <katco> axw: i'll send you an email