=== menn0 is now known as menn0-afk
davecheneywaigani: i'm looking at your style guide PR00:22
davecheneyhave you pushed the latest version ?00:22
waiganidavecheney: give me 5 - just about to push latest00:23
waiganidavecheney: just writing an example00:23
davecheneywaigani: ok, protip, don't comment on the review saying "i've done this", unless you've pushed the code00:24
davecheneyit's confusing00:24
* thumper agrees00:24
davecheneysorry, i know this is yet another github fail00:25
davecheneyand it requires you to juggle responding to comments and pushing branches concurrently00:26
davecheneybut it is what it is00:26
davecheneythumper: https://github.com/juju/juju/pull/244/files00:43
davecheneyi need advice on how to handle this00:43
* thumper loks00:44
thumperdavecheney: kick it up00:46
thumperdavecheney: email alexisb and mramm00:46
thumperthey will say what is good or not00:46
thumperI'm not sure what our agreement is00:46
davecheneythumper: will do00:46
thumperschool holidays are cramping my style00:51
* thumper will be in and out for the afternoon00:51
=== psivaa_ is now known as psivaa-off
davecheneythumper: consider yourself nagged about a policy of using the juju/errors package00:56
thumperdavecheney: I just pass that nagging on to fwereade00:57
davecheneythumper: then everything is going swimmingly01:02
davecheneyhey, being a reviewer does suck01:04
davecheneyyou cant' edit the title or the labels01:04
gsamfiradavecheney: but its very much appreciated :)01:09
waiganidavecheney: sorry for the confusion. It's up now.01:11
davecheneywaigani: oh man01:11
davecheneyyou just undid all the stuff I said to drop01:12
davecheneyshit man01:12
davecheneyi wanna land this today01:12
davecheneyi'm going to write some comments01:12
davecheneyyou're going to do what i write01:12
davecheneythen we're going to merge this01:12
davecheneyok ?01:12
waiganidavecheney: ok01:12
=== menn0-afk is now known as menn0
=== thumper is now known as thumper-afk
menn0davecheney: I wish you hadn't hit merge on my PR... I was about to push an extra unit test (as per my comment)01:26
davecheneymenn0: sorry01:31
menn0davecheney: it just means I have to do another tiny PR with the test change01:31
davecheneymenn0: ok01:32
menn0davecheney: https://github.com/juju/juju/pull/27301:32
davecheneymenn0: what does \d{14} do ?01:35
davecheneyi'm not used to that form01:35
menn0"\d" == "[0-9]"01:35
jrwren14 digits01:35
menn0{14} means "exactly 14 of the last thing"01:35
jrwren\d will match utf8 digits as well, in most regex engines01:35
menn0you can also do {A,B} which means between A and B of the last thing01:36
davecheneyyeah, thati's what I thought01:37
davecheneybut how does \d{14} assert that it's 7 bit ascii01:37
menn0davecheney: the fact that we're checking the entire string exactly with a regex that only allows characters that are 7-bit ascii01:38
menn0I hadn't heard of the UTF-8 digits thing01:38
menn0it's not mentioned in the docs: https://code.google.com/p/re2/wiki/Syntax01:39
davecheneymenn0: change looks good01:39
davecheneybut the comment doesn't match the description01:39
davecheneywould it be simpler to say, the backup file must be juju-backup_NNNNNNNNNNN01:40
davecheneyor something like that01:40
menn0davecheney: yeah, I did consider that the comment could be confusing ... I just want to make sure that future changes ensure that the filename remains 7-bit ascii01:40
menn0maybe I'll add a more explicit test...01:41
davecheneymenn0: you keep saying 7 bit ascii01:41
davecheneybut don't you mean juju-backup_14decimals ?01:41
menn0davecheney: juju-backup_14decimals happens to also be 7-bit ASCII :)01:41
menn0that's what I'm getting at01:41
menn0obviously it's not clear enough01:41
menn0let me rework the PR a little01:42
davecheneythere are 117 other valid ascii symbols that also fit in the 7 bit range01:42
jrwrenhttps://code.google.com/p/re2/wiki/Syntax has an [:ascii:] character class for that.01:46
davecheneywaigani: did you push that PR again ?01:46
jrwrene.g.  \d matches 101:48
jrwren[0-9] does not match 101:48
=== Guest41889 is now known as wallyworld
wallyworlddavecheney: bug 1336891 - can that be fixed by changing our code like in other cases so as not to have to have the backported compiler fix?01:53
_mup_Bug #1336891:  intermittent panic: runtime error: invalid memory address <gccgo> <ppc64el> <juju-core:Triaged> <juju-core 1.20:Triaged> <https://launchpad.net/bugs/1336891>01:53
menn0davecheney: look again please: https://github.com/juju/juju/pull/27301:55
davecheneywallyworld: no01:55
davecheneywallyworld: we need the compiler fix to land01:55
menn0davecheney: I've removed mention of 7-bit ASCII01:55
davecheneymenn0: LGMT01:55
menn0davecheney: the test now checks what is actually important and only accepts a minimal set of characters01:55
davecheneythat amkes much more sense01:56
wallyworlddavecheney: ok, i was asked to check because john "fixed" other issues like that by tweaking our code not to tickle the bug01:56
davecheneywallyworld: matt sent me the dmesg output01:56
bigjoolswhich one of you guys is working on LXCs?01:56
davecheneythis is the runtime bug we were hitting several months ago01:56
davecheneybigjools: that's thumper-afk01:56
menn0davecheney: cheers. merging now.01:56
wallyworlddavecheney: ok, thanks, will update the bug01:56
davecheneywallyworld: are we talking about different bugs ?01:56
wallyworlddavecheney: the bug i typed above01:57
davecheneyi just put an update on the bug i think we're talking about01:57
davecheneywallyworld: same bug01:57
davecheneysee comment #901:57
wallyworlddavecheney: ah sorry yes, my browser copy was out of date01:57
davecheney:sad chrome01:57
menn0jrwren: are you sure \d matches unicode digits? http://play.golang.org/p/fyInDC26-W02:01
menn0jrwren: you've got me curious :)02:02
=== thumper-afk is now known as thumper
thumperbigjools: whazzup?02:33
bigjoolsthumper: there's a new API call coming in the next release that allows you to get a static IP02:33
bigjoolsmaas release, I mean02:33
bigjoolsso I figured you might want that for containers02:34
thumperbigjools: nice, but that is networking, so dimeter :)02:34
thumperbut handy, thanks02:34
* bigjools watches hot potato flung to Malta02:35
* thumper blows on his hands02:35
sebas5384did somebody make some benchmarking into the lxc conatiners created by juju-local ?02:35
thumpersebas5384: at some time... yes02:35
sebas5384thumper: hey o/02:35
sebas5384because i'm running php using a drupal charm02:36
sebas5384and I did a benchmarking script02:36
sebas5384better say, i execute a benchmarking script in php02:37
sebas5384and it seems that its at least 4 or 5 seconds more than the host02:37
thumperwhich version?02:38
thumperversion of juju02:38
sebas5384same version in the two machines02:38
sebas5384the last stable one02:38
sebas5384that one02:41
sebas5384its strange because there are no constraints02:42
waiganidavecheney: pushed again now02:48
sebas5384thumper: maybe is something about filesystem ...02:49
davecheneywaigani: ta02:49
davecheneyfood, then review02:49
thumpersebas5384: the lxc containers are being treated exactly like machines, they do apt-get update/upgrade at the start (for that version)02:50
thumpersebas5384: not surprising they are the same02:50
sebas5384well yeah, but why then php is slower into the lxc containers :/02:51
jcw4davecheney, actionWatcher refactor related to our discussion yesterday : https://github.com/juju/juju/pull/27503:03
davecheneyjcw4: ta03:13
jcw4davecheney: pushing one quick change in 15 seconds03:13
=== vladk|offline is now known as vladk
jcw4davecheney: okay; updated03:14
=== jcw4 is now known as jcw4|afx
=== jcw4|afx is now known as jcw4|afk
davecheneyjcw4|afk: sorry mate, thats a -1 from me03:38
stokachudo i pass a 0 to ToMachineSpec in the ServiceDeploy api in order to have it automatically deploy to a new machine?03:38
stokachuor do i need to add-machine and deploy to that machine via the api03:38
davecheneystokachu: well, i'd start with ""03:40
davecheneyand i'd open a bug that the documetation on that method is crap03:40
stokachuor will setting NumUnits to 1 make it deploy to a machine whether allocated or not?03:40
stokachuok lemme try these options first03:40
stokachuthen ill create a bug03:40
davecheneystokachu: you cannot specify a number of units less than 103:41
stokachuso right now if i set the numunits to 1 the ServiceDeploy will just deploy the charm without assigning it to a machine03:42
stokachunew or one allocated (local provider)03:42
davecheneystokachu: sorry, i can't help you03:42
davecheneyi tried reading the documentatoin but is it crap03:42
stokachulol it is kind of difficult to use the api with the docs :(03:43
stokachuon a brighter note ive got a python3 juju client working pretty well https://github.com/Ubuntu-Solutions-Engineering/macumba03:44
stokachugonna use this for the openstack installer03:45
stokachudavecheney, hah, setting ToMachineSpec = "" worked :|03:47
davecheneypossibly if you don't supply that tag in the json message, you'd get the same effect03:48
stokachui think i tested that and it didnt work, lemme try again03:49
stokachuah my mistake that does work03:50
stokachudavecheney, ty03:50
davecheneyif cannot afford to provide ToMachineSpec then one will be provided for you.03:51
=== jcw4|afk is now known as jcw4
jcw4davecheney: it sounds like my description of the solution is the problem right?03:59
jcw4davecheney: let me try again and see if it makes more sense03:59
waiganidavecheney: https://github.com/juju/juju/pull/267 has merge conflicts04:09
waiganiaccording to gh04:09
davecheneywaigani: ta04:11
davecheneyi'll repropose04:11
davecheneysomeone must have landed something over night04:11
davecheneynot surprising04:11
jcw4davecheney: I updated the PR message, should I update the actual git commit message with the same details too?04:24
davecheneyjcw4: nah, nobody reads those04:25
davecheneywell, maybe not nobody04:25
jcw4davecheney: lol04:25
davecheneybut it's fn hard to get access to them04:25
davecheneyfrom the PR screen04:25
davecheneyjcw4: good PR message04:28
jcw4davecheney: tx04:28
davecheneymy repl set tests are getting worse and worse04:30
davecheney[LOG] 0:37.995 DEBUG juju.testing tls.Dial( failed with dial tcp connection refused04:30
davecheney    s.dialAndTestInitiate(c, root, getAddr(root))04:30
davecheney    c.Assert(err, gc.IsNil)04:30
davecheney... value *errors.errorString = &errors.errorString{s:"cannot get replica set status: can't get local.system.replset config from self or any seed (EMPTYCONFIG)"} ("cannot04:30
davecheney get replica set status: can't get local.system.replset config from self or any seed (EMPTYCONFIG)")04:30
davecheneynever seen that error before04:30
davecheneyi assume it's a race setting up the repl set04:30
=== vladk is now known as vladk|offline
=== vladk|offline is now known as vladk
jcw4tx davecheney ; I'll address the comments that I think are correct.  If you'd like to review my responses on the others and let me know if they make sense to you?04:39
davecheneyjcw4: ok04:40
=== vladk is now known as vladk|offline
=== jcw4 is now known as jcw4|ZZzzzz
menn0davecheney: Backups in environment storage: https://github.com/juju/juju/pull/27805:04
menn0davecheney: reviewing your PRs now05:05
menn0davecheney: LGTM for #26505:15
davecheneymenn0: ta05:17
davecheneymenn0: i added a lot of snotty comments on your PR05:17
davecheneyin general it looks good05:17
menn0davecheney: thanks for looking. I'm just responding now.05:19
menn0davecheney: I've responded to your snotty comments. Some answers from you are required.05:29
menn0davecheney: :-p05:30
menn0davecheney: thanks. any thoughts on the first comment?05:43
davecheneymenn0: waiting for you to push a new revision05:55
davecheneyalso, i just fixed a bug in my PR05:56
davecheneywhen someone calls05:56
davecheneyback.Backup("some password", "", ...)05:56
davecheneythat is user is ""05:56
bigjoolsdimitern: around?05:58
=== uru-afk is now known as urulama
dimiternbigjools, yeah, barely (need coffee :)06:01
bigjoolsdimitern: well get some coffee and then I can show you Cool New Stuff™06:01
menn0davecheney: I see the fix in your PR. Should that mongo connection field even be called Tag?06:03
menn0it's all about the user that's connection to mongo06:03
menn0is it really a tag?06:03
bigjoolsdimitern: dogwalking, speak in a bit06:05
davecheneymenn0: it's not a user06:06
davecheneyyou say user, you think human06:07
davecheneybut we say tag06:07
davecheneyas in 'some actor in the juju universe'06:07
davecheneynow, some tagable things don't make sense to have credentials to talk to mongo06:07
davecheneylike networks and environments06:07
davecheneybut machines and units do06:07
davecheneymenn0: i agree this is apeshit crazy06:08
menn0davecheney: I get what tags are about... I guess what I'm missing is what our mongo DB usernames look like06:08
davecheneywhich is why i'm trying to fix it06:08
davecheney^ these are mongo usernames06:09
menn0right. so we are using tag strings as mongo username06:09
menn0I didn't realise that.06:09
davecheneyit gets so much more horrible06:09
davecheney"" < this is the admin user06:09
menn0your change makes sense now06:09
menn0davecheney: I've pushed updated revs for my PR06:09
* menn0 nods.06:09
menn0davecheney: related topic: does it annoy you that it's Backup(password, tag, ...) and not Backup(tag, password, ...)?06:11
davecheneymenn0: no06:11
davecheneyit annoys me that it creats a temporary file on disk when this shoud return a io.Reader06:11
davecheneybut that is for another day06:11
menn0davecheney: shouldn't it /take/ an io.Writer instead of returning a io.Reader? I don't see how it can return an io.Reader without either holding the entire backup in RAM or using a temp file.06:13
dimiternbigjools, sure06:14
davecheneymenn0: you can do it either way06:14
davecheneyyou have goroutines, remember06:14
davecheneybut you are correct, passing the http response write to backup would be simpler06:15
davecheneyi was thinking some odball io.Pipe thingy06:15
menn0davecheney: that PR has an extra rev for a one line fix to the existing code that I noticed btw06:15
=== vladk|offline is now known as vladk
=== vladk is now known as vladk|offline
=== uru_ is now known as urulama
bigjoolsdimitern: so got 5 mins09:26
dimiternbigjools, sure09:26
bigjoolsdimitern: I added an API in maas 1.6 so you can claim an IP09:26
bigjoolsapparently this is useful to you?09:27
dimiternbigjools, great! so we can reserve container addresses explicitly09:27
bigjoolswill point you at a doc, one sec09:28
dimiternbigjools, how does it work? does the api call return the address it allocated?09:28
bigjoolsdimitern: so api calls are:09:32
bigjoolsGET /api/1.0/ipaddresses/09:32
bigjoolsto list ones you reserved09:32
bigjoolsPOST /api/1.0/ipaddresses/ takes op=release or op=reserve09:32
bigjoolsand I shall dig up more concrete representations of the data you get back in a moment when I can find the damn thing09:33
dimiternbigjools, thanks09:33
dimiternbigjools, is this advertised via the capabilities/version api call?09:33
bigjoolsdimitern: good question09:34
bigjoolsdimitern: oh HA, I already sent you an email on the 2nd about this09:37
bigjoolswith the API description09:38
bigjoolsdimitern: the latest beta package for 1.6 is in ppa:maas-maintainers/testing09:38
dimiternbigjools, sorry, it seems I've missed it09:38
bigjoolssubject "proposed release 1.6.0"09:38
bigjoolsI'll forward it again09:39
dimiternbigjools, i'll look into the api docs, cheers09:39
bigjoolswell there's no docs yet, it's just an email09:39
* bigjools EODs09:40
rogpeppedimitern, wallyworld, jam, davecheney: any chance of a review of https://github.com/juju/charm/pull/15 please?09:44
TheMuerogpeppe: looking09:46
rogpeppeTheMue: thanks09:47
TheMuerogpeppe: nice one, +109:52
rogpeppeTheMue: thanks!09:52
* TheMue uses the time for reviews when his tests run09:53
TheMuejam: ping09:53
jamTheMue: yes?09:53
TheMuejam: I’m still at the bootstrap/endpoint issue, because I came a step nearer09:54
TheMuejam: but I don’t find a solution, maybe you’ve got an idea09:54
jamTheMue: so we've had a second report of it, so it does sound like something that is a reasonably serious issue in 1.20. It might be associated with some people saying it took *minutes* for mongo to initialize09:54
TheMuejam: I can simulate it, when I consume some time between ensureMongo and openState09:55
jamwallyworld and axw think that mongo delay is because of the oplog, which they had a patch for09:55
jamTheMue: so that triggers it?09:55
TheMuejam: yes, tried other cases, but no error. but when adding just some seconds between those too steps I get the same error on local09:55
jamTheMue: so the fix (IMO) is that we shouldn't let things like the API actually respond to requests until we have a backing DB09:56
jamso something like the "afterUpgradeStep" stuff does today.09:56
jamthough you might run it by axw or wallyworld since I think they've touched the code most recently09:56
TheMuejam: thought in this directy too, yes09:57
TheMuejam: ok, will contact them09:57
jamTheMue: btw, congrats to Germany... I don't know how they smashed it so hard, but kudos09:59
TheMuejam: thanks, we had a lot of fun yesterday evening, and have been astonished on our own10:01
TheMuejam: I expected a close fight, maybe 1:0. but 7:1, wow10:01
=== vladk|offline is now known as vladk
TheMueWow, need a few seconds, a weird JS almost killed my machine.10:45
dimiternvladk, jam, standup?10:48
=== urulama is now known as uru-food
TheMuejam: never done this on LP before. how do I merge https://code.launchpad.net/~niedbalski/golxc/fix-1329049/+merge/225584 and are any steps regarding our dependencies required?12:14
jamTheMue: I believe the process for golxc (not managed by a bot) is that you: "bzr co lp:golxc; bzr merge $THEMUESCODE; go test ./...; bzr commit -m "merge $THEMUES most awesome of patches"12:15
jamTheMue: actually, the first step you need is to create a kanban card :)12:16
TheMuejam: ok, ok, will do this now :D12:17
TheMuejam: so, can continue, card exists 8)12:20
vladkdimitern, jam: interfaces file in MaaS: https://docs.google.com/a/canonical.com/document/d/18okVil9JGJgGxNlig5Gk8wwtNDSoXtT-Qta6B7gxTI8/edit#12:21
dimiternvladk, thanks, will take a look12:22
=== uru-food is now known as urulama
=== jcw4|ZZzzzz is now known as jcw4
jamTheMue: there is a slight "it depends" which matters on how you have Bazaar set up locally.12:23
jambut basically, you want to turn your $GOPATH/src/launchpad.net/golxc back into a pristine trunk12:23
jammerge your code into it12:23
jamrun the test suite,12:23
jamcommit it, and then push it back to launchpad12:23
jamvladk: thx12:24
Egoistis it possible that juju don't use containers on openstack?12:52
TheMuejam: thx for your help, e’thing worked fine as wanted12:53
jamTheMue: great13:10
jamEgoist: "is it possible" you don't have to use containers with openstack, though openstack itself is doing virtualization, right?13:10
jam(IIRC, LXC is one of the possible ways that openstack creates virtual machines.)13:11
axwkatco: #131947513:19
_mup_Bug #1319475: Juju should support new signing format <ec2-provider> <juju-core:Triaged> <https://launchpad.net/bugs/1319475>13:19
=== urulama is now known as uru-afk
=== viperZ28__ is now known as viperZ28_
perrito666I just notices that I didnt see a few of natefinch comments13:51
alexisbdimitern, jam : I sent you both a note on IPv6 I need a response asap14:08
dimiternalexisb, sure14:10
alexisbthanks dimitern14:11
perrito666ok natefinch axw I think I addressed all the comments14:14
natefinchperrito666: link me again?14:16
alexisbwallyworld, ping14:16
perrito666natefinch: ln -s natefinch ./natefinch14:16
perrito666natefinch: anything else?14:16
wallyworldalexisb: i think i know what you're going to ping me about14:17
alexisbHi there wallyworld - well done to the team for whipping through a bunch of bugs14:17
alexisbon this bug:14:17
_mup_Bug #1338179: juju 1.20.x slow bootstrap <bootstrap> <cloud-installer> <local-provider> <status> <juju-core:Fix Committed by axwalk> <juju-core 1.20:Fix Committed by axwalk> <https://launchpad.net/bugs/1338179>14:17
alexisbwas there an actual fix?14:17
alexisbgiven we couldnt repo?14:17
wallyworldalexisb: yes, the size of the mongo oplog file was reduced14:18
axw(for local only)14:18
alexisbare we sure that addresses the bug that was reported?14:18
wallyworldalexisb: it was deduced that that was the most likely cause of the extra 2 minutes of bootstrap time14:18
wallyworldalexisb: no14:18
alexisbok, thanks14:18
wallyworldbut it was the most likely and plausible cause14:18
alexisbjust want to make sure I report back the status correctly14:18
alexisbwallyworld, I could use about 30 minutes of your time today if you can spare it14:19
TheMuewallyworld: so it may also fix lp:1338511?14:19
alexisbI need an education on bootstrap stuff for the SRE proposal14:19
wallyworldbug 133851114:20
_mup_Bug #1338511: api-endpoints fails if run just after bootstrap <landscape> <juju-core:In Progress by themue> <juju-core 1.20:Triaged> <https://launchpad.net/bugs/1338511>14:20
wallyworldTheMue: we are looking tat that now14:20
wallyworldalexisb: sure, any time14:20
TheMuewallyworld: can simulate it on local when consuming some time between ensureMongo and openState14:20
alexisbwallyworld, I will send you an invite14:20
wallyworldTheMue: thanks, we're looking to see if the previous fixes have helped and what else needs to be done14:21
TheMuewallyworld: great, I’ve got to thank14:23
axwTheMue: you mean the error occurs when you add a sleep between those two things? or adding the sleep stops it from happening?14:23
axwI'm able to get the EOF just about time I bootstrap && api-endpoints14:23
TheMueaxw: on local I had it only once, since then no more. but when adding a sleep between those two I get it consistently14:24
axwhuh, weird.14:24
axwTheMue: thanks14:24
TheMueaxw: yw14:24
* TheMue listens and looks outside, we’ve got a warning for heavy thunderstorms14:25
bodie_would love a quick LGTM on https://github.com/juju/juju/pull/163 -- it's a little overripe14:29
bodie_( mgz ?)14:30
TheMueNow it starts14:30
* bodie_ is jealous of TheMue 14:31
bodie_We had a great one earlier this week -- some of the largest bolts I've ever seen14:31
bodie_one must have been about a block away14:32
alexisbsinzui, when are we targeting the 1.20.1 release?14:32
wallyworldwhen the devs get their stuff together :-)14:33
sinzuialexisb, Ahead of schedule...When I created the milestone, last week, I set 2014-07-17.14:33
wallyworldhopefully today we will have the final change committed14:33
alexisbsinzui, I am not trying to push, I just want to set expectations from those asking14:34
sinzuialexisb, I will release when CI passes the fixes. Maybe tomorrow, more likely Friday14:34
wallyworldwe will not get the restore stuff done, but there's fixes that we must release asap14:34
alexisbsinzui, ack14:34
sinzuialexisb, We can say before Tuesday...and I may have jury duty that day14:34
alexisbwallyworld, is the plan to release the restore stuff in 1.20.2?14:35
wallyworldalexisb: yeah, unless we can somehow make it done for 1.20.114:35
wallyworldalexisb: ordinarily we could delay 1.20.1, but the fixes we need to release are critical14:35
alexisbok, cool, thanks for the info guys, sorry I am behind :)14:35
alexisbwallyworld, I agree that 1.20.1 has enough critical fixes we need to get it out as soon as we can14:36
wallyworldalexisb: especially since the users can't even bootstrap a maas env etc14:36
sinzuiwallyworld, my new stable release rule is to set the next release 2 weeks after the last release. We aren't committing to it, but users can see the milestone and know we are thinking about it14:36
wallyworldsinzui: +1 to that. in this case we are pulling the trigger early due to those bugs14:37
wallyworldsinzui: so maybe we could even target 17/7 for the 1.20.2 release still14:38
sinzuiyep, and we look good because we delivered early14:38
wallyworldso stable releases a week apart14:38
wallyworldjust this once14:39
wwitzel3wallyworld: https://github.com/wwitzel3/juju/tree/020-policy-deployer14:39
katcoaxw: https://github.com/crowdmob/goamz16:08
alexisbjam, I will be late to the cloudsigma call16:30
=== vladk is now known as vladk|offline
katconiemeyer: hey, got a sec?16:59
niemeyerkatco: Heya, sure17:00
katcoi was beginning to look at implementing v4 of AWS sigs17:00
katcoand i happened upon a fork w/ the code written: https://github.com/crowdmob/goamz17:00
katcoi saw that you had made some commits; trying to figure out what the best path forward is... bringing this back into the project proper, or letting this fork go? opinions?17:01
niemeyerkatco: I don't really have a great opinion on the subject, mainly for lack of information17:04
niemeyerkatco: I don't know how that fork has changed in comparison to what we have at goamz, or which incompatibilities were made17:04
niemeyerkatco: I know we have fundamental issues in goamz's API which need fixing, though17:04
niemeyerkatco: I've been wishing to work on goamz2 for some time, and even started to solve a few orthogonal problems that are dependencies of a better API17:05
niemeyerkatco: But haven't started the v2 proper yet17:05
niemeyerkatco: I do know how it should look like, though17:05
katconiemeyer: gotcha17:05
katconiemeyer: so it sounds like, best not to merge this work back in; it appears to have changed significantly enough that a v2 from you would be duplicate work17:06
niemeyerkatco: It's not that.. I don't know what "this work" comprehends17:07
niemeyerkatco: I do know that if they're just improvements on top of the current API, they are not solving the real issues17:08
niemeyerkatco: Well, that's not correct either, sorry17:08
niemeyerkatco: They are not solving the fundamental issues which v2 ought to..17:08
niemeyerkatco: (certainly solving *someone's* real issues)17:08
axwniemeyer: right now we're just interested in implementing the v4 signing algorithm (to support cn-north-1, I think?)17:09
jcw4niemeyer: is your vision for v2 captured where I can read it?17:09
niemeyerjcw4: I have talked before in written form.. just need to find the notes17:09
niemeyerjcw4: The basis is that we have a static API that prevents access to underlying information of what is an extending API17:10
niemeyerjcw4: We need a mechanism that both offers arbitrary new fields to be injected in requests, and arbitrary new fields to be pulled from answers17:10
jcw4hmm, and gracefully handling 'schema' differences?17:11
niemeyerMy plan is to have a two-layered approach, similar to what I've done in the Launchpad package (lpad), but not quite exactly it17:11
niemeyerSo that we have a static layer offering comfortable access to functionality, which sits on top of a dynamic layer which offers access to anything17:12
niemeyerand they both can be used together, rather than having to opt for one or the other17:12
jcw4niemeyer: +1 I see17:12
niemeyerThis allows people to get unblocked quickly when they face limitations, instead of having to fork every time17:12
jcw4the dynamic layer sounds like it could possibly be abstract enough to be reused in different APIs?17:13
niemeyeraxw: If the fork has v4 implemented, we should merge it in17:13
niemeyeraxw: Just need to see if the contribution agreement was signed17:14
niemeyerjcw4: Yes, definitely17:14
niemeyerjcw4: Part of the dynamic layer is xmlpath17:14
niemeyerjcw4: Or rather, based on it17:14
niemeyerjcw4: That's why it was written, in fact17:15
jcw4very interesting17:15
katconiemeyer: apologies; at a sprint, and lots of unrelated talk in the room :p17:15
niemeyerjcw4: But I never moved forward with the rest of the plan17:15
jcw4niemeyer: priorities I guess, but it does sound useful and cool17:15
katconiemeyer: so you're proposing to only merge the v4 code in from the fork?17:16
niemeyerkatco: Well, I'm just proposing that if v4 is indeed necessary and it is in the fork, we should merge it assuming that whoever wrote it signs the contribution agreement17:17
niemeyerkatco: I'm not saying "just" that17:17
niemeyerkatco: It really depends on what else is at stake17:17
katconiemeyer: understood. the only snag i can see is that the way they're doing signing is quite different (i.e.: better) than how we are doing it atm17:18
katconiemeyer: so it would be a drop-in replacement.17:18
katconiemeyer: and then there's the contributor question; maybe it's just easier to dupe the change and wait on your v217:19
katconiemeyer: i'll get some more input from the team, and we'll make a decision. tyvm for your input :)17:20
niemeyerkatco: Thanks for asking17:23
axwTheMue: I've gotten to the bottom of the EOF error, and have a fix. Going to reassign to myself18:03
TheMueaxw: great, good to hear. will make the maas folks happy18:10
axwTheMue: would you please review this tiny change? https://github.com/juju/juju/pull/27918:18
axwoops, I left a debug statement in18:18
TheMueaxw: ping me when it’s removed18:20
axwTheMue: fixed18:20
TheMueaxw: ok, looking18:20
TheMueaxw: no?!? that easy?18:22
axwTheMue: indeed :)18:22
TheMueaxw: fantastic, +1, reviewed18:23
TheMueaxw: yw, I’ve got to thank you18:23
=== lifeless1 is now known as lifeless
sinzuiwallyworld, the functional-backup-restore tests failed for the 1.20.1. Either we fix it soon or we give up and say no stable release of juju ever supported backup18:58
wallyworldsinzui: that functionality is still being worked on18:59
wallyworldsinzui: perrito666 is the man to bug :-)19:00
perrito666wallyworld: oh that was so managerial :p19:00
wallyworldrestore is being translated from bash -> go19:00
wallyworldand with the rewrite, robustness etc added19:00
perrito666wallyworld: sinzui I am sprinting once again on this thing19:00
wallyworldsinzui: it will be ready for 1.20.2, perrito666 promises19:01
perrito666wallyworld: I swear you are growing pointy hear19:01
perrito666sinzui: please remind me, --ha or --ha-backup?19:02
perrito666found it19:03
sinzuiwallyworld, perrito666. Don't make those promises. Not because I don't have faith in your skills, but restore was broken by HA development...someone else is always breaking this feature. I don't think unit tests can be written to prevent a HA change from breaking restore an hour after you fix it.19:04
sinzuiperrito666, --backup and --ha-backup as of today :(19:04
wallyworldsinzui: i forgot the smiley before, here it is now :-)19:04
TheMueSo, last job of today before looking football again (to see who’ll be our competitor on Sunday) a PR https://github.com/juju/juju/pull/28119:28
* TheMue waves a good bye to the channel o/19:29
axwenjoy :)19:29
TheMueaxw: thx, will do19:30
wallyworldaxw: https://github.com/juju/juju/pull/28220:32
axwwallyworld: looking20:33
thumpermramm: you around?20:46
thumpermramm: got time to chat?20:55
mrammthumper: sure20:55
waiganithumper: I want to setup a fake pub key to test. I've got keysDir := c.MkDir(). Now, I want to copy testing/sshdata/authorized_keys/id_rsa.pub into it. Are there any test helpers I should be using?22:04
thumperwaigani: not sure, take a look at the auth keys tests22:04
thumpersee if there is something reusable there22:04
waiganithumper: good point, ok22:04
waiganithumper: next question. I'm not sure how to create a failing test. ProvisionMachine will always succeed, even if you pass it a non-standard path to check for keys - even if that path does not exist. It just tries to find keys to upload and if it can't it continues on without them.22:11
thumperwaigani: mock out the api22:12
waiganithumper: okay, and check that the file path being passed in is found?22:13
thumperwaigani: well, what are you trying to test?22:13
waiganithumper: that ProvisionMachine will upload a public key passed to it, that is not in ~/.ssh22:14
thumperbug reference?22:15
waiganithumper: https://bugs.launchpad.net/juju-core/+bug/127046622:16
_mup_Bug #1270466: allow specifying a key when doing manual provisioning <manual-provider> <ssh> <juju-core:Triaged by waigani> <https://launchpad.net/bugs/1270466>22:16
waiganithe fix is straight forward22:16
thumperwaigani: can you push your branch as it is to github so I can see it?22:19
waiganithumper: sure22:19
waiganithumper: sorry here is the diff https://github.com/waigani/juju/compare/manual-prov-key22:27
waiganithumper: I have not touched the CLI part yet obviously, but that will just be adding an extra flag/arg to pass in the key22:29
* thumper looks22:30
waiganiactually, maybe I should be adding the test to authkeys22:31
waiganiactually, custom keys have already been covered in authkeys tests...22:33
thumperif the adding of custom keys has already been tested elsewhere, don't do it again22:38
thumperthe real test is that the right key is read, and passed through22:38
waiganiyep, I don't see that actually being tested in authkeys22:39
waiganiso I'll copy testing/sshdata/authorized_keys/id_rsa.pub into a tmp location and pass that in and check that it gets read22:40

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!