[08:18] <SpamapS> hrm.. https://launchpadlibrarian.net/90171425/buildlog_ubuntu-lucid-i386.charm-tools_0.2%2Bbzr119-7~lucid1_FAILEDTOBUILD.txt.gz .. especially confusing
[08:19] <SpamapS> + dpkg -l openssh-client
[08:19] <SpamapS> + awk /^ii/ { print $3 }
[08:19] <SpamapS> + ssh_version=
[08:20] <SpamapS> especially that part, since running that locally in a lucid chroot produces ssh_version=1:5.3p1-3ubuntu7
[09:39] <rog> TheMue: morning!
[09:39] <TheMue> rog: Hey, Roger, good morning. Returned well?
[09:40] <rog> TheMue: yeah, journey back good, and i had a nice rest-of-weekend too, thanks
[09:40] <rog> TheMue: you?
[09:41] <TheMue> rog: everything fine, thx. took a first look for the may trip. that will be a harder journey. ;) at least about 25h from here, maybe more.
[09:42] <rog> TheMue: i should do the same. better organise it with a couple of days of recovery time at the start...
[09:43] <rog> TheMue: i was in the heathrow lounge with a few of the guys going back to Oz, and their flight was delayed by 11 hours!
[09:43] <mpl> where will it be in may?
[09:43] <rog> mpl: oakland, i think
[09:43] <mpl> florida?
[09:43] <TheMue> rog: yep, landing on sunday evening to start fresh on monday morning won't work.
[09:44] <mpl> oh no, CA
[09:44] <TheMue> oakland, california
[09:44] <mpl> ouch indeed.
[09:45] <mpl> TheMue: it shouldn't take that long though. I think my flight from Madrid to Chile took me 13 hours.
[09:46] <TheMue> mpl: from Germany I only found two stop trips with about 23h travel time (w/o driving from/to airport and check-in)
[09:47] <mpl> TheMue: right, everything included it might account for as much
[09:47] <mpl> oh, w/o you said.
[09:47] <rog> TheMue: flying to oakland or SFO?
[09:47] <TheMue> yip
[09:47] <TheMue> OAK
[09:47] <rog> TheMue: you might find flying to SFO is a better option
[09:48] <mpl> oh well, you'll see. good luck with that though. the way in was ok for me, but I found the way back (from chile) pretty horrible indeed.
[09:49] <TheMue> rog: Good tip, looks better.
[09:50] <mpl> rog: that's actually why I missed one full morning of the last iwp9. I was so tired from the trick back to Chile, and with a cold from the horrible constant a/c during the flight, that I overslept and didn't hear at all my phone alarm.
[09:50] <mpl> s/trick/trip/
[09:50] <rog> TheMue: yeah, SFO looks about 5 hours better for me, at least (14h vs 21h)
[09:51] <TheMue> rog: same here, and one choice is taking the A380 (nice)
[09:53] <mpl> TheMue: I've been told one big advantage of the A380 is it's way less noisy (you don't hear the reactors/engines so much as in the other ones).
[09:54] <TheMue> mpl: Would be a great benefit to at least have the chance for some napping during the flight.
[09:55] <mpl> TheMue: exactly.
[09:56] <mpl> can anyone come to the ubuntu meetings or does one have to be a canonical employee?
[09:58] <TheMue> mpl: http://uds.ubuntu.com/
[09:58] <mpl> TheMue: thx
[10:00] <mpl> cool. I might try to go to the next one happening in europe then.
[10:02] <SpamapS> While OAK is closer to the venue, SFO has tons of direct flights from Europe.
[10:03] <SpamapS> mpl: you can apply for sponsorship.. juju will get just a few slots to sponsor community members.
[10:04] <SpamapS> mpl: plane noise can be handled quite effectively by simple ear plugs. :)
[10:04] <SpamapS> oh my, its 2am, I should sleep
[10:07] <mpl> SpamapS: the price is not the main issue for me. I just try to not travel by plane too far away when I can, for ecological reasons.
[10:08] <mpl> SpamapS: and yeah I never travel anymore without earplugs, as I usually have a very light sleep.
[14:01] <niemeyer> Helloooo
[14:01] <niemeyer> How's everybody doing this fine Monday?
[14:13] <hazmat> niemeyer, pretty good
[14:13] <hazmat> ;-)
[14:13] <niemeyer> hazmat: Still here :)
[14:14] <hazmat> its a national holiday in the US as well
[14:14] <niemeyer> My connectivity is a bit on the bad side at the moment
[14:14] <niemeyer> hazmat: Oh, double holiday then :)
[14:15] <hazmat> niemeyer, we're having a good discussion in juju-dev on charm upgrades
[14:15] <hazmat> looks like no history bot on the channel though
[14:15] <rog> niemeyer: mornin'!
[14:16] <rog> niemeyer: flight go ok?
[14:16] <hazmat> niemeyer, here's the log fwiw http://pastebin.ubuntu.com/806218/
[14:16] <niemeyer> hazmat: Thanks
[14:16] <niemeyer> rog: Hey!
[14:16] <niemeyer> rog: Yep, all good
[15:09] <niemeyer> Hmm.. hungry.. lunch time.
[15:46] <gmb> Hi folks. Is there a way to run the juju test suite without it (apparently) trying to connect to EC2?
[15:46] <gmb> I assume I've missed something screamingly obvious here.
[15:48] <hazmat> gmb, it doesn't actually connect to ec2 for the test suite
[15:48] <hazmat> gmb, it may ask for values for the account but you can put in dummy values
[15:49] <gmb> hazmat: Ah, really? It breaks because my AWS credentials aren't in the environment, which seems a bit odd.
[15:49] <gmb> Thanks for clarifying though.
[15:49] <hazmat> gmb, yeah.. that's a test setup bug, feel free to file a bug on it
[15:49] <gmb> hazmat: Righto, will do. Thanks!
[16:02] <_mup_> Bug #917228 was filed: ./test --functional errors if EC2 credentials aren't in the global environment <juju:New> < https://launchpad.net/bugs/917228 >
[16:47] <niemeyer> gmb, hazmat: ./test *--functional* should probably connect to EC2, though?
[16:50] <SpamapS> niemeyer: perhaps the tests that need AWS_ should be skipped with a helpful message.. "set AWS_... to run these functional tests"
[16:51]  * SpamapS says.. from "holiday" :)
[16:52] <niemeyer> SpamapS: The error message seems pretty ok (Could not find AWS_ACCESS_KEY_ID)
[16:52] <niemeyer> SpamapS: But I agree, it could be improved
[16:55] <SpamapS> I have never been very comfortable with directing people to use those variables, especially since the aws cmdline utilities don't even use them anymore (which is quite baffling to me that they changed them)
[16:55] <mpl> niemeyer: yo. do you have a couple minutes today to discuss ssh ?
[16:56] <niemeyer> SpamapS: Huh.. when were they changed?
[16:56] <niemeyer> SpamapS: I still use them
[16:56] <niemeyer> mpl: Sure, what's up?
[16:59] <mpl> niemeyer: I haven't looked at it again, but basically I don't see how to connect the ssh connection with a zk Dial call. Because exp/ssh only gives us net conn to work with as a reader/writer. and otoh zk.Dial() does some things that I don't see how to replicate in terms of smaller operations.
[16:59] <SpamapS> niemeyer: ec2-* doesn't seem to listen to AWS_* anymore
[17:00] <SpamapS> niemeyer: I have to set EC2_... now
[17:00] <niemeyer> SpamapS: Which version of the tools are you using?
[17:00] <niemeyer> SpamapS: I'm probably missing an upgrade then
[17:00] <niemeyer> mpl: What we want to do is exactly what the Python code is doing, except we'll be using the internal package rather than delegating to an external process
[17:01] <niemeyer> mpl: The overall functionality and the reason why it works is exactly the same
[17:01] <niemeyer> mpl: The server side behavior is also the same.. we'll be connecting to sshd, not to something else
[17:01] <SpamapS> niemeyer: 1.5.0.0
[17:02] <SpamapS> niemeyer: I didn't use them really at all between Ubuntu 10.10's release and early in 11.10's cycle.. so it changed somewhere in there
[17:02] <hazmat> niemeyer, indeed but those tests are likely bit rotted
[17:02] <SpamapS> niemeyer: since euca-* is typically faster to start up.. being in python instead of java. :)
[17:02] <niemeyer> SpamapS: Ok, I'm using 1.4.3 still..
[17:02] <niemeyer> hazmat: How's that related to the bug?
[17:03] <hazmat> niemeyer, not.. i hadn't realized the --functional flag was being used.. which does indeed require  a valid ec2 account creds
[17:03] <niemeyer> SpamapS: Hmm.. not really
[17:03] <mpl> niemeyer: I understand how to pass along bytes from point A to point B through the established ssh tunnel. what I don't get is how to pass a "valid" dial as the zk server is expecting it.
[17:03] <niemeyer> SpamapS: I'm using the 11.10 version, I believe
[17:05] <niemeyer> mpl: It's normal TCP connection forwarded through the ssh channel
[17:05] <SpamapS> niemeyer: hrm, actually I may have been confused how ec2-* worked.. it never read AWS_* ... it was respecting EC2_PRIVATE_KEY / EC2_CERT ...
[17:07] <niemeyer> SpamapS: Ah, ok.. those are different indeed
[17:08] <SpamapS> Does that mean that the ec2-* utils only use the SOAP interface, while euca-* uses the RESTful API?
[17:08]  * SpamapS has never much cared how they worked.. only that they do. :-P
[17:10] <SpamapS> niemeyer: also its definitely a *lot* slower to use the ec2-* on my system.. http://paste.ubuntu.com/806410/
[17:10] <SpamapS> 4.0s vs 0.6s
[17:10] <mpl> niemeyer: ok. but do you agree that with the way exp/ssh works, I cannot uze zookeeper.Dial() and I have to replicate whatever it is doing (tcp connection plus whatever else it's doing)?
[17:10] <mpl> *use
[17:15] <niemeyer> mpl: No, I disagree
[17:15] <niemeyer> mpl: You seem to be thinking about implementing something else, that is not what the Python code is doing
[17:15] <niemeyer> mpl: Imagine you had to change the Python logic to get rid of the fact it uses the external "ssh" command for establishing connections to the server
[17:15] <niemeyer> mpl: That's all you have to do
[17:16] <niemeyer> mpl: zookeeper's concerns are completely irrelevant
[17:16] <mpl> niemeyer: it's not the same. because since the python code uses the external ssh, it just gets for free a new port that is listened on to connect to
[17:16] <mpl> niemeyer: while with exp/ssh, you just get a net conn on which to read and write
[17:17] <niemeyer> mpl: "because since the python code uses the external ssh, it just gets for free a new port that is listened on to connect to"
[17:17] <niemeyer> mpl: Yep, and that's what we want to do as well!
[17:17] <niemeyer> mpl: Again, you're thinking about implementing something else that is not what the Python version does. Think about the behavior of ssh, and imagine how you'd mimic it
[17:19] <mpl> niemeyer: ok, I have my mimicking ssh tunneling listening on that new port. I receive a zk.dial() call, so I establish the tunnel to the sshd.
[17:19] <niemeyer> mpl: Right
[17:19] <mpl> niemeyer: but then how do you
[17:19] <mpl> sorry
[17:20] <mpl> how do you forward that dial that was received?
[17:20] <mpl> hmm
[17:21] <niemeyer> mpl: Programming? ;-)
[17:21] <mpl> maybe I don't have to do anything because when establishing the ssh tunnel there alread was a dial to the zk server.
[17:22] <mpl> niemeyer: I'm sorry, I just get the feeling that I had already tried those things and failed. I guess I'll just retry.
[17:22] <niemeyer> mpl: It's certainly possible.. I think Dave even had some code doing something similar already
[17:23] <niemeyer> mpl: But I'm afraid I can't help you much more than that :(
[17:23] <niemeyer> mpl: The way I can figure exactly how to do it is by sitting down and doing it
[17:23] <niemeyer> mpl: Alternatively, we might look at some other task that you'd feel more comfortable about
[17:24] <mpl> niemeyer: as you wish. if you feel like you need it done soon, then maybe. but I'm happy to retry.
[17:25] <niemeyer> mpl: We're not in a rush about this just yet, so you certainly have some extra time to retry if you wish
[17:25] <mpl> (I don't like giving up in general)
[17:36] <SpamapS> woot, 2 big boxes of juju t-shirts just arrived at my door. :-D
[17:36] <SpamapS> jcastro_: ^^
[17:38] <SpamapS> correction, *3* big boxes of juju t-shirts
[17:43] <koolhead17> SpamapS:
[17:43] <koolhead17> jcastro_: where is my kit sir!! :P
[17:49] <SpamapS> koolhead17: what kit?
[17:49] <koolhead17> SpamapS: there is a event where am thinking to take 1-2 hr session on juju workshop, like charm school
[17:50] <SpamapS> koolhead17: ahh. Well we're still just now getting juju shirts/stickers/etc. made for the first time. So please have patience. :)
[17:50] <SpamapS> juju only spun up its first EC2 instance about a year ago. ;)
[17:50] <koolhead17> SpamapS: feb11 is event date
[17:51] <koolhead17> SpamapS: yeah i know, i am planning to show juju/charm magic on/via LXC
[17:54] <SpamapS> koolhead17: cool :)
[17:54] <koolhead17> i have  added a talk too. :)
[17:54]  * koolhead17 has fingers crossed 4 talk getting selected
[18:03] <mpl> niemeyer: sorry, had to talk with my boss. so yeah, I'll retry, thanks for the talk. going home now.
[18:15] <hspencer> koolhead17, i updated more info on juju with Walrus
[18:15] <hspencer> put a bug in with txAWS
[18:15] <hspencer> waiting to hear back
[18:15] <hspencer> but yea
[18:15] <hspencer> somethign is up with txAWS
[18:16] <koolhead17> hspencer: hey dude
[18:16] <hspencer> how was your weekend?
[18:16] <koolhead17> hspencer: but are we able to get juju working on euca
[18:16] <koolhead17> SpamapS: hspencer is euca developer
[18:16] <hspencer> err i didn't try it...i just looked at txaws
[18:16] <koolhead17> hspencer: weekend was cool, beer/strangers and music
[18:16] <hspencer> juju uses txaws to talk to Walrus correct?
[18:17] <hspencer> does it do it by default?
[18:19]  * koolhead17 looks at SpamapS hazmat niemeyer
[18:26] <SpamapS> hspencer: yes all AWS interaction in juju is done via txaws
[18:27] <SpamapS> hspencer: I think I responded to your txaws bug report about how its unfortunate but ultimately running S3 commands under a prefix path is not quite the same as S3...
[18:31] <hspencer> SpamapS, yea, it is.  But what i found more interesting is that it looked like there was some syntax checking done on on the URL before it was being sent.  I used ngrep on my Walrus component to track traffic occurring on the wire. I didnt' see anything there
[18:31] <hspencer> is there some syntax checking?
[18:32] <hspencer> just curious?
[18:33] <SpamapS> hspencer: I don't have a deep understanding of txaws, I just observed the result of the actions and confirm that the code is working as intended. I don't have a working walrus to test against though I think we need to get a eucalyptus system setup to run juju's functional tests against at some point.
[18:34] <hspencer> if you want, you can test it against the ECC
[18:34] <hspencer> you just need an account on there
[18:34] <hspencer> you don't need to launch an instance
[18:35] <hspencer> i just used the txaws commandline tools directly
[18:35] <hspencer> figured if they work that way, then it shouldn't be a problem to get it working with juju
[18:35] <hspencer> know what i mean?
[18:40] <hazmat> yup
[18:41] <hazmat> i can setup an account with ecc and have a look then..
[18:46] <koolhead17> fwereade: hey there
[18:56] <fwereade> heya koolhead17
[18:56] <koolhead17> fwereade: how are you?
[18:57] <koolhead17> hspencer: so we are almost there?
[18:57]  * koolhead17 is hopeful to get juju on euca too :)
[18:57] <fwereade> koolhead17, not too bad, just finishing up some python stuff so I can get to work on go :)
[18:58] <koolhead17> :P
[18:58] <hspencer> koolhead17, workin on it
[18:58] <hspencer> :)
[18:58] <koolhead17> hspencer: ;-)
[18:58] <hspencer> workin with hazmat
[18:58] <koolhead17> hspencer: SpamapS and hazmat are always here
[18:58] <koolhead17> yes
[18:58] <hspencer> its good stuff
[18:58] <hspencer> lookin forward to tit
[18:58] <koolhead17> indeed.
[18:58] <hspencer> to it
[20:07] <hspencer> for the folks workin on debugging txAWS issue: EMI on ECC is uploaded and available
[20:07] <hspencer> IMAGE	emi-BBBB188B	oneiric-uec/oneiric-server-cloudimg-amd64.img.manifest.xml	hspencer	available	public		x86_64	machine	eki-81ED1733	eri-E8B51886	instance-store
[20:08] <hspencer> if you have an account on the ECC, you should be able to launch and run that instance
[20:30] <hspencer> hazmat, SpamapS, niemeyer: just an FYI, oneiric image up on ECC.  Working on getting natty image up there as well.  Let me know if you guys have any issues
[20:31] <niemeyer> hspencer: Sweet, thanks
[20:31] <hspencer> yep, no problem
[20:31] <hspencer> i want to help get this issue fixed. shouldn't be having this issue
[20:31] <hspencer> plus, i would like to have another tool to use with Walrus
[20:31] <hspencer> outside of s3cmd
[20:32] <hspencer> any way i can help
[20:32] <hspencer> just lemme know
[21:41] <hspencer> hazmat, SpamapS, niemeyer: natty is on the ECC.  The image information for natty and oneiric are as follows
[21:42] <hspencer> IMAGE	emi-55CC1725	natty-uec/natty-server-cloudimg-amd64.img.manifest.xml
[21:42] <hspencer> IMAGE	emi-BBBB188B	oneiric-uec/oneiric-server-cloudimg-amd64.img.manifest.xml
[21:42] <hspencer> lemme know if you guys have any issues running those images on the ECC
[22:21] <SpamapS> hspencer: so, question.. is walrus normally setup as a sub directory in Eucalyptus, or is that just one option?
[22:32] <_mup_> Bug #917405 was filed: charm upgrade from charm store can leave units out-of-date and unfixable <juju:New> < https://launchpad.net/bugs/917405 >