=== rogpeppe2 is now known as rogpeppe === ant_ is now known as Guest47305 === admcleod_ is now known as admcleod [10:06] hey admcleod I am getting this: http://pastebin.ubuntu.com/19259834/ [10:06] this is on zeppelin and everything in on apache-* charms [10:06] any thorughts? [10:07] kjackal: interesting, so lzo is an issue. can you also pastebin your juju status? [10:08] http://pastebin.ubuntu.com/19260221/ admcleod [10:09] kjackal: in /usr/lib/hadoop can you run this: bin/hadoop jar share/hadoop/common/hadoop-common-2.7.1-tests.jar org.apache.hadoop.io.TestSequenceFile -seed 0 -count 1000 -compressType RECORD xxx -codec org.apache.hadoop.io.compress.LzoCodec -check [10:09] kjackal: on the spark unit, and on the slave [10:13] admcleod: on the spark node: http://pastebin.ubuntu.com/19260497/ [10:16] admcleod: and here is running it on the slave: http://pastebin.ubuntu.com/19260637/ [10:17] kjackal: ok so its possible that spark doesnt know where the hadoop libs are. (also, does the log that you're trying to create the spark context with actually exist?) [10:18] admcleod: the error is the same regardless the file I am tryin to access [10:18] kjackal: https://gist.github.com/andrewdmcleod/e32e8982c0b8f8083442 < check these files to see if they have the appropriate paths.. if not, add, restart the pyspark interpreter [10:19] kjackal: the classpath including the specific lzo jar should already be in one of those files [10:22] kjackal: can you also try to run 'spark-submit' at the cli on the spark unit and pastebin if there is an error? [10:23] spark-submit gives no errors and sparkpu.sh runs ok [10:24] kjackal: are those paths in the files (re the gist)? [10:24] so. no! [10:25] kjackal: thats interesting. im pretty sure kwmonroe put those in somewhere. [10:25] some of the files in the gist are not there (compute-classpath) and none of the rest contain the lines the gist wants to inject [10:25] kjackal: it may be possible he (someone) put them in different files [10:30] kjackal: are you trying to add them manually to see what happens? [10:30] admcleod: yeap waiting for the job to finish [10:31] nope-same error.... [10:32] it is as if pyspark does not load the classpath from the expected routes. [10:32] admcleod: ^ [10:32] kjackal: and you restarted pyspark interpreter, try restarting zeppelin copletely? [10:32] +m === CyberJacob is now known as zz_CyberJacob === zz_CyberJacob is now known as CyberJacob [13:38] kjackal: any luck? [13:39] admcleod: not realy... just documented the problem we have in a card and going to look at this in more depth now [13:40] do you have the commit where we add lzo to our classpath? [13:40] kjackal: no, sorry [13:40] kjackal: i looked and couldnt find anything [16:03] ping magicaltrout [16:04] yo lazyPower [16:05] magicaltrout - i need to introduce you to some people. You've taken point on the dcos/mesos work right? [16:05] it does appear i'm that man [16:05] awesome, i'll hit you with an email introduction then. [16:05] cool [16:11] magicaltrout - are you still tom at analytical thingy dot whatever? [16:11] yeah lazyPower [16:11] ta [16:11] anything you have is current [16:50] Saw this post : http://www.pasteall.org/pic/105055 Would this be the OpenStack Charms Meeting, or not ? (also attending #ubuntu-meeting) Sorry If I am Late. [17:04] Suergally that appears to be correct [17:04] ddellav ^ can you confirm? [17:05] lazyPower Suergally yea, looks like it. gnuoy can confirm [17:06] lazyPower: ddellav Not sure about the details .. but I was in the 'meeting' for half an hour with no correspondance. [17:07] ah good point, i see 5pm - i wonder if that calendar is UTC... [17:08] I was there from 5.30 pm so UTC or not - it is UTC irrelevant. [17:13] ddellav: gone are the days when we should have mis-billed calendar events - so can you confirm the iCal entry is correct or not, and possibly post something to correct that, cheers. [17:17] Hello Team, My sample charm and it is pending for long time on unknown state with the log http://paste.ubuntu.com/19194830/ and code files are under this branch https://code.launchpad.net/~prabacha/charms/trusty/sampledatabase/devel As per this reactive code i am not able to set the state as 'mysql-config.done'. Could some one help me on this issue? [17:18] hi Prabakaran, i'll take a look..i just deployed your layered charm and will dig into the mysql state issue. it'll take me a while though. [17:19] hey cory_fu, remember that "config.x" states get removed sometimes? suchvenu figured it out.. when leader-elected runs, it seems to clobber the config.x. [17:21] it's pretty easy to repro: http://paste.ubuntu.com/19289554/ [17:28] ddellav: Did you get that or do you need more+ time ? [17:29] Suergally i do not have control over that calendar. perhaps gnuoy or lazyPower can make those changes [17:30] lazyPower: Did you get that message I gave some 15 minutes ago ? === mup_ is now known as mup [17:30] Suergally - Yes. I'm speaking with openstack charmers. You're continued patience is appreciated :) [17:31] lazyPower: What channel are they attending ? Cheers. [17:32] Suergally - the meeting has moved both channel location and time/date. I don't have any further details at this time [17:32] kwmonroe: Is the charm looking for config.changed or config.default/set ? Because config.changed is documented to go away and the end of the hook, even if it's not handled due to other conditions. [17:33] * Suergally kinda -doesn't know- who everyone is 'talking to' :( [17:33] cory_fu: http://bazaar.launchpad.net/~ibmcharmers/charms/trusty/layer-ibm-db2/trunk/view/head:/reactive/ibm-db2.sh#L528 [17:33] config.set.X [17:34] hi Suergally - apologies for the stale calendar entry. that just changed as we became an official openstack project yesterday. next meeting will be: Monday July 24 at 17:00 UTC in #openstack-meeting [17:35] beisner: terrificco ! Can I have an I'Cal entry to add to my calendar off me google calendar thingy, then , please ? [17:36] kwmonroe: I am quite surprised that the config.set state would drop off, but I would recommend removing lines 157 and 158 from layer:basic: https://github.com/juju-solutions/layer-basic/blob/master/lib/charms/layer/basic.py#L157 [17:36] See if that fixes it, by chance. [17:36] will do cory_fu [17:37] Suergally, indeed we will update that, but it may not be updated yet today. [17:37] beisner: So I dont have to add anything ? I'm lost (?) [17:37] Suergally, correction. new channel location is set for #openstack-meeting-4 [17:38] Suergally, Monday July 24 at 17:00 UTC [17:38] Suergally, but i can't update the ical yet so please take note of the time/place. [17:39] beisner: Can this be brought to the attention of the mailing-list, then ?# [17:39] wow i'm winning with date calcs here. Monday July *25* at 17:00 UTC [17:43] beisner: That's the same time as this meeting (security) http://www.pasteall.org/pic/105070 Or is that DoDo too ? [17:43] BRBack (coffee needed ;-) [17:44] Suergally, it may very well be. typically different audience and community. the openstack project meeting schedules are the scheduling confines that openstack-charm meetings now follow. [17:50] beisner: That's a shame that they cross-match. Surely it's easier to have it an hour either side (?) 'just sayin [17:52] cory_fu: that worked [17:52] That's very strange [17:53] kwmonroe: Is there any chance that the leadership-elected hook in that particular charm is overriding the hook implementation from the base layer? [17:54] beisner: And 2ndly - that's not mentioned here yet : https://wiki.ubuntu.com/ServerTeam/OpenStackCharmsMeeting \or anywhere I can trace ! [17:54] doesn't look like it cory_fu http://paste.ubuntu.com/19292199/ [17:54] Suergally, apologies there. our challenge is to coordinate among 148 (as of now) other openstack recurring meetings. [17:55] kwmonroe: I'm fine with removing those two lines, as they aren't ideal anyway, but maybe you could step through the leader-elected hook (specifically though init_config_states) to figure out why they weren't getting re-added? [17:55] Suergally, rest assured, we are working on updating things following yesterday's change. [17:56] beisner: I think most of your users only follow the 'fridge' calendar :) In terms of wikis, google calendars and #channel-links to the aforementioned meetings. [17:56] kwmonroe: I'm a bit concerned that they should have been re-added at the start of the leader-elected hook and if that happens to run before config-changed (can happen), then the states will get missed [17:57] IOW, I'm concerned that there's some other bug that removing those lines is masking [17:59] ack cory_fu [17:59] beisner: How can I help or a/improve ? [18:00] and I think the openstack charm dev/community docs are still a work in progress so feedback pull requests could be welcome there [18:01] cargonza: do you have a link to the latest? [18:06] oh hold up cory_fu [18:06] this isn't right http://paste.ubuntu.com/19292199/ <-- that's not the leader elected from layer-basic [18:06] (there is no basic.init_config_states) [18:07] Suergally, we've got folks on the necessary content updates. it will not update today, but all areas will accurately point to the right place very soon. thanks again for bringing it to our attention. [18:07] kwmonroe: Right you are, my good man. Some other layer must be overriding it [18:07] even money on the leadership layer [18:07] stuuuuuuuubbbb [18:07] kwmonroe: That'd be my guess, too [18:08] beisner: No problem .. just thought you could point me to a mailing-list so we can talk about it (?) [18:08] kwmonroe: Yep: https://git.launchpad.net/layer-leadership/tree/hooks/leader-elected [18:08] cory_fu: so is the best way forward for stub to remove those hooks and let the basic layer provide them, or sync those hooks with what's in layer basic? [18:09] arosales, which doc are you looking for? [18:09] i vote the former.. i don't see anything in there that's not covered by basic's leader-x hooks. [18:09] kwmonroe: Yes. Those files were put in the leadership layer before charm-tools created that hook, so they now need to be removed [18:10] roger that [18:10] cargonza: the openstack charm dev doc [18:11] upstream doc: http://charm-guide.readthedocs.io/en/latest/ [18:13] cargonza, ah yes. the issue here is we have two other locations with stale content pointing to the old meeting location and meeting time: http://fridge.ubuntu.com/calendars & https://wiki.ubuntu.com/ServerTeam/OpenStackCharmsMeeting [18:14] ok have addressed both... I think. We'll get the cal updated asap. [18:20] ooo you guys grabbed some prime url real-estate with that one [18:21] cargonza: The details are to be pointed at the #openstack-meeting-4 #channels I presume ? Perhaps look-out for what else can be averted/binned. [18:23] ack, thx Suergally. we'll fix it asap [18:23] Cheers, where the juju mailing list ? [18:24] **where's the .... [18:25] stub: i learned lp git! MP for your eyeballs at your convenience: https://code.launchpad.net/~kwmonroe/layer-leadership/+git/layer-leadership/+merge/299986 [19:35] hey whats up guys? Few questions concerning aws cloud network spaces [19:36] is there anyway to make juju aware of pre-existing spaces ? [19:36] or ehh [19:36] pre-existing aws networks [19:36] or can juju only create new spaces? [19:38] cory_fu: look ok? https://github.com/juju-solutions/interface-mysql-root/pull/1 prabakaran's deploy is g2g with this change [19:38] kwmonroe: Apparently it's fine [19:39] But yes, it also looks good to me [19:39] Thanks, marcoceppi : [19:39] :p [19:39] yowzer marcoceppi.. thanks! [20:02] sorry, I was going through my email, saw it, merged it [20:25] #noragrets [20:42] S.O.S. -> http://imghub.org/image/EyHX [20:43] can someone take a look at what I'm doing wrong when adding my subnets ? I'm in the middle of a time crysis, and got stumped on this .... I just need to know the way!!!! [20:45] bdx: what are you doing currently? could you map out what you have for VPCs? [20:45] rick_h_ alexisb_ ^^ can we get some help from the networking side? [20:46] marcoceppi, unfortunately most of our networking folks are out for the day, email may be best at this point [20:46] but we can try [20:46] thumper, you around? [20:47] aye [20:47] I have a feeling I'm not creating the spaces first ... [20:47] can you take a look at bdx's sos and see if there is anything you can help with [20:47] * thumper looks [20:47] thumper: can I screen share w/you real quick? [20:48] I'm not sure how helpful I'll be, but sure [20:49] bdx: which version of juju? [20:49] devel [20:49] 2.0-beta11-trusty-amd64 [20:49] and aws? [20:49] trying to use an existing vpc? [20:49] ya [20:49] * thumper takes a quick look at the code [20:50] thumper: I see "juju add-subnet" and "juju add-space" to be the commands I need to get the vpc added as a space righ? [20:51] I think when you bootstrap the aws controller, you need to specify the existing vpc there [20:51] ooooh [20:51] :-) [20:52] go on ... [20:52] vpc-id [20:52] Description: "Use a specific AWS VPC ID (optional). When not specified, Juju requires a default VPC or EC2-Classic features to be available for the account/region.", [20:53] Yes! where did you find that>?? [20:53] um... in the code [20:53] shoot ... [20:53] I wonder where it is documented for real users [20:53] it is in the config docs for the ec2 provider [20:53] i've read that somewhere before [20:53] and i don't read the code :) [20:53] I'm sure it is somewhere [20:55] bdx: I *think* for aws, you can specify the subnets but it won't create them for you [20:55] you still need to do that manually [20:55] not sure on what level of discovery the ec2 provider has [20:55] when dealing with existing vpcs [20:55] bdx: when bootstrapping, the provider will check to see if the vpc is sufficient for juju's needs [20:56] it does have a force option, but TBH not sure why one would want to force it to use a VPC it doesn't think is god [20:56] good [20:56] thumper: nice, ok ... so the deal is, I need to provision some instances in a pre-existing vpc .... do I then need to bootstrap with that vpc-id? [20:57] yes [20:57] I believe so [20:57] ahhh, I see [20:57] trying now [20:59] bdx: I do recall that there was a bug somewhere with beta 11 and aws [20:59] something like us firewalling off the controller by mistake :) [20:59] but only in particular circumstances [20:59] you may or may not hit it :) [21:00] thumper, bootstrap is underway === natefinch is now known as natefinch-afk [21:00] in my vpc [21:00] thank you, thank you, thank you [21:00] no worries [21:01] thumper: thanks for the heads up on ^^ too [21:07] thanks team for the support! [21:32] beisner: if you ask the news team to update the fridge, I'm sure they'll be happy to (it's #ubuntu-news ;) ) [22:47] sos follow up -> http://imghub.org/image/W1EU :-) === urulama is now known as urulama|___