[01:26] axisys: shot in the dark, but can you login with those creds here: https://jujucharms.com/login/ [02:13] kwmonroe: yes I can.. I just did.. [02:13] but still same failure with conjure-up [02:14] axisys: can you file a bug at https://github.com/conjure-up/conjure-up [02:17] I can deploy a kubernetes model from the jujucharms gui.. no problem [02:17] using aws [02:18] axisys: so you can do it without logging in through jaas [02:19] stokachu: yes .. using firefox [02:19] even google chrome [02:19] I just exported the model yaml.. [02:20] yep.. checked and my aws now has 3 instances of ec2 running.. [02:20] No I mean use conjure-up but not go through the hosted jaas [02:20] so from terminal it did not work with cojure-up kubernetes as it failes in SS) [02:20] SSO [02:21] failed* [03:10] axisys: there are 2 ways to deploy, one is through SSO (jaas) which is what you are picking, the other can go diretly through AWS [03:10] so when you select which controller to utilize you pick the second option *self hosted* [03:10] at least until we get this jaas issue sorted [06:57] kjackal: Sorry for late reply, we are prepping for midsummer here.... [06:58] no worries erik_lonroth_ [06:58] kjackal: I'm still begugging why spark won come up [06:58] do you have any logs we can look at? [06:59] I'm able to find the message you linked to before. Yeah, I'll send some output. sec. [07:00] https://pastebin.com/bqWKGvZE [07:02] juju status gives: https://pastebin.com/DZey05k0 [07:02] As you can se, spark is still in "maintenance". [07:14] kjackal: A completely different thing.... I've just bought two nanopi 64 bit ARM (4 cores!) to play with. How would I do to put those as usable by juju ? =D A fun side project perhaps. [07:17] nice erik_lonroth_ :) I haven't done anything like that. I would start with https://www.ubuntu.com/download/server/arm I think SaMnCo has done some experimentation with rpi he might have blogged about it [07:17] let me see [07:19] well I mostly did MAAS on rpi but I backed off around 2.1 bc we did not release binaries for arm anymore [07:19] https://askubuntu.com/questions/431526/guide-for-using-juju-in-deploying-to-a-raspberry-pi [07:20] Oh, so you don't release binaries for ARM anymore? [07:21] Was there any technical reason for that? I have a friend that happily would endeavour into that. [07:25] erik_lonroth_: which bundle did you deploy? [07:26] juju deploy cs:bundle/hadoop-spark-37 [07:27] Let me know what I can investigate [07:28] thanks [07:40] Could this be another proxy issue? [07:40] https://pastebin.com/4RKMmBqW [07:46] erik_lonroth_: there are some resources we would rather have access to: https://github.com/apache/bigtop/blob/master/bigtop-packages/src/charm/spark/layer-spark/config.yaml [07:46] they are under https://s3.amazonaws.com/jujubigdata [07:48] Oh, "resources" - thats a chapter in the juju world I've just recently got in touch with... [07:49] So, you need the "config.yaml" from me? I've used the default and have not changed anything in there yet. [07:49] so... erik_lonroth_ can you try a deployment with the bench mark disabled? https://github.com/apache/bigtop/blob/master/bigtop-packages/src/charm/spark/layer-spark/config.yaml#L16 [07:49] actually you might be able to do that right now [07:50] since your charm is not on error state [07:50] .. OK. Can I "redeploy" easily without having to re-install all machines etc? [07:50] juju config spark spark_bench_enabled=false [07:50] I'll test immediately [07:51] but your charm is not on an error state... strange... [07:52] Woho! "spark" is now "active".... But I agree that getting stuck in the "maintenance" mode is not very informative. === frankban|afk is now known as frankban [07:53] kjackal: How to I enable the resources so that I can get the benchmark examples to run ? [07:54] you would need to have access to those s3 buckets on amazon [07:54] When I try, I get "access denied" [07:54] AccessDeniedAccess Denied609A3CEEE51904DEkg/5Bjg6iytNAlLW5tXY8hP493G7l5kfFPnVrGHXa8nxF40rMncVF9dyJqudPuEs [07:56] erik_lonroth_: can you please file a ticket for not going into an error state? [07:56] I can reach https://s3.amazonaws.com/jujubigdata/ibm/noarch/SparkBench-2.0-20170403.tgz from here [07:56] I could reach that link, but the first one gives me the access denied. [07:57] what is the "first one" ? [07:57] https://s3.amazonaws.com/jujubigdata [07:58] erik_lonroth_: a yes, https://s3.amazonaws.com/jujubigdata should give you access denied you cannot see what is there [07:58] Ah Ok [07:58] I'll file that ticket for you. What do you want me to point out? [07:59] ... also where do you like me to post the issue? [08:00] erik_lonroth_: what you can do is download the benchmark binary from here https://s3.amazonaws.com/jujubigdata/ibm/noarch/SparkBench-2.0-20170403.tgz then host it somewhere you know is accessible from your infrastructure and then when you deploy spark you can set the new url of the resource by providing this config param: https://github.com/apache/bigtop/blob/master/bigtop-packages/src/charm/spark/layer-spark/config.yaml#L33 [08:02] erik_lonroth_: for the ticket describe how you reached that deployment. Provide the logs, and say that you did not reach the ready state. Based on the error you have seen on the logs you would expect an error state but you only got the maintenance [08:02] let me tell you where to file the ticket [08:04] erik_lonroth_: do you have access here, or do you need to create an account? [08:05] Account where? [08:05] I've got a github account [08:05] erik_lonroth_: https://issues.apache.org/jira/secure/CreateIssue!default.jspa [08:06] Ah, I don't think I have an account [08:06] the spark layer is upstream, part of bigtop [08:06] erik_lonroth_: what about here: https://github.com/juju-solutions/layer-apache-bigtop-base [08:06] erik_lonroth_: can you file an issue there? [08:07] I'll try file it on github [08:08] thank you erik_lonroth_, kwmonroe will be pleased to hear you found the issue. Many thanks for the tcket [08:17] Can I reference you in github ? [08:17] Whats your usernames there so I can get it in there? [08:30] kjackal: https://github.com/juju-solutions/layer-apache-bigtop-base/issues/59 [08:31] thank you erik_lonroth_, would you mind if I linked the pastebins you showed me? [08:32] Thats no problem. [08:32] thanks [14:58] thanks for the call kwmonroe, very helpful! [15:02] thanks for showing up magicaltrout! and especially for the ping -- i was sitting in that room by myself like "dang, nobody wants to hangout with me". the priority discussion was super helpful for me. i'll fire off a minutes email so all the jealous people here know what we're talking about. [15:03] awww such love [15:05] you're just jealous admcleod_ [15:08] it is not even a joke to say that is true. === scuttlemonkey is now known as scuttle|afk [15:48] magicaltrout: I am crashing that meeting next time! [17:00] nice kjackal [17:00] ask kwmonroe what his golf handicap is [17:01] i don't like to brag, but i typically shoot in the 60s. [17:01] not bad for the front 9, #amirite? [17:51] o/ juju world [18:20] /w/win 7