| hml | https://bugs.launchpad.net/juju/+bug/1732665 | 08:09 |
|---|---|---|
| mup | Bug #1732665: Firewaller issues on vmware vsphere <firewaller> <vmware> <vsphere-provider> <juju:In Progress by hmlanigan> <juju 2.3:Confirmed> <https://launchpad.net/bugs/1732665> | 08:09 |
| === dames is now known as thedac | ||
| WesleyM | Hello, I'm kinda stuck on my juju deployment. I'm trying to deploy juju into my MAAS node and everytime I get the error "ERROR unable to contact api server after 1 attempts: unable to connect to API: Forbidden" does someone know the solution to this? | 15:12 |
| rick_h | WesleyM: can your laptop speak to the network that the maas nodes are on? | 16:07 |
| rick_h | WesleyM: that's usually an issue where MAAS brings up the node, installs the juju agent, but the local client (running juju bootstrap) cannot reach that new controller | 16:08 |
| WesleyM | Rick_H: It's all running on the same machine, my laptop is connected with a SSH connected to the machine in that range | 16:22 |
| bdx | heyyyo | 18:05 |
| bdx | for me, `juju deploy postgresql --storage pgdata=ebs,2000G` gives http://paste.ubuntu.com/p/K7bTjJ2RD2/ | 18:31 |
| bdx | anything over 1000G seems to just hang | 18:33 |
| kwmonroe | bdx: does aws impose ebs limits for your account? i know, for example, that default accounts can only spin up X machines of type Y. curious if you have to request to get an ebs limit raised. | 18:36 |
| kwmonroe | either way, it'd be nice to see an error instead of waiting -- but maybe something to check on your end. | 18:37 |
| bdx | kwmonroe: I can create 5T, 10T vloumes via cli or aws console | 18:37 |
| kwmonroe | hmph | 18:37 |
| bdx | yeah | 18:37 |
| kwmonroe | and bdx, this hangs the same way whether you use 2000G vs 2T? | 18:37 |
| bdx | yea | 18:39 |
| bdx | heres my controller log http://paste.ubuntu.com/p/KnjJ7PkqBn/ | 18:39 |
| kwmonroe | thanks for a 60K line pastebin. please hold while i watch my browser crash. | 18:41 |
| bdx | LOL right | 18:57 |
| bdx | Im surprised pastebin took it | 18:57 |
| bdx | mine still hasn't finished loading | 18:58 |
| bdx | its 10MB | 18:58 |
| bdx | should machine-0.log get rolled at some point? | 19:01 |
| bdx | logsink.log gets rolled .... but machine-0.log isn't being rolled | 19:04 |
| bdx | kwmonroe: I can deploy an instance `juju deploy postgresql` then add a 2000G volume and attach it manually and it works just fine | 21:12 |
| bdx | https://paste.ubuntu.com/p/jjQfbVBSmW/ | 21:12 |
| bdx | its only when I try to use juju to preform the operation that it fails | 21:12 |
| kwmonroe | you're out of my league bdx.. i'm not sure what's going on when you run juju storagey things. rick_h, who's our storage point-person? | 21:14 |
| bdx | I just need verification that `juju deploy postgresql --storage pgdata=ebs,2000G` fails for someone other than myself so I can get some legitimacy on this bug | 21:16 |
| kwmonroe | oh, hell, i can do that.. gimme a minute. | 21:16 |
| magicaltrout | just drop the boring data..... | 21:17 |
| kwmonroe | quiet you | 21:17 |
| magicaltrout | use juju; drop table kwmonroe; | 21:18 |
| bdx | this bug https://bugs.launchpad.net/juju/+bug/1751909 | 21:19 |
| mup | Bug #1751909: turn up the volume - artificial quotas? <juju:New> <https://launchpad.net/bugs/1751909> | 21:19 |
| magicaltrout | i hope you're cold kjackal | 21:35 |
| kjackal | freezing! magicaltrout | 21:35 |
| magicaltrout | good | 21:35 |
| kjackal | Anything for you ;) | 21:35 |
| magicaltrout | when you see Uros... tell him he didn't reply to my email he said he'd reply to a week ago | 21:36 |
| magicaltrout | *sob* | 21:36 |
| igor | AWS deployment | 21:49 |
| kwmonroe | figured you out bdx... default "ebs,2T" is using the prev gen of aws storage: https://aws.amazon.com/ebs/previous-generation/ | 21:49 |
| kwmonroe | which taps out at 1tb | 21:50 |
| magicaltrout | goooo kevin | 21:50 |
| bdx | oh what | 21:50 |
| kwmonroe | whatcha need to do is use the new gp2 volume type: juju create-storage-pool mybytes ebs volume-type=gp2 | 21:50 |
| kwmonroe | juju deploy postgresql p2 --storage pgdata=mybytes,2000G | 21:50 |
| bdx | oh dude | 21:50 |
| kwmonroe | er, s/p2// | 21:50 |
| bdx | you are the man! | 21:51 |
| igor | AWS kubernetes cluster deployed with juju | 21:51 |
| kwmonroe | i know, right? | 21:51 |
| bdx | dude | 21:51 |
| magicaltrout | igor: you need to formulate a question with those words | 21:51 |
| bdx | aha | 21:51 |
| igor | Unable to mount volumes for pod | 21:51 |
| igor | this is the error when when I try to deploy wordpress chart | 21:52 |
| igor | the cluster looks good and functional | 21:52 |
| kwmonroe | bdx: that should unblock you, but the bug you opened is still relevant in that i think surely there's a message at the provider saying why 2tb can't be fulfilled. we should capture that and inform the user (vs waiting around like log bumps) | 21:53 |
| igor | docker image deployed fine, but failing when it's claiming storage in cluster. the cluster has dedicated controller as per juju deployment wizard | 21:54 |
| igor | any post configs that should/can be applied to fix https://usercontent.irccloud-cdn.com/file/qXCr2sly/image.png | 21:56 |
| igor | magicaltrout: woulkd this be the right forum for juju kubernetes deployer in AWS | 22:00 |
| magicaltrout | yeah igor although i'm not very familiar with it personally | 22:02 |
| kwmonroe | yup igor, you're in the right place... ryebot, any ideas on igor's "unable to mount volumes for pod" when deploying a wordpress chart with k8s/aws? | 22:02 |
| * ryebot catches up | 22:02 | |
| ryebot | Hmm let me take a look at something | 22:04 |
| igor | sure | 22:04 |
| ryebot | igor: What kind of storage are you using? | 22:06 |
| TheAbsentOne | Hey guys, to fully understand how juju (and the reactive framework) fundamentally work I want to recreate the phpmyadmin charm to the reactive framework and find out a way to log all the hooks and processes the juju controller makes. Any tips on how to approach this, or anyone knows some other sources or documentation then the juju docs site? | 22:06 |
| igor | https://www.irccloud.com/pastebin/VatLAPJW/ | 22:07 |
| ryebot | igor: looks like it's pending - did you set up a persistent volume? | 22:08 |
| igor | ryebot: volume.alpha.kubernetes.io/storage-class: standard and I've tried with volume.alpha.kubernetes.io/storage-class: default | 22:08 |
| igor | no sure | 22:08 |
| igor | how do I check? | 22:08 |
| ryebot | igor: You'd know, I think. | 22:09 |
| ryebot | igor: I think you'll need to use native integration and EBS like this: https://docs.conjure-up.io/devel/en/cni/k8s-and-aws#how-to-setup-ebs-for-kubernetes | 22:09 |
| igor | this is sample chart to test deployment | 22:09 |
| TheAbsentOne | can you mount manually on aws? (are you sure where the problem resides?) | 22:10 |
| igor | you mean from container itself manually? | 22:11 |
| ryebot | igor: conjure-up can set up native integration for you | 22:12 |
| igor | great ryebot: should I just follow your link to the doc above to do that? | 22:13 |
| TheAbsentOne | yeah, just a random thought though, nvm it actually it's late for me xD | 22:14 |
| ryebot | igor: Yeah I'd perform an installation with conjure-up to enable the cloud integration for aws and follow those instructions such that they fit with what you're trying to do. | 22:19 |
| bdx | kwmonroe: sc1 | st1 .... not on juju's terms | 22:29 |
| igor | ryebot: just did that and got the same error | 22:33 |
| igor | https://usercontent.irccloud-cdn.com/file/nLn64W5s/image.png | 22:33 |
| bdx | kwmonroe: https://paste.ubuntu.com/p/2rnz86Q2pS/ | 22:33 |
| bdx | support for st1 and sc1 - https://bugs.launchpad.net/juju/+bug/1753593 | 22:40 |
| mup | Bug #1753593: support st1 and sc1 ebs volume-types <juju:New> <https://launchpad.net/bugs/1753593> | 22:40 |
| kwmonroe | ack bdx, that was a wag on my part. so it seems for now you'll need to use a new pool with gp2 (or possibly use the default ebs-ssd pool) if you want to get > 1tb. i'm guessing that's not great for the pocketbook :/ | 22:42 |
| kwmonroe | +1 on your latest comment in that bug | 22:42 |
| bdx | totally .... I only *need* this 5T elasticsearch for the next 2 weeks I think, so I'm not to worried, but going forward it will be super nice having st1 and sc1 for sure | 22:44 |
| bdx | got me unblocked though - https://paste.ubuntu.com/p/qJ62FFH6KR/ | 22:45 |
| bdx | I dont even know if the st1 or sc1 would support the iops for this es ... Im going to see how that ^ m5.2xlarge with the 5T gp2 holds up | 22:46 |
| bdx | hooking it up now | 22:46 |
| kwmonroe | ugh bdx, next time tell me *before* you run these kinds of models.. https://finance.yahoo.com/quote/AMZN/ -- i'm assuming your tomfoolery drove AMZN up +23 today. | 22:48 |
| bdx | darn .... that really makes me want to run this somewhere else | 22:51 |
| * bdx (just angry I don't have any shares) | 22:54 | |
| magicaltrout | shares? doesn't everyone buy bitcoin these days? | 22:57 |
| magicaltrout | ICO's baby *\o/* | 22:57 |
| igor | ryebot: I've installed with conjure-up kubernetes and select AWS | 22:57 |
| igor | is there anything else to do during/after install to enable the cloud integration? | 22:58 |
| igor | I've got dedicated controller node and 14 cluster nodes build fine in AWS | 23:00 |
| igor | ryebot: looks like AWS permissions are fine: | 23:26 |
| igor | https://www.irccloud.com/pastebin/42gzgUqV/ | 23:26 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
