A place for non-work-related flimflam, faffing, hodge-podge or jibber-jabber you’d prefer to keep out of more focused work-related channels. Archive: https://archive.sweetops.com/random/
“Home datacenter. A 3-node Raspberry Pi cluster running kubernetes, and a dedicated Pi with a connected 433 mHz receiver for talking to some temperature sensors around the house.!” – thought this was a pretty cool setup so thought i’d share haha
thats good shit
@Matthew Good find. Very cool stuff
”Kubernetes the Hard(ware) Way”
LOL. the accuracy.
is kubernetes the shim we need to get to mainstream #serverless by 2020?
packer + windows = good ansible + windows = great packer + ansible + windows = nasty crap
why is that @justin.dynamicd ?
Mostly the Ansible provisioner for packer is just poor. It works by invoking a custom ansible driver that creates a local proxy to send all packer commands through which causes all sorts of connection reliability problems. I ended up just using shell-local instead :/
It’s s nightmare to troubleshoot when the inventory is constantly spewing out that it’s talking to 127.0.0.1 so it’s up to you to remember if that task is actually local or if it’s proxied to the host.
But it’s what the client wants, so gosh darn it I find a way to make it work
so instead I’m now running powershell on the host that generates an inventory file based on settings as queried, then i download the finished file and run a local ansible-playbook referencing it.
You want more fun stuff @antonbabenko ?
Does anyone use AWS ECR?
we have a few modules that use it
Terraform module that implements a web app on ECS and supports autoscaling, CI/CD, monitoring, ALB integration, and much more. - cloudposse/terraform-aws-ecs-web-app
Terraform module to build Docker image with Jenkins, save it to an ECR repo, and deploy to Elastic Beanstalk running Docker stack - cloudposse/terraform-aws-jenkins
I’ve literally never had this issue before, docker image built on my local machine gets uploaded to ECR and when i deploy that image, it comes across corrupt with the exact same configuration and it makes 0 sense. It appears my image being uploaded is corrupt and i’ve been troubleshooting for hours
Doesn’t get thrown when i build locally, did a full docker system prune -a as well so it is built from scratch
I usually use DockerHub and not ECR so what i’m thinking is my image being uploaded is dropping half the image or getting corrupt somehow
Hrm, I used ECR exclusively for about 8 months and never ran into an issue of that sort
& troubleshooting is tough because that exact same Dockerfile gets built locally and works as designed and this image worked 3 days ago.
Does the issue appear when you push to DockerHub?
I also hit this error yesterday with ECR and did absolutely nothing but re-push today and it worked properly
I have not tried this image with DockerHub just yet because i’m building an EKS cluster and would like to use ECR for simplicity
ECR was very smooth and working very well for me just 1 week ago, then today my image just isn’t deploying properly even has a smaller image size and I can’t make sense of it. Deleted the entire repository and have pushed through plugins as well as manually uploaded
Class 'Phar' not found in /usr/local/bin/composer:23 sounds like missing Linux libraries
I try to use the Silex framework as base for my web application. However, if I try to include the *.phar archive, PHP throws the following error: Fatal error: Class ‘Phar’ not found in /var/www/
Yes but the sad thing is, is the image has those dependencies just when it’s deployed they are lost and dropped
When i build the exact same image locally it works
where do you build the image before pushing to ECR?
are you using
:latest or a specific release tag?
This exact same image worked 3 days ago but today doesn’t and it really makes me nervous because i’m trying to get things ready for production and all of a sudden the most random behavior ever occurs
specific release tag
no, like semver
can you share your kubernetes deployment/pod spec?
have you set the
Yes let me get that for you
Looks like Captain Picard to me
That's because Moore's Law really is over, he says: "We are now a factor of 15 behind where we should be if Moore's Law were still operative. We are in the post–Moore's Law era." This means, Patterson told engineers attending the 2018 @Scale Conference held in San Jose last week, that "we're at the end of the performance scaling that we are used to. When performance doubled every 18 months, people would throw out their desktop computers that were working fine because a friend's new computer was so much faster." But last year, he said, "single program performance only grew 3 percent, so it's doubling every 20 years. If you are just sitting there waiting for chips to get faster, you are going to have to wait a long time."
We are in the post–Moore’s Law era.
@Erik Osterman (Cloud Posse) Yeah, that line caught me off-guard
Anyone ever have troubles with ECR images behaving differently that on your local machine? Image works on my local machine, did a force rebuild even and then when uploaded to ECR and deployed the container dies with errors.. Literally can’t make sense of why the image uploaded wouldn’t behave exactly the same
I’ve never had this problem with DockerHub and we are exploring into ECR and i’m uncertain why it’s behaving in such a way
try specifying image sha to get exact version… just to make sure that you are running exactly the same version
hold on - will give an example
so you might try do docker run and specify sha to get exact version
I haven’t tried it with run command though - only with dockerfile
but still worth a try
@Matthew also, if your image uses any config or env vars, make sure they are ok inside the container - the problem might be not in the image, but in the settings/config
her is more about getting exact version https://email@example.com/container-deployments-a-lesson-in-deterministic-ops-a4a467b14a03
Surprise! Determinism rears its beautiful head again.
i’m dropping off now - its late in my time zone. But please post what it was
In one timeline a quick path to clarity. In the other a long and painful journey trying to understand the obscure intent of a line of code. The only difference between these two realities? The revision history.
This is a talk about writing maintainable code. But rather than the code itself, we’ll see the impact a codebase’s history can have on its maintainability. We’ll explore the differences between a useful history and an unhelpful one, and you’ll learn about the practices, tools and techniques that can make the difference.
Tekin Suleyman is a freelance consultant who’s been shipping Ruby code for over a decade. He’s worked with teams, large and small. He also runs the North West Ruby User Group in Manchester.
A colleague just shared this but during a time where I’m unsubscribing from like everything, I can’t bring myself to register to read the entire article. If someone feels like taking one for the team and posting a brief here, that’d be amazing: https://www.infoworld.com/article/3304282/ci-cd/jenkins-cicd-is-in-trouble-so-its-founder-wants-to-split-it-up.html
The autonomous nature of the Jenkins community has made it unable to solve some issues, which are becoming more pronounced with the project now more than ten years old
They’re pretty slick. They’re truncating the article’s content on the backend unlike most. Can’t just element inspect this problem away
http://www.jsfuck.com/ (don’t try that at home )
I have just tried inbox for the second time and I don’t like UI still.
Four years ago, Google took a fresh look at email. With Gmail looking more and more like Inbox, it’s declaring “mission accomplished.”
Yea, me too
Have a great week-ends! Dive into the week-end - https://www.facebook.com/ODDSbible/videos/679380252428829/
gmail like interface for github notifications
maybe i can get to zero?
JSON’s license only allows use for good, so when Google dropped their “don’t be evil” slogan they had to build protobuf
Businesses often have one big question for us: How can they better understand their website visitors and deliver more relevant, engaging experiences?
You were in my hood and didn’t stop by?!?
Oh snap! That’s too bad.
Free tech e-book everyday. Sometimes worth $30-50
Download a new free programming eBook from Packt every day! Every 24 hours we will offer a new eBook on topics like Python, DevOps, Angular, AWS, C#, C++ and Java. Available to download in PDF, ePub or Mobi. Develop your tech skills and build your knowledge on software development, websites, infrastructure management and IT operations. Claim yours today.
Aka “not thinking”
I’m probably more excited by this than I should be:
We’re excited to announce multiple features that deeply integrate HashiCorp Consul with Kubernetes, including an official Helm Chart, automatic syncing of Kubernetes services with Consul (and vice versa), auto-join to join a cluster in Kubernetes, injectors so pods are automatically secured with Connect, and more.
that’s cool - so a nice way to bridge two worlds
Yeah, consul agents even easily install on windows so it’s a great way to make a singular SD that can grow into a envoy mesh as things migrate
or even for shops that like to deploy ‘n’ k8s clusters to be able to have a common SD/Mesh framework
And, yes, we were blocking exit as sign said