#random (2019-01)

“Microsoft has dropped a 40-foot long data-center pod onto the seafloor off the coast from the European Marine Energy Centre in Orkney, north of Scotland.”

this sounds like a thought experiment: schrodinger’s cat if a data center pod is on the sea floor loses connectivity…. how do you know if it’s operating or not? if you open it, it will let in the water and fry the computers haha it’s a new form of quantum computing https://en.wikipedia.org/wiki/Schr%C3%B6dinger%27s_cat

https://www.bbc.com/news/technology-44368813

https://www.datacenterdynamics.com/news/microsoft-azure-suffers-outage-after-cooling-issue/

https://thenextweb.com/dd/2019/01/05/github-now-gives-free-users-unlimited-private-repositories/

WOW

took them long enough

but I think this is great.

that combined with GitHub actions reduces the competitive advantages of GitLab.

Now, if they would allow project folders with multiple repos, I’ll be totally content.

I hate the flat organization with 300 top-level folders.

https://queue.acm.org/detail.cfm?id=3300018

If you are facing some opposition in your company while trying to rollout DevOps style workflows, this might be a helpful read. He makes some good arguments.

Morning people, I’ve published a new blog post about docker and data persistence on the host, I hope you’ll find it interesting and helpfull https://cloudlad.io/Docker%20host%20data%20persistence%20pattern

Q: When a file gets created by the process that runs inside the container, which UID and GID will it get assigned?
A: It will have the UID and GID of the user that is set with the USER instruction in the Dockerfile or 0 (the root user) if the USER instruction is not specified.

Ah yes, we’ve had this problem in geodesic for some time

i’ve been torn on how to fix it. at first, I fixed it this way

but then we had a non-root container and it was a PIA since as a human-operator we cannot run apk add ...

(our use-case is a bit unusual though for docker sicne we’re using docker as a thin shell)

check this out! https://github.com/mumoshu/variant

it’s a “universal cli” interface that is awesome for integration of lots of tools

you use it like this: https://github.com/mumoshu/kodedeploy/blob/master/kode

basically, you can now have a consistent interface to your most common workflows

while this can be achieved with make (we do it), I like the structure of this more since it supports args

in make, we often use environment variables like args (named parameters), e.g. make plan SOMEVAR=somevalue

with @mumoshu’s task runner, it would look like kode plan --application=foobar --environment=baz

it’s kode because the $0 script is called kode

another way to think about this is “executable documentation”

it’s documentation of how to use the commands which is so good, it’s executable

502 on zoom should url rewrite to boom

yea, this has got to be a massive failure

Zoom is having an outage FYI

variant looks very neat! I like it more than https://github.com/tj/robo which is very similar but less powerful. Good finding, Mr. @Erik Osterman (Cloud Posse)

yes, I also was mulling over using robo for the same reason

Hi, Can someone give the terraform module to provide IAM access to Cloudwatch

https://www.reddit.com/r/aws/comments/aegh6y/fargate_performance_issues_update_with_minimal/

If anyone wants to take this for a test: https://github.com/pecigonzalo/loro

let me know your thoughts

@pecigonzalo I’ve used awslogs like this - awslogs get /var/log/syslog ALL --watch to watch logs. It works. What is your tool trying to solve?

awslogs does not support some of the features, EG: log formatting like -o '[ {{ uniquecolor (print .Stream) }} ] {{ .TimeShort }} - {{ .Event.ecs.cluster }} - {{ .Event.decoded.message }}'

only log query/filtering, which I intend to support as well

I also had issues with its time filtering

It also supports global config files

$HOME/.loro.yaml to set some defaults, like since and etc

The idea is to combine what I believe is the best of the 3 linked solutions: https://github.com/segmentio/cwlogs https://github.com/jorgebastida/awslogs https://github.com/mmcquillan/lawsg

all great projects, but IMHO all were lacking X or Y so im trying to bring it all together

loro can also print JSON or raw messages, so you can parse with JQ

Would it not be better to add this missing functionality into one of those tools? Rather than another tool with largely overlapping use cases. I totally get it if this is a learning exercise

It says on the README why it was not added to cwlogs which is the closes to what im trying to get to

but if you want me to expand further: lawsg the code structure of lawsg is not great, i would have to basically rewrite the entire tool awslogs code is much better, but I believe A) binary distribution is better for this tool, B) golang is a more powerfull and faster language when dealing with many logs

cwlogs is great, but its supporting segment.io log format and that makes sense and works great, but its not generic enought for many other cases or supporting other log formats

I was literally just looking for something like this on friday

Give it a look, so far I found some cases when --follow stops following new streams

I think its because i do not discover new streams, ill check tomorrow

^ that should be fixed now and added non-json messages basic support

New slack logo is ugly

At first I thought some users uploaded new avatars, which were odd…

Yeah just seen that

Anyone know of a cmdline tool around which does this ? Acquires lock from different KV , SQL, Dynamodb / S3 exit 0 when lock is acquired exit 1 when lock is not acquired or timeout

@joshmyers has a tool

https://github.com/joshmyers/dynolocker

cool thx

I never contemplated the risks of lets encrypt going donw

https://letsencrypt.status.io/

https://github.blog/2019-01-17-lets-talk-about-open-source-sustainability/

Hah! first time I’ve seen this:

congrats @Jan @maarten @Nikola Velkovski!

https://i.imgur.com/vIf0lAD.gif

https://github.com/wagoodman/bashful

https://github.com/wagoodman/bashful/raw/master/.data/demo.gif

hahaha wait what

I talk too much?

lol, looks like Ansible to me!

bashsible

https://open.spotify.com/user/1124279316/playlist/0MJBni0UzdnML1amikx0Rc?si=NvAYMZFHT8ui0d9KxXAxaA

I’m using https://checkvist.com as a personal TODO management solution

please share how you do it on your side! (I’m sure that there are a lot of smart solutions)

@Erik Osterman (Cloud Posse) probably “productivity” channel is needed?

@tolstikov totally….. I’m all about productivity hacks.

created #productivity

ugh timezones suck. Am in New Zealand for a conference and jet lag is kicking my arse.

Trying to finish writing a talk I’m giving in about 4 hours..

What’s the talk on?

@endofcake is in NZ

https://linux.conf.au/schedule/presentation/241/

Talk is essentially talking about how automation shouldn’t be the first thing you show junior staff

and that they need to understand what the automation is doing behind the scenes

I interview people all the time

and they’ll tell me all about how they use puppet and terraform and so on

but when I ask if they can do it manually

or understand what it’s actually doing for them

I get blank looks

hahah yea, I know the feeling

Like don’t me wrong I love writing automation

and using it

but I want to understand what the hell it’s doing

Yea, i’ve noticed the same thing

I assume every one knows what an HTTP request looks like

but then ask them, and they’ve only ever used curl or some library to wrap it

I’ll be honest I’m a bit rusty on the specifics of it

because I haven’t needed to do that in quite a while

i guess it’s just about what point we jump on the rails

yep

I’d get back in the swing of it pretty quickly tbh

it’s what I do

I get handed things that are on fire

and I get to figure out how to extinguish said fire without actually breaking shit further

but yeah I planned to write the talk before I came

but I got pulled into a DC migration 2 days before I left

as the people planning it had done a fantastic job of screwing it all up

Or just basic subnet math

I find debugging questions to be great

Yea those are great

Mixed with high level architecture questions

Does anyone have recommendations for “applicant tracking systems” aka ATS?

I’ve used http://www.greenhouse.io/ in the past,

never used anything else so can’t compare but it fulfilled our needs

Sounds like you have a good problem to solve! More applicants than you can handle via emails or in memory. :)

https://hire.google.com

This looks nice too

https://pi-hole.net/

I found it after this proposed change https://bugs.chromium.org/p/chromium/issues/detail?id=896897&desc=2#c23

@Erik Osterman (Cloud Posse) not a software recommendation but rather a framework to build hiring process which we use since some time is https://www.amazon.com/Essential-Guide-Hiring-Getting-Hired/dp/0988957418/

https://blog.travis-ci.com/2019-01-23-travis-ci-joins-idera-inc

I never heard of Idera

Me neither

I wonder if they’re gearing up to compete more fully with GitlabCI and potentially with whatever integrations Microsoft adds to GitHub

From other comments on HN, sounds more like Idera is about buying up SaaS products with a large base of legacy users, but not necessarily innovating.

basically ensuring the longevity of the product, but not trying to out do the competition

I lol’d a lot harder than I probably should have :p https://www.reddit.com/r/pcmasterrace/comments/aj5aks/this_is_your_ram_and_this_is_your_ram_on_chrome/?utm_source=reddit-android

https://www.reddit.com/r/Terraform/comments/ajggfw/mask_sensitive_data_in_output_from_terraform_plan/

https://github.com/auchenberg/vscode-browser-preview

https://ohshitgit.com/

anyone else catch themselves typing git commands into slack by accident all the time?

~for me it’s always ls~

ls

I do Ctrl+w and close tabs/windows all the time when trying to delete a word

http://viorel.me/2018/the-dot-ro-gold-rush/

Anyone every worked with https://concourse-ci.org ? Would like to hear opinions.

https://podman.io/releases/2019/01/16/podman-release-v1.0.0.html

@maarten used concourse briefly. Was nice but seemed very tied into CloudFoundry/bosh

Have you looked @ Drone?

Hi @joshmyers, a potential customer wants to use it, would be part of the assignment.

It is a bit of a learning curve IIRC

@maarten re concourse - I spent some time POCing it mid 2017. in general it didnt fit my idea of good. from skimming my notes (I apparently didnt keep a copy of the ticket where I summarized this for a client…): scaling (auto or otherwise) is a significant challenge (perhaps easier if your a bosh shop). secrets management is nonexistent. yml configs don’t ship and branch etc with the code (ala circle, travis, most of shippable, etc), you actively push them to the concourse server. docker support inside a project seemed painful (no native support) (ie, if you wanted to build a container then use it in your build flow, that was less than sane to achieve)