Sysadmin needing advice

I don't really know where to begin here, but I need help.

I have been a professional Linux admin for the past 15 years. I come from a background of managing large fleets of various *nixes, and in my last gig I am the principal architect for the on-prem Linux environment of the company that I work for. We deliver services to in-house customers in the form of applications that are mostly COTS, but we also do a bit of development internally. I have adopted Ansible, Git, and Docker (mainly Compose stacks) for most of those services, though some continue to run natively on VMs. I have implemented monitoring, alerting, logging, near zero-touch provisioning, etc. So I know my way Linux, and I know my way around development fundamentals.

But I have not had the chance to play with K8S yet, though I read on the primitives (Pods, ReplicaSets, PVs and PVCs, Ingress, etc) and understand containerization in general. I also have no formal cloud experience. I understand the abstraction layer that it provides for datacentre infrastructure, and I understand that most of the services that I have to babysit today are managed for me out of the box. I also explored Terraform a little bit, and understand the idea behind immutable infrastructure. I tried to implement some of the same principles on-prem, but datacentre automation is still years behind what you get in the cloud, despite the efforts of HP, Dell, Red Hat, VMware, and alike (we need to manage several bare metal hosts as well as VMs).

I feel like I am missing a big part of the "modern IT" experience. Things like CI/CD pipelines, Serverless computing, cloud-native applications. I can grasp the concepts behind them, and given time I could certainly pick any of it up. But there is no hope to introduce any of that where I work, and I mostly learn when I have to use something in real life (I can follow tutorials but most are very artificial, and having a toddler at home makes it hard to dedicate time to train myself outside of work). There is also a lot of risk in jumping ship and likely take a pay-cut to start off as a junior DevOps Eng (I know, not a job title, but that's what's advertised) or RSE somewhere else. And the market for DevOps is so global that you could get anyone to work remotely for less than what would have to be the minimum salary I can accept without having to sell our house and cancel our mortgage (not something my family are in a position to do).

I am not even sure if [Dev\]Ops is something I want either, nor think it has a lot of life left. I enjoy programming but I don't really want to retrain as a professional programmer. And I do believe declarative infrastructure is a bit limiting, so I wouldn't be surprised if something like Pulumi eventually takes over. Yes, there will still be plenty of jobs in Ops, and plenty of jobs for Linux SysAdmins. But they might be seen the same as COBOL developers are seen today: a relic of the past, still well paid because there are so few of them left and legacy never dies.

In a way, I feel trapped in a silo of legacy technology with no way out. I could easily wing it for the next several years without ever having to do nothing but the minimum to keep up, and my company would still be happy. But I wouldn't. None of my colleagues seems to be bothered by the fact that they will cease to be relevant in the next 10-15 years. I have no other friends in this space. I literally have no one to talk to that understands how things are moving and where they're going.

Please help. What can I do?

https://redd.it/o5jbhp
@r_devops

Differences between Azure site Recovery (AZR) & Zerto Virtual Replication (ZVR)

Cloud consumers generally ask “What are the differences between Azure site Recovery (AZR) and Zerto Virtual Replication (ZVR)?’” The very first reason consumer ask this potential question is because they are familiar with one of the technologies and not the other and on the surface, they sound similar.

Here’s our step by step guide, why ASR and ZVR are very different from each other and meet different levels of performance and usability. 

Azure Site Recovery (ASR) and Zerto Virtual Replication (ZVR) are both great products that can help any IT organization to improve the success rate of their disaster recovery plan. The purpose of this document is not to be a kill sheet for one or another product, but rather to highlight the pros and cons of each product in relation to VMware virtual machine replication in the public cloud of Microsoft Azure.

Both products are very capable of fulfilling the above-mentioned task at a high level, but as this article will show, each product has advantages over the other depending on the workload that needs to be protected.

For example, while it takes a little more time to initially configure Zerto Virtual Replication, it includes more time – saving workflows not only for disaster recovery but also for migration and testing. Failback is also another area where Zerto shines above ASR as the required networking is already configured and a single check mark can be used to enable reverse replication.

One of ASR’s major advantages is that it can protect Azure’s physical workloads. Zerto has no support for physical workload and is limited in terms of supported workloads from the source. Today only the source is supported by VMware vSphere and Microsoft Hyper – V. In theory, if the operating system is supported in Azure, any physical or virtual machine can be replicated with ASR. This is because both physical workloads and workloads based on VMware are based on ASR. 

The installation requirements for each product are unique and ASR is a clear winner in terms of ease of installation and speed for customers who only need to protect some VMs on a single Hyper – V host. But as the number of protected workload increases, the necessary ASR infrastructure can become overwhelming and the scalability ease of ZVR becomes much more evident.

When compared to a monthly subscription bases, ASR and ZVR are very similar in terms of costs. ASR’s list price is $ 25/month per protected VM and ZVR is approximately $ 21/month per protected VM. Both solutions also require you to pay for the storage that your replicated workload consumes, but for Zerto there are a few additional costs that ASR does not need. One such cost is the compute instance running the Zerto Cloud Appliance, and the other is the cost of connecting to a VPN or ExpressRoute. So, remember that ZVR may be richer in features, but it may also cost a little more than ASR as well.

Azure Site Recovery

Four major components are needed to replicate the Azure Public Cloud from Azure Site Recovery on – site. These components are downloaded and installed on a Windows machine in your on-site datacenter from the Azure Vault configuration wizard.

A Config Server is the first server you install. A Config Server is a central administration server that communicates with the Azure Vault and performs on-site jobs. Two other components, a Process Server and a Master Target Server, will also be installed on your main Config Server by default.

The component of the Process Server is what processes and transports the protected data to the Azure Vault. Depending on the size of the workload to be protected, process servers have CPU, memory and disk requirements. If a single Process Server is unable to scale up large enough to protect your workload, it is possible to take a scale-out approach and deploy additional process servers.

Master Target Servers are required only during operations of failback. It should be noted that a Master Target Server can only fail servers of the same type of

operating system so that a Windows Master Target Server cannot fail Linux servers. A Linux Master Target server also needs to be installed for Linux servers.

A Mobility Services Agent is the last component of the ASR topology. This agent is installed on each virtual machine that is protected and is responsible for sending changed data to the Process Server so it can be shipped to the Azure Vault. In summary, ASR installs multiple premise components that differ depending on what workloads you are planning to protect. The amount of planning and the number of components may be acceptable for small to medium-sized deployments. A customer may have to invest large amounts in planning for larger deployments as it will require more than one process server and more than one master target server. This can make deployment more complex and means you have to maintain multiple machines.

Zerto Virtual Replication

Zerto has two main components, Zerto Virtual Manager (ZVM) and Zerto Virtual Replication Appliances (VRA). Each physical host in your environment receives a VRA if it contains protected VMs or receives data for protected VMs. This usually means that all hosts in a cluster will receive VRAs.

For multi-site deployments, each site receives a ZVM and for each of the physical hypervisor hosts, clusters with protected data receive one VRA.

Zerto leverages a combined ZVM / VRA architecture called a ZCA or Zerto Cloud Appliance for public cloud sites such as Azure and AWS. For Azure, the ZCA can be deployed from the Azure Marketplace or you can build a MyZerto installation package from scratch with a Windows VM and Zerto for Azure. Since the ZCA uses an embedded VRA, the only way to scale Azure’s replication capacity is in a fashionable scale. So an additional ZCA will have to be deployed once a ZCA has reached its maximum (about 100MB / s throughput).

In summary, the architecture of Zerto contains the same number of components regardless of how many on-site VMs you protect. Planning is also straightforward because it scales along with your environment, which means that if you add an ESXi host or a Hyper – V host, you also add a VRA.

Read More : https://www.taliun.com/differences-between-azure-site-recovery-azr-zerto-virtual-replication-zvr

https://redd.it/o5jby6
@r_devops

Preparation failed: Got permission denied while trying to connect to the Docker daemon socket at unix:///var/run/docker.sock



Trying to transform a project in gitlab to docker image. The gitlab is selfhosted. This is the error I get:

Running with gitlab-runner 13.12.0 (7a6612da) on test -KnwQXuT Preparing the "docker" executor ERROR: Failed to remove network for build ERROR: Preparation failed: Got permission denied while trying to connect to the Docker daemon socket at unix:///var/run/docker.sock: Get https://%2Fvar%2Frun%2Fdocker.sock/v1.25/info: dial unix /var/run/docker.sock: connect: permission denied (docker.go:858:0s)

I added a runner,registered it, created a gitlab-ci.yml with docker template which I left unmodified. As solutions I tried chmod 666 /var/run/docker.sock
before and after adding the runner but it did not work. I also added sudo before running and registering the docker runner but still no success. I found another option to add my docker user to group docker but all my users are in the docker group. The docker runner is run as a docker image. I did not know what to do...Please help. I`m am trying for about 10 days with no success... Thanks in advance

​

https://stackoverflow.com/questions/68081163/preparation-failed-got-permission-denied-while-trying-to-connect-to-the-docker

https://redd.it/o5jlqu
@r_devops

Developer Research Project - Participate in a Virtual Interview

HubSpot's marketing team would like to interview professional web developers to gather information about their skillset, career path, and any educational resources they use. Participants will be compensated for for time.

What does participation look like?

If you are selected to participate, one of our researchers will schedule an hour long call. During this time you will be asked a number of questions about your background and experience as a developer. Your name and answers will only be used internally within HubSpot.

Who are we looking for?

\- Full-time professional web developers

\- Front-end, back-end, or full-stack experience

\- Developers who work for a single employer, an agency, or are self-employed

\- Must be comfortable answering questions in depth via a call

How to Participate:

Complete this form and we will reach out if you are a good fit for this project.

Thank you for your time!

https://redd.it/o5luhb
@r_devops

Need advice on building a container provisioning system (think HackTheBox)

Hi there,

Gonna keep the deets of this to a minimum for confidentiality but also just for sake of reading time. I am working on building an infrastructure/platform where a user can provision "labs" (basically one "play" machine and one testing machine, or a few play machines within a lab etc. Now I'm mainly looking for a high level advice on how I can set such a thing up. I was researching container services, and also looking at ways to facilitate this. I know that automation is going to be a big thing, perhaps Ansible / Terraform / Kubernetes. I am well versed in Python, however I've not worked on something of this magnitude. The provisioning methods can probably be done as functions which can then be used by other teams working on the front-end website.

Maybe one of you devs could give some sort of idea on how you would go about doing something like this? Any advice/tips/insight would be deeply appreciated.

https://redd.it/o78sgy
@r_devops

Hii all, i am new to devops.. I am currently working as an devops intern. I feel like creating yml-ci files and understanding stuffs are lil difficult. Is auto devops really helpful?. Kindly give me some suggestions where to start and best way to learn it.

Rookiee..!!!

https://redd.it/o5i61g
@r_devops

I NEED HELP WITH GRAFANA DASHBOARD PROVISIONING!

>EDIT - THIS PROBLEM WAS SOLVED - TL:DR SOLUTION -
>
>Make sure your target names in Prometheus match with Variables in Grafana Dashboard JSON
>
>Unlike the popular $DS_AEROSPIKE_PROMETHEUS missing or not found error, such as these **tickets**. I did not encounter any of these after following the above step.
>
>The version of Data Source and Grafana mentioned in the JSON files do not matter, it's a common thought that might flash you, please don't waste your time on it like I did.
>
>And finally, I'd love to thank @**AnachronisticAdmin** for helping me out, I really appreciate the patience.

​

Problem is - Dashboard loads, but data does not,

I am nearing a deadline to set up a Grafana Dashboard that monitors Aerospike Cluster from Prometheus, my dashboards are being taken from here, and versions under use are as follows:

Prometheus - 2.26.0
Grafana - 8.0.3 and
[Dashboards from here](https://github.com/aerospike/aerospike-monitoring/tree/master/config/grafana/dashboards)
Step-by-Step documentation followed from Here

Now, steps followed are:

1. Installed Grafana using .deb installation, and dpkg-i
2. Created the /var/lib/grafana/dashboards/<dumped-a-few-jsons-here>
3. Provisioned Datasource as Prometheus in /etc/grafana/provisoning/datasources/all.yaml
4. Provisioned Datasource as Prometheus in /etc/grafana/provisoning/dashboards/all.yaml
5. Permissions and ownership set at /etc/grafana , /var/lib/grafana and /var/log/grafana level
6. Starts Grafana Server , set password,
7. I can view the datasource I created, and by dashboards are loaded as well

CHECKS:

1. Data Source is Working
2. Prometheus is up and running displaying healthy targets on its UI
3. Individual Panels from the same Data Source is also working.

&#x200B;

BUTTTTT

THERE IS NOT DATA ON THEM :(

Fixes I have tried are,

Common issue of $DS\_AEROSPIKE\_PROMETHEUS or $DS\_PROMETHEUS not found was attempted by replacing $DS\_AEROSPIKE\_PROMETHEUS to Prometheus in all the JSON files I have.
Tried to play around with the Variables from UI , but to no luck , please help me out here.

https://redd.it/o5i9vc
@r_devops

What does DevOps do? (Alliteration intended)

Hey! I'm working on a couple of products at my new job. The products are all about website monitoring and status pages.

I'm in marketing (tech isn't really my strong suit) and I've been tasked with understanding what the DevOps role involves and how the products we build can help engineers like you guys. If you could give me your take on the role, what you're responsible for (without divulging any secret details), the kind of tools you use, and how tools like Jira come into the picture, it would definitely give me a leg up.

P.S.: More than happy to talk about the products on request. Right now I'm trying not to make this post too promotional.

P.P.S.: I really hope I don't sound like a terrible spy trying to gather dirt on the competition.

Cheers!

Edit: Removed the links to my products

https://redd.it/o5j4i6
@r_devops

News VSCode extension "Blockman" to Highlight nested code blocks with boxes

Check out my VSCode extension - Blockman, took me 6 months to build. It's free. Please help me promote/share/rate if you like it. You can customize block colors (backgrounds, borders), depth, turn on/off focus, curly/square/round brackets, tags, python indentation and more.....

https://marketplace.visualstudio.com/items?itemName=leodevbro.blockman

Supports: Python, Dart, R, Go, PHP, JavaScript, JSX, TypeScript, TSX, C, C#, C++, Java, HTML, CSS and more...

This post in react.js community:

https://www.reddit.com/r/reactjs/comments/nwjr0b/idea\_highlight\_nested\_code\_blocks\_with\_boxes/

https://redd.it/o7cefu
@r_devops

How to identify a count index in Terraform

I want to deploy multiple resources in Azure with Terraform, I'm using the count index to specify how many resources I want but I would like to deploy some extra disks in one VM but it gives me errors when I tried to execute it

My configuration looks like this



\# Create a virtual machines 
resource "azurerm_virtual_machine" "vm" {
count = 2
name = element(var.vm_names_grupo1,count.index)
location = var.location
resource_group_name = azurerm_resource_group.rg.name
network_interface_ids = [element(azurerm_network_interface.nic.*.id, count.index)\]
vm_size = element(var.vm_size_grupo1,count.index)
storage_os_disk {
name = element(var.vm_osdisk_name_grupo1,count.index)
caching = "ReadWrite"
create_option = "FromImage"
managed_disk_type = "Standard_LRS"
disk_size_gb = element(var.vm_osdisk_size_grupo1,count.index)
  }
storage_image_reference {
publisher = element(var.image_reference_publisher,count.index)
offer = element(var.vm_image_offer,count.index)
sku = element(var.sku,count.index)
version = "latest"
  }
os_profile {
computer_name = element(var.vm_names_grupo1,count.index)
admin_username = var.admin_username
admin_password = var.admin_password
  }
os_profile_linux_config {
disable_password_authentication = false
  }
}
resource "azurerm_managed_disk" "example" {
count = length(var.extradisks)  
name = element(var.extradisks,count.index)
location = var.location
resource_group_name = azurerm_resource_group.rg.name
storage_account_type = "Standard_LRS"
create_option = "Empty"
disk_size_gb = 10
}
resource "azurerm_virtual_machine_data_disk_attachment" "example" {
count = 2
managed_disk_id = element(azurerm_managed_disk.example.*.id, count.index)
virtual_machine_id = element(azurerm_network_interface.nic.1.id, count.index)
lun = "10"
caching = "ReadWrite"
}

The problem is in this line

virtualmachineid = element(azurermnetworkinterface.nic.1.id, count.index)

I don't know how to specify a vm

https://redd.it/o7cu62
@r_devops

Want to go into DevOps and AWS. Any project to begin with? Any material for a beginner to study the AWS certification?

I'm a web developer with knowledge on PHP, NodeJS and SQL databases. Last year I've started using Docker to create a development environment for my projects and last week I've finished a course of Docker + Kubernetes with a project being deployed to Google Cloud. I'm interested to get into DevOps because I like managing services servers and analyze metrics (I've hosted many gaming servers on old PC with linux), so I'm looking for projects to run in Docker and deploy them on GCP while I still have the 300 free credits.

Also, since a lot of companies are asking specialists with knowledge on AWS I'm looking material to study for the certification. Since I've never used AWS and I have 0 experience on cloud computing I'm asking for a Coursera course or if the AWS official docs are enough for the certification

https://redd.it/o7dyem
@r_devops

does devops mean something else or what? how do you define devops?

I am not sure if I understand the term DevOps correctly even though it's in my job title. Somebody corrected me with another definition in a post a month ago and got more upvotes than me, so I am confused (of course, I could be wrong)

1. https://old.reddit.com/r/cscareerquestions/comments/ncbq99/juniorsweatbigtechthisisnotwhatiexpected/gy4465s/?context=3

Please read the original post as well to see where I was coming from.

And what do you think DevOps is?

https://redd.it/o7evuj
@r_devops

What is the best way to learn about devops?

Could you please tell me the best way to learn about devops?
The tools, software, etc...?

https://redd.it/o7eq2o
@r_devops

Which one of these certificates will help me for DevOps career

Hi, my school is offering one free course for the below certification. Which one do you think will be helpful during applying for a job in DevOps or a similar role such as release engineer? They take at least 12 days to complete.

1. Microsoft Certified: Azure Developer Associate

2. Microsoft Certified Azure Administrator  

3. Microsoft Certified: Azure Data Fundamentals

4. Microsoft Certified: Azure Administrator

5. Microsoft Certified: Azure Al Engineer Associate


&#x200B;

&#x200B;


Also not related but from the same school:

1. APU Certified PHP & MYSQL Developer

2. APU Certified ASP.NET Developer

3. APU Certified Data Analyst

4. APU Certified Java Developer

5. APU Certified Cybersecurity Engineer

https://redd.it/o7hbif
@r_devops

Refactoring obsesssive Dev

So be me, helping with terraform, aws, various processes and operations stuff in a three men team plus manager within rapidly developing startup

Not everything I can do on time, some ideas get stale, firefighting questions burns down, backlog exists and compliance pressure is high, deadlines are tight.

Now backend dev starts to walk through pipelines, terraform code and beautify them, increase readability and parametrize everything. Touches drafts and concepts only, avoids prod code

He is not asked doing so
He has features backlog started sprints ago
There were talks on this behavior - does not help
Management squeezes biz value out of his code efforts
Started blaming on me blocking his coding testing etc with some absent integrations, but dies when asked shall the code work if he has integration or requirement x in 15 minutes

After his refactoring i have strong feeling i am that bad i don’t deserve work, profession, food
Got less sleep also

Looked at his code, ran sonar, created tickets on some findings regarding unification across projects, docs and so on just out of rage and to see his reaction to his own issues - got devaluated for “creating tickets is ineffective communications”

Frankly speaking I dont want to leave the company, other people are fine.

Im not new to operations sre and devops, but this whole thing is a bit too much for my paper-thin skin and i feel like i need to go back to school

Manager is already looking for secondary developer (or says so)

Any survival advice appreciated

https://redd.it/o7icm4
@r_devops

Do any Ci/Cd Tools support a parallel build feature for many concurrent jobs?

I've spent the last few months re-inventing the wheel apparently, just learned about CI/CD tools yesterday. The last 24 hours were spent researching which would best meet my fairly unique needs, but I am still overwhelmed by all of the options.

In a nutshell, my goal is to build a web-based firmware compiler. Simple page, it has one TextArea where the end user can adjust a configuration file as needed. When this is submitted, the multi-part form data is sent to the backend, compiled, and the binary is returned for download.

Until now, I was building a RESTful server in Python with Flask. It would handle the compile requests as described, but I quickly realized I needed a queue system for concurrent jobs, authentication security, etc. Basically a woefully underwhelming version of CI/CD.

Ultimately I need an API that allows for initiating builds remotely (seems a non issue with most), but more importantly, one that can run the same build as different instances. For example, 5 website users all try to compile firmware at the same time, I need to either process these at once, or queue them up to build one after the other. Are there any CI/CD tools better suited for that than others?

I've been looking at Jenkins, Gitlab and Amazon CodeBuild thus far, but would greatly appreciate any advice or suggestions.

https://redd.it/o7j88e
@r_devops

For those wanting to get into DevOps

This is oriented towards those new to the industry moreso, but been seeing a lot of posts asking how to learn DevOps and the skills.

Establish a learning routine. Pick a series of projects from the internet like setting up a minecraft server with AWS, launching a static web page, basic Linux administration etc.

Preferably pick a group of projects that cover Cloud, Linux, IaC, CI/CD. It doesn't have to be complex. It can be simple.

Then do them everyday. Everyday. Until you can do it without watching the videos. Until you can bring in your mother and explain to her in detail whats going on and why and how. Do this routine before you tackle on something new. Before you watch a new tutorial.

Obviously as your skills grow you should apply and add to this routine, but the most important thing is that you do it. Certs help you get interviews sometimes, but they will never speak for you. At the end of the day you need the experience of doing it constantly to help you do the talking.

Doing something everyday consistently is experience.

Doing something once and moving on is a experience.

If you can understand the difference then you're already ahead of the curve.

https://redd.it/o7jvoi
@r_devops

Penetration testing tools cheat sheet



Penetration testing tools cheat sheet, a quick reference high-level overview designed as a quick reference cheat sheet providing a high-level overview of the typical commands a third-party pen test company would run...



https://highon.coffee/blog/penetration-testing-tools-cheat-sheet/

https://redd.it/o7krw2
@r_devops

Would you take the job?

From talking to friends in this company it sounds like the management is trying to improve their processes especially after a mass exodus but no one appears to know how to run a software development shop even though they excel at manufacturing. They’ve got all the problems, it sounds like they were copied out of the Phoenix project before any lessons were learned. I’m being considered for the second Devops position and while I’m fairly young in the field, 3 years since undergrad, it appears to me that my experience to-date has equipped me with knowledge which no one there has yet. For example keep WIP low, reduce unplanned work, get someone between the devs and marketing to keep promises realistic, use something simpler than the classic Gitflow, reduce constraints optimizing for speed not cost, GitOps, etc.

What kind of questions would you ask during the technical interview to sound out the managerial willingness to listen to input and openness to change? I’m all for a chaotic job if there’s the possibility of bringing order, but I don’t want to change jobs and get stuck following people who don’t know how to manage the flow of work into and out of a software team. It makes everything miserable for everyone. Being a large and fairly old corporation, I especially anticipate that the lack of official certs and years on my resume will make my opinion appear sophomoric. Thoughts? What would you ask? (Obviously I’ll be asking diagnostic questions to get a better feel for organizational health, but I need to figure out the managers.) What would you want to see before you took the job?

https://redd.it/o7m9f4
@r_devops

Am I good enough for a devops role?

I have about 6 years of experience working in an automation/QA role with a lot of other stuff including some devops sprinkled in. I have been creating and managing the whole test automation infrastructure of different projects through mostly internal tools . This is has made me very comfortable with

Linux (all OS flavors),
bash
python .

I have intermediate experience with

jenkins
ansible
docker
AWS
Probably few other things that I am forgetting .

BUT I have done all of this only through the command line and not through higher levels of abstraction like kubernetes , terraform , etc . Also most of my work has been making these tests and its automation systems from scratch and then deploy & monitor them which means I do not have experience working along with other people.

I want to transition to a full time "devops" role but I do not know if what I know is enough to get my foot in.

If what I know seems insufficient or I just want to boost my chances , what do you guys think will help me in getting there ?

https://redd.it/o7ng95
@r_devops