A+ helped me get a $12/hour job 7 years ago. $40K entry helpdesk jobs are out there, but I've mostly just heard about them in unionized jobs, and someone applying internally can cross departments and will automatically get the job instead of you because of seniority.

I'm kind of torn about the whole internal SMB IT thing. The problem is there are a lot of bad MSPs out there, and most owners don't know how to get a good MSP. I think internal IT is better than a bad MSP.

However, once a SMB is gotten into ship shape, the workload for an SMB tech is so small. I got into development because there was so little of my time taken up but technical IT related tasks. I tried getting into improving business workflow, but that didn't get very far because I wasn't trusted as more than a tech.

For internal SMB IT to really be worth it, IT has to be deeply involved in the business side of things. There's just not enough straight technical work.

When I was internal SMB IT, my counterpart in our sister company in the states had their terminal servers managed by someone else, their network covered by someone else, and even convinced them to hire a help desk tech. They had 50 users.

I can see how SMB IT can get sucked into this. When having a conversation with a VAR they specifically try to direct the conversation to areas they suspect you are less knowageable in and try to make you feel inadequate. Their sales pitch grinds to a halt when you prove you know more about it than their sales pitch covers.

Currently I'm 'using Docker' in Gitlab in order to run unit tests when I make a commit on an application I develop personally. I also my free Gitlab CI/CD minutes to run my application in production... since it doesn't have to run 24/7 I have it scheduled for when it needs to run, so when it's scheduled to run it grabs my Docker image from the registry and runs what it needs to run.

I want to start using docker to start testing Salt config changes I commit on Gitlab. There's a few things I need to get figured out first though, like how to create a Ruby Gem because Kitchen-docker needs some modifications in order to run docker-in-docker. This would be more readily do-able if I was using my own Docker cluster.

At work I'm trying to push for us to start using Docker for Windows, because the current plan for de-coupling our one web service would cost a fair amount of money in Windows Server licences otherwise.

It sounds like they are pretty confident they can implement a good roaming profiles setup and are going to go ahead with it. It's their problem and not mine, so I'm going trust they know what they are doing.

@dafyre said in Can Roaming Profiles be a good idea?:

If you are going to be doing RDS, I would recommend using User Profile Disks (UPD). It eliminates some of the complexities of roaming profiles. Apps that don't work well with roaming profiles work file with UPDs because it connects a VHDX to C:\Users\usersname as opposed to storing all the files on \someserver\someshare\username.

Interesting that this never came up in my mcsa.
But it looks like scale out file server would be required in this situation

Scenario: 6 terminal servers in 3 locations. Application used on the terminal server has a profile folder for each user. This profile is currently stored on one file server. The application performs best when the profile is actually on the disk of the terminal server that the application is run on at the time.

The original plan was to do DFS replication for the central file server, but a proposal was made to have a file server on in DFS replication as a location for Roaming Profiles. The Roaming Profiles will ensure that the Application profile is local.

When reading this proposal, red flags come up with me on seeing Roaming Profiles. However, I have never used Roaming Profiles and have just gone with the general consensus that they are prone to slowness problems. But is there a way to do them well? I know there are some out there that manage it with Roaming Profiles + folder redirection. Do Roaming Profiles require the whole profile to be roaming? (thus requiring the folder redirection for things like Desktop and Documents)

My initial thought on reading the proposal was that we would be basically duplicating our replication. i.e. the Roaming folder gets synced to each terminal server and the roaming profile gets synced between DFS-R members. Why not just sync the data between the terminal servers?

I believe this is why the CEO of Gitlab is adamant about being able to code from tablets. Kids these days don't have a computer accessible to them, but they need to be given the ability to create.

I think we are seeing a shift from mobile devices just being consuming devices to actually using them for work. Businesses are now trying to integrate tablets into their workflow and there is a push for LoB applications to start supporting mobile devices.

There are high school graduates entering the workforce today who don't know how to navigate a file system. There is now a gap between what businesses are using for computing to what is taught in school.

The future is further abstractions so that we can work and create from phones and tablets, but we're not there yet, even though we already need to be.

It can be quite handy if you're stuck working from a Windows desktop. I remember that I used it in order to get my Test Kitchen setup working at my previous job. Something to do with some things not working with Ruby for Windows. And then it was easy to connect it up to LXD

Yep, sure do. You have to add hex together in order to get the proper value in order to keep 1.0, 1.1,and 1.2 working

I think I will make a proposal, thanks for joining the community and reaching out @multiOTP

@multiotp said in RDP - Whitelist IP address with 2 step authentication?:

Hello,
What do you mean exactly by " in SMB, people would find it annoying to have to get a code every time" ?
Any proposal welcome, and we will adapt our Credential Provider (https://github.com/multiOTP/multiOTPCredentialProvider)
Regards,
Andre

What I mean by that is, with a small business that does not have it's own IT staff, the business owner often is the person who determines the security that is used based on their own personal availability tolerance. If they find it too annoying (aka availability/accessibility is too low) to enter in a code, they will simply tell a MSP they don't want that. The MSP has to either oblige, convince them otherwise, or drop them as a client.

Small businesses with IT staff runs into a pretty similar situation. Might actually be worse because there is not a peer-to-peer relationship between IT staff and owner, and the alternative is for the IT staff to quit.

@multiOTP Welcome to the community!

@scottalanmiller said in UniFi Home Lab vs Campus:

@flaxking said in UniFi Home Lab vs Campus:

Idk, for $10, and a little bit of time, my Meraki AP running OpenWRT was quite the bargain for a home AP

Except you can't get the unit for $10.

The UART adapter was the $10, and I didn't even have to do the free webinar, it was a colleague that did it.

Someone somewhere out there probably has found a channel to collect unlicensed Meraki APs (partially thanks to them giving them out like free candy) and is flashing them and using the open source version of CT to create ultra cheap wireless networks.

Idk, for $10, and a little bit of time, my Meraki AP running OpenWRT was quite the bargain for a home AP

@scottalanmiller said in MS SQL Server Requirements:

@flaxking said in MS SQL Server Requirements:

How do we make this clear on our system requirements info?

Tell people to treat it like any database and do their own due diligence. It's MS SQL Server, make them get their own people to do it like any IT shop would. You don't need to tell people the requirements for the OS, you don't need to for database basics either.

This would be ideal, but I fully expect a lot of pushback

@black3dynamite said in MS SQL Server Requirements:

Do you have a least the minimum requirements to start with?

I don't think we actually list minimum requirements for the SQL Server, I believe the recommended requirements are 12GB RAM and then the note about database size

I work for a software company and sales wants an update to our system requirements.

Specifically, they want an update to our recommended memory requirements for the MS SQL Server.

The way everything is structured, I'm not really sure it makes sense for us to be giving out recommended requirements for the MS SQL server, since technically we don't take responsibility for their SQL server, that's all in their tech's responsibility.

Practically though, potential clients need to be able to see what is required to use our software.

Right now, we have a recommended RAM requirement that is more designed towards physical servers, and making sure they are starting out with enough. We then have a note about needing more RAM as the database gets bigger. And then we have an example of how much RAM you might need if you have a bigger database.

This doesn't really work because 1. it gives the impression that we have a large base requirement of RAM for a blank database, which would mean the software is does some crazy stuff, which it doesn't.
2. It gives the impression that we have a formula based on the database size for how much RAM is recommended.

Sales wants this magical formula that we don't have.

Some of our clients have 100GB database that are only utilizing 6GB of RAM (by the SQL buffer cache), while our biggest client has a 400GB database and 200GB of RAM.

There's too many factors to be able to come up with some kind of formula. How do we make this clear on our system requirements info?

I once had to work with a database that the only connection we had to it was through Telnet.

The queries could return xml output, however it was a nested database, any queries utilizing nested relations wouldn't preserve the relationship in xml.

So I had to capture the xml, then work with in python and create the associates, filter it how i wanted it, and then export to csv.

The csv part was the easiest.