r/sysadmin

GitHub Official X Post

"We are investigating unauthorized access to GitHub’s internal repositories. While we currently have no evidence of impact to customer information stored outside of GitHub’s internal repositories (such as our customers’ enterprises, organizations, and repositories), we are closely monitoring our infrastructure for follow-on activity."

Dark Web Informer says "GitHub source code allegedly offered for sale: Internal orgs and private repositories claimed

A threat actor using the alias TeamPCP claims to be selling GitHub source code and internal organization data.

The actor claims the dataset includes around 4,000 private repositories and says samples can be provided to interested buyers to verify authenticity.

━━━━━━━━━━━━━━━━━━━━
Target: GitHub
Country: United States
Sector: Technology / Software Development / Source Code
Incident Type: Alleged Source Code Sale
Claimed Exposure: Around 4,000 private repositories
Actor: TeamPCP
Price: Offers over $50,000
━━━━━━━━━━━━━━━━━━━━"

Edit: adding xcancel link, thanks jykke!

Large(ish) ISD. School admin wants a "house points" system with a lot of hesitancy about how to do it and a lot of faculty hesitation.

I (Computer Science (among other things) teacher) build a series of Sharepoint Lists and PowerAutomate flows to make a points system work (List X talks to List Y when Form Z is submitted). Nothing spectacular or particularly interesting.

While building, I realize that my own personal M365's tenant account (not personal as in outside the tenant, but my work account from within the tenant) might hit a 6k PPR/day ceiling, so I request either a service account for transparency/"bus" factor or insight into how I might get a higher rate limit if that's not possible.

CTO goes into red alert. In a private meeting from which I am excluded, mentions me "reverse engineering" the Ron Clark House Points app (lol what? It's literally get/append/update flows), how Sharepoint Lists aren't as "secure" as other parts of our Azure tenant, and how not even student email should be stored in Sharepoint due to security concerns.

I feel like I'm being gaslit, but I'm hoping those with experience can help me get some insight.

I am a senior systems administrator / projects guy and the other guy is help desk. We support a large machine shop that has some PNCs, CNC machines, various other OT equipment, and about 200 on-site end users in an office building on the other side of the property. The rest are remote users.

Last month HR informed my manager that we were just awarded an absolute metric shit ton of business and we are going to need to scale quickly. Right now, my coworker and I are pretty much maxed out in terms of daily bandwidth. 15 tickets a day on a slow day, as high as 25 or 30 on a chaotic day. Currently sitting at 71 tickets in the queue. I'm currently solo implementing three large-scale projects that is slow going because I have to dedicate at least 2 or 3 hours a day to just about doing tickets in order to help out the help desk guy keep the ticket level at a manageable number.

We've already hired about 10 people in the last month and we are being stretched even further. I came to my manager and told him my concerns and he was bluntly dismissive and told me that before they hired me, he was doing all of his managerial duties as well as laptop deployments and help desk tickets.

Am I the one who's off base here or are we on our way to being completely fucked once those additional hundred people get hired in?

Accepted a 6 figure Senior L3 Engineer / Team Lead role at an MSP today. Found out they do time tracking in Autotask for non billable time down to the 5 minutes, billable is 15. I haven't tracked non billable time in 20 years. They want 40 hours on my timecard every week. How does this work for things like checking emails, context switching, mentoring a junior, multitasking, ramp up/ down time, making coffee, taking a leak, etc? He said it's not for punitive measures it's to see where the business is spending time. I already don't want to work there because of this. Is this normal?

Im feeling so stuck and like im being lead on.
I have been in IT since 2019 and at my current company coming up on 3 years soon. Other coworkers were promoted over me to level 2 with way less experience. I finally spoke up earlier this year and my manager acknowledged I was more senior and said he would create a path and sys admin position + more money for me.
Now he did create said position but in another office and said my growth was put on pause bc the team had some cleaning up to do. That I wouldn’t be moved up to sys admin but level 2 support (like my less experienced coworkers and I had already said I didn’t want level 2 which is why he said he would create the sys admin role).

I feel so defeated, again. I have done the work to prove and show I can do more. Which he acknowledges!!??? I have spoken up and tried advocating for myself, and now I think he just said whatever to shut me up. I’ve applied to sys admin positions elsewhere but the job market is crap. Idk what else to do. I’m tired and embarrassed. I feel like I tried but keep failing. If you made it this far, ty for reading my rant lol

https://krebsonsecurity.com/2026/05/cisa-admin-leaked-aws-govcloud-keys-on-github/

There's just too many failings to recount them all.

• Keys and plaintext passwords
• Public repo called 'private'
• Ignored warnings from security researcher

Background: I have always worked alone in IT, I'm self-taught, and my largest env was 300 students and 40 staff at a small, private K12. I have ZERO experience with "standard" IT envs.

How do IT depts typically record when actions are taken, such firmware updates, configuration changes, or other "internal" events that are noteworthy but not specifically a support ticket?

I can jerry rig an existing tool to make it work or vibe code something from scratch, but I don't want to reinvent the wheel. Shirley, this is a normal part of responsible IT management.

Edit < 10 minutes later after getting several replies: Apparently, this is just an ordinary, mundane part of IT management, and I am seriously out of the loop. I feel slightly embarrassed now, but I'm gonna leave this up for others who may be too shy to ask.

Every time I contact Microsoft support, I get connected to outsourced Indian support agents who only give generic scripted responses and then quickly close the ticket on the very first contact without actually solving the issue.

It feels like they barely read what was written before replying with: “clear cache”, “reinstall the app”, “wait 24 hours”, “we understand your frustration”.

Then the ticket gets marked as solved even when nothing was fixed.

I understand first-level support follows scripts, but the current experience feels completely disconnected from real technical support.

Has anyone here actually managed to get their issue escalated to someone who genuinely investigates the problem?

Hey all, I was wondering if someone could help me out with a rough estimate of what they’re paying for their VMware hosts in 2026. We need to do a hardware refresh but some of the quotes we’ve been getting are crazy. Our team was looking for a dual core r760, but for 512GB of RAM dell were quoting like $15K for the RAM only.

Comes out at about $35,000 a server total. Is that normal in today’s market or am I completely confused? Any help would be great even a ball park figure is helpful!

Hi. I currently work for a non profit who has a Dropbox plan of 3 users for 15Tb total and it costs about 700 a year after the non profit discount.

The staff is interested in moving storage to microsoft and move our slack and storage all into teams.

Currently, it seems our non profit account gets 1Tb of storage, and in the admin panel I went to buy more storage to match dropbox and it says it costs .96/Gb per year. That seems to indicate that it would costs $13,440 to match Dropbox's $700 plan?

I looked at Google non profit website and it says the free non profit account get 100Tb. To make microsoft match that would cost almost 100k a year compared to totally free with Google?

Am I missing something here?

Dropping off a donation and popped inside a thrift store. Found a brand new looking Panasonic KXP 1080i dot matrix printer for $15. Last saw one like that in 1992 so I bought it for no reason whatever. Splashed another $10 for a USB-Centronics adapter. When I heard the song of the dot-matrix again I was inexplicably filled with joy. BTW, tractor feed paper has gotten EXPENSIVE and greenbar is almost unobtainable in 9.5x11.

I'm bringing it to the office today to share my happiness.

Update: total whiff on the USB adapter, this IEEE1284 USB adapter is incompatible with the ooold SPP parallel port. It would fail during the negotiation and reset. I ordered an Epsonnet print server off fleabay that I'll direct run to an empty ethernet port on my box. Now I have to get it working, my pride is on the line.

I've been noticing a trend in various organizations, and I'm curious if you've seen it from the trenches: The default assumption that "AI will fix it" for virtually any business challenge, often before anyone has truly drilled down to the root cause.

It's like management hears "AI," and suddenly every structural inefficiency, communication breakdown, or outdated manual process gets rebranded as an AI opportunity. The risk isn't just wasted budget on complex solutions for simple problems; it's also diverting attention from critical systemic issues that a shiny new algorithm can never truly solve. You end up with an AI layer on top of a broken foundation, and guess who has to make it work?

I'm thinking about the value of having a go-to resource – essentially, a curated directory of proven, non-AI operational workflows and systemic solutions for common business problems. Something indexed by problem type, offering a baseline comparison. The idea is to help quickly identify if an issue is genuinely a structural inefficiency needing a process overhaul, rather than just a lack of automation.

Have you experienced this "AI-first" pressure? What are your thoughts on needing more accessible, peer-reviewed non-tech solutions to counter the hype and ensure we're solving the right problems effectively?

Disclosure first because the mods will catch it: I wrote this, MIT licence, link at the bottom, no upsell.

School IT background, but the use case generalises to any small-to-mid-size M365 estate (schools, small businesses, charities, roughly 20-500 users) wanting proper drive-letter mapping for OneDrive and SharePoint where the OneDrive sync client isn't the right shape. Wanted a drive-letter sync setup with a few specific things: image thumbnails that actually render in Explorer for cloud-mounted files (most existing tools don't), accurate cloud-quota numbers in the drive-properties dialog (refreshed from Graph), a local-first design where files always exist as real NTFS first and sync to the cloud in the background (rather than presenting the cloud as a virtual filesystem), one config.json that deploys to the whole estate, and zero telemetry. Didn't find one off the shelf with all of those, and was also curious whether I could build it myself. Side project that turned into a real thing.

It's called OneSync. .NET 8, Dokan for the FS bridge, MSAL + Graph SDK for auth and sync. Self-contained exe, Intune-deployable.

Highlights:

- True drive letters via Dokan (not subst, not "OneDrive sidebar entry under TenantName - LibraryName")
- One config.json deploys to the whole estate. OneSync probes Graph at startup and silently skips drives the signed-in user can't access, so different roles (staff/students, sales/finance/ops, whatever your split is) share one config
- Accurate cloud quota shown in Explorer drive properties (refreshed from Graph)
- Image thumbnails (JPG/PNG/HEIC) render like a local disk
- LRU eviction below a free-space threshold so shared/hot-desk laptops don't fill up over time
- Office desktop AutoSave + co-auth on docx/xlsx/pptx (via file-association redirect to ms-word:ofe with the direct SharePoint URL, slightly cursed but works)
- Zero telemetry. No analytics, no licence-server pings, no anonymous usage stats. Only outbound calls are to graph.microsoft.com and login.microsoftonline.com.

One war story while I'm here: Dokan's FindFiles callback enters synchronously from the kernel, so anything that blocks in there hangs Explorer. The first version called Graph synchronously to enumerate folders via .GetAwaiter().GetResult() (I know, I know). Worked fine until Graph went into 429 cooldown, at which point opening a folder hung Explorer for ten minutes while the cooldown lifted. I learned a lot in those ten minutes about how Explorer's UI thread feels under those conditions. Current behaviour: if Graph is in cooldown, folder browse returns immediately with on-disk placeholders and file hydration kicks off via Task.Run. Obvious in hindsight.

Repo: https://github.com/madeyouclickstudio/OneSync (DEPLOYMENT.md covers silent install switches and Intune detection rules)

Honest caveat: I'm running it with a handful of pilot users, not a full estate. The fleet-scale stuff is engineered for it but not battle-tested. Anyone deploying this to thousands of users is a pioneer at their own risk. Bug reports very welcome.

Yes I know about the OneDrive sync client. We tried it. The reasons it didn't fit (no real drive letters, every library is its own sidebar entry, no way to cap disk usage on shared laptops, no central deployment for which libraries each user sees) are exactly the reasons this exists. If those don't bother your fleet, OneDrive sync client is fine and you should ignore me.

Tell me what's wrong with it.

Hi everyone,

Our IT team is currently working on integrating Copilot to assist our accounting team. The primary use case will be summarizing Excel spreadsheets and performing calculations.

I would appreciate any advice or recommendations on what we should consider during implementation, particularly regarding permissions, policies, SLAs, and data protection.

Are there any legal risks we should be aware of?

Our main concern is ensuring that our data remains secure. We cannot risk any client information being exposed or used to train AI models. Is there a way to ensure that all data remains strictly within our tenant? While we understand that 100% isolation may not be achievable, we would like to know what steps we can take to maximize security and minimize risk.

Any guidance would be greatly appreciated.

Fifteen years in IT. This is a new one.

Guy comes in last week, been on holiday for a month, came back and couldn't remember his password. Tried a few times, locked himself out, called us. Fine, happens. We start the recovery process.

Midway through he mentions he actually has the password written down somewhere. Good i say, where?

Notepad file. Desktop.

I stopped typing for a second.

Me: so the password is saved in a file you can only open once you're already logged in?

Him: ...yes.

He got it immediately to be fair. Didn't try to argue about it or anything, just kind of sat there processing it. We finished the recovery, set a new password, i made him repeat it back to me twice before he left.

On his way out he said he'd write it down properly this time. I asked where. He said probably another Notepad file but in a folder this time so it's more organized.

I said great and wished him a good day becuase there's nothing else you can do really.

Genuinely though. Paper on the fridge at home. Nobody has ever been hacked through their fridge. Zero recorded incidents. Completely air gapped. I recommend it to everyone and i'm only half joking.