SysAdmin Heaven

Real infrastructure. Real failures. Real fixes.

SysAdmin Heaven

notes from running real infrastructure

Real infrastructure. Real failures. Real fixes. Notes on running, breaking, investigating and repairing systems across Windows, Linux, cloud, networking and everything in between.

Read the articles

I run servers, and I write down what breaks.

Most infrastructure problems are not hard because the concept is difficult. They are hard because the system tells you something that is technically true and completely misleading — a firewall rule that is correct but never matches, a config file that validates but describes the wrong thing, a fix that appears to work and quietly opens a hole.

Latest articles

All articles →
12 min read

The SCCM client that was running, registered, and ignoring every deployment

Why a GPO and a script on a fileshare quietly underperforms, and what to replace it with

A machine stops getting deployments. The console says Active. CcmExec is running. It registered this morning. Every check you would think to run comes back clean, and the machine has not installed anything in two months. Most shops end up with the same answer: a GPO and a remediation script on a fileshare. It half works, nobody entirely trusts it, and the manual triage never actually goes away. This is about why that happens and what to do instead.

Topics

Why this exists

Everything here describes a system that was actually built, actually broke, and was actually investigated and fixed. The commands ran. The logs are real. The dead ends are included, because the dead end is usually the part that would have cost you the afternoon.

Cross-platform by default — Windows and Linux, on-premises and cloud. The interesting problems rarely respect a vendor boundary.