Technical Anshu
Tutorials Blog
✅ Pre Deployment Checks

Pre Deployment Checks That Stop Failed Rollouts Disk Space Version and Running App

A failed rollout usually fails on facts that were knowable yesterday: the disk was nearly full, the old version was still there, the app was open. Pre-deployment checks turn those surprises into a small skipped list.

✍️ Technical Anshu🪟 Windows endpoints⏱️ 7 minute read
🎥 Watch pre-deployment activities being added to a package in the video

Endpoint Central packages support pre-deployment activities and conditions — things that must be true, or must happen, before the installer runs. Used well, they are the cheapest failure insurance in the whole deployment process: the machines that would have failed are skipped with a reason instead of failing with an error, and you fix them in daylight instead of during the rollout.

console — pre-deployment conditions on a package
Free disk spaceAt least 2 GB free Checked
Existing old versionUninstall first Handled
Application runningPrompt user to close
Machines skipped3 of 50 — listed with reasons
Three skipped machines with clear reasons beat three failed machines with cryptic errors — every time.
1

Disk Space — the Silent Killer of Big Installs

A half-installed office suite on a full disk is worse than no install at all: the old version is broken, the new one is incomplete, and the user has neither. Set a minimum free space condition sized to the package — the installer needs room for itself, its extracted files, and the installed result, roughly two to three times the installer size for large suites. PCs that fail the check are skipped and listed, and you will usually find they were overdue for a cleanup anyway.

  1. Check what else eats the disk first — runaway temp files and ancient downloads are the usual tenants
  2. A fleet-wide pattern of full disks is a hardware conversation, not a deployment problem
2

Version Present — Install Once, Upgrade Deliberately

  1. If the target version is already installed, decide what the deployment should do — skip it, or repair it — instead of letting the installer improvise
  2. If an older version is present, choose consciously: in-place upgrade, or uninstall-then-install as a pre-deployment activity. Some vendors support one and corrupt on the other
  3. Pull the version data from Endpoint Central's inventory before building the configuration, so the target group reflects reality
3

Running Application — Close It or Wait

Installers cannot replace files that are open. If the application you are updating is running — and for a browser or office suite, it always is — the deployment will fail, hang, or half-replace the program. Your honest options are: deploy after hours when the app is closed, notify the user to close it, or configure the package to check and skip machines where it runs. The one option that does not exist is hoping nobody has it open at 11 AM on a Tuesday.

4

The Smaller Checks That Save Specific Rollouts

  1. Pending reboot: a PC already waiting to restart for a previous install will stack problems; let it reboot first
  2. Operating system fit: a 64-bit-only package aimed at a forgotten 32-bit machine fails predictably — filter by inventory
  3. Power: laptops on battery are bad install candidates; prefer machines on mains for big packages
  4. Conflicting software: the old antivirus must leave before the new one lands — sequence removals before installs
5

Read the Skips as a Work List, Not a Shame List

After the rollout, the skipped machines are your follow-up list, each with its reason attached: free some disk, close the app, clear the old version. Clear them in small batches and re-deploy only to them. Over a few rollouts you will notice the same machines skip every time — that short list is your fleet's honest maintenance backlog, gift-wrapped by your own deployment process.

✅ Checks are not bureaucracy: every condition you add converts one future failure into one explained skip. Failed machines generate tickets and confusion; skipped machines generate a tidy list you can finish by Friday.

🎥 Watch the Full Deployment in the Video

See package conditions and pre-deployment activities in the full video.

▶️ Watch Full Video
❓ Frequently Asked Questions

What is a pre-deployment activity in Endpoint Central?

An action or check attached to a package that runs before the installation — for example uninstalling the previous version, checking free disk space, or running a script. It prepares the machine so the main install has the conditions it needs.

How much free space should I require before a big install?

A safe working rule is two to three times the installer size, because the installer file, its extracted contents and the installed program all exist at once. For small utilities a flat minimum like 2 GB free is plenty.

A PC got skipped by a check. How do I deploy to it later?

Fix the reason the check reported — free the disk, close the application, remove the conflicting version — then retry the configuration on that target. The check will pass and the install proceeds; there is nothing to undo.

This guide is based on the Technical Anshu video on remote software deployment with ManageEngine Endpoint Central • Stuck on a step? Ask in the video comments and you will get a reply 🙏