azcomputerguru/claudetools
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

sync: auto-sync from HOWARD-HOME at 2026-06-21 13:31:49

Browse Source 
Author: Howard Enos
Machine: HOWARD-HOME
Timestamp: 2026-06-21 13:31:49
This commit is contained in:
Howard Enos
2026-06-21 13:32:50 -07:00
parent eeaaa566d7
commit 1c8fc09590
4 changed files with 39 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
.claude/memory/feedback_rmm_longops_fire_and_forget.md Normal file
View File

@@ -0,0 +1,20 @@
---
name: feedback_rmm_longops_fire_and_forget
description: Long-running RMM endpoint ops (software installs, big downloads) must be fire-and-forget, not live-monitored
metadata:
type: feedback
---
Long-running endpoint operations via RMM (software installs, large downloads,
BD/AV deployment) must run **unattended** — dispatched fire-and-forget, NOT
babysat with live polling loops. Howard's directive (2026-06-21): "this needs to
happen without us monitoring it."
**How to apply:** run the long op under the endpoint's **Task Scheduler** as
SYSTEM (`schtasks /Create ... /RU SYSTEM /RL HIGHEST /SC ONCE /F` then
`schtasks /Run`), or a detached process that logs its own result to a file. The
RMM command returns immediately. **Verify later by OUTCOME** (e.g. the endpoint
appearing in GravityZone inventory, a result/log file, a service check) on a
separate check — never sit in a poll loop for minutes. This mirrors how Syncro /
real RMM deployments work: fire the job, walk away, check status afterward.
See [[feedback_bitdefender_unattended_install]].