server keeps rebooting after patch install failure

TL;DR

Repeated unplanned reboots after patching have three usual drivers: a patched boot-start driver crashing early, an update policy forcing restarts to finish a failing install, or a service recovery setting that reboots the box. Stabilize first: disable automatic restart on failure to capture the stop code, boot the last known good state or previous kernel, and read the crash or event evidence before re-applying anything.

server keeps rebooting after patch install failure

Use this when

• The server reboots repeatedly since its patch window, with short uptimes.
• Reboots happen without any user or admin logged in.
• The pattern started with a specific patch wave and affects similar hosts.

Not for this skill when

• The reboots follow a published maintenance schedule. That is policy working, not a failure.
• The facility had power events in the same window. Rule out power before blaming patches.

Steps

1. Stop the cycle: on Windows disable automatic restart on system failure and note the stop code; on Linux boot the previous kernel and check the watchdog timer. Verify: the machine halts at an error or stays up, instead of looping silently
2. Collect evidence: Windows Event Viewer system log and minidumps, or journalctl for the previous boots and kernel oops lines. Verify: the reboot trigger (bugcheck, oops, policy, watchdog) is identified from logs
3. Roll back the implicated change: uninstall the suspect KB or driver, or remove the suspect package version on Linux. Verify: the rollback completes on the stable boot
4. Reboot and monitor uptime through the period that previously failed. Verify: uptime exceeds the old failure interval with services healthy
5. Hold that patch for this hardware class and record it in the wave report. Verify: the hold is registered so the next wave skips it automatically

Variant phrasings

server reboot loop after patch bugcheck driver

The Windows root-cause phrasing.

linux server keeps rebooting after update watchdog oops

The Linux phrasing.

automatic restart after failed update policy loop

The policy-driven phrasing.

Compatibility: Compatibility: Windows Server and Linux servers; the stabilization steps differ by OS but the evidence order is the same.

Why it happens

A machine that reboots itself is destroying its own evidence each cycle, so the first job is stopping the loop safely. Patch waves trigger this when a boot-start component in the patch is incompatible with the hardware, or when restart-to-complete policies keep forcing attempts at an install that crashes the machine each time it progresses.

Edge cases / pitfalls

• In virtualized fleets, check the hypervisor event log too; a host-level reset looks identical to a guest reboot from inside.
• Kernel watchdog and IPMI watchdog timers can reboot a hung (not crashed) machine; a patch that makes boot slower than the watchdog timeout creates this pattern.
• Do not re-push the same wave to the rest of the fleet until one affected host yields a root cause.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst__Phu8Tp3eiNJVZOXCk5gSA