Hello jonatanencarnacion,
We completed a higher-fidelity laboratory test using three PowerShell Universal 2026.2.2 nodes, independent repository clones, one-way Git Sync, and a shared SQL database.
Across 25 Git synchronization cycles, all four test schedules remained registered, retained their Next Execution values, and completed 134 executions. We did not reproduce the schedule-registration loss.
We observed two transient behaviors for further analysis: one node processed shared jobs before its local Git configuration finished loading, and another node temporarily used Git settings associated with a different node. Neither behavior removed the schedules, so there is no confirmed root cause at this time.
If the issue occurs again, please avoid reloading the configuration or restarting the services until the following is collected from each node:
* Exact incident time and time zone
* Automation > Schedules and Automation > Jobs > Diagnostics screenshots
* Current Revision
* Git HEAD and `git status --short`
* SHA-256 and `LastWriteTimeUtc` of `.universal\schedules.ps1`
* PSU logs covering approximately 15 minutes before and after the incident
* Relevant Windows Application events for PSU, .NET, SQL, or Git
* The Git settings associated with each node, if visible
Please post only sanitized summaries and screenshots in the forum. Complete logs can be sent privately to:
ruben.tapia@ironmansoftware.com
Suggested subject:
F-55574 - Scheduler incident logs - <date and time>
Even by email, please remove passwords, tokens, private keys, license keys, cookies, and connection strings containing credentials.
Best regards,
Ruben Tapia
