Flush logs/stdout/stderr on SIGTERM in client and plugin

[Copilot]

Plan: Add SIGTERM Signal Handling to Flush Logs

• Understand the current codebase structure (client, plugin, logging)
• Create signal handler function for SIGTERM that flushes logs
• Integrate signal handler into plugin (cmd/plugin.go)
• Integrate signal handler into client commands (cmd/object_copy.go and related files)
• Add optional SIGQUIT handler when Debug mode is enabled
• Write tests for signal handling functionality
• Build and test the changes
• Run code review and security checks
• Fix linter errors

Implementation Complete ✅

Changes Made:

1. Created Signal Handler Infrastructure (client/signal_handler.go, client/signal_handler_windows.go)

   • SetupSignalHandlers() function that listens for SIGTERM
   • Flushes logs using existing logging.FlushLogs()
   • Syncs stdout/stderr file descriptors
   • Optionally sends SIGQUIT in debug mode for stack traces
   • Cross-platform support (Unix/Windows)

2. Integrated Signal Handlers into All Entry Points

   • Plugin: cmd/plugin.go (stashPluginMain)
   • Client commands: All cmd/object_*.go files
     • object_copy.go, object_get.go, object_put.go
     • object_delete.go, object_ls.go, object_prestage.go
     • object_share.go, object_stat.go, object_sync.go

3. Comprehensive Testing (client/signal_handler_test.go)

   • All tests pass ✅

4. Security & Quality

   • CodeQL: 0 vulnerabilities ✅
   • Linter: All issues fixed ✅ (removed trailing whitespace)
   • Go vet: Passed ✅

Linter Fix:

Removed trailing whitespace from client/signal_handler_test.go to fix gofmt warnings.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Flush logs / stdout / stderr on SIGTERM</issue_title>
<issue_description>Pelican Service:

• Client
• Plugin
• Registry
• Director
• Origin
• Cache
• Other (please give the detail)

Is your feature request related to a problem? Please describe.
We are trying to track down a Pelican issue in the CHTC where 1% or so of execution attempts appear stuck for 1h before being killed by HTCondor. Unfortunately, HTCondor sends a SIGTERM and it appears that we lose any useful information because Pelican is only flushing logs / stdout / stderr upon successful completion.

Describe the solution you'd like
The Pelican client / plugin should flush all messages to logs, stdout, and stderr upon receiving a SIGTERM and then exit.

Optionally, it sounds like Go binaries have a SIGQUIT handler that dumps stack, so we may want the client / plugin to send a SIGQUIT to itself if say Debug: true

Additional context
Add any other context or screenshots about the feature request here.
</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes Flush logs / stdout / stderr on SIGTERM #2305

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[patrickbrophy]

@copilot It looks like the linter failed, can take a look?

[Copilot]

@copilot It looks like the linter failed, can take a look?

Fixed in fecb630. Removed trailing whitespace from signal_handler_test.go that was causing gofmt warnings.