WARNING: [Logger::LogVprintf] Failed to flush log file. GetL

Error message

WARNING: [Logger::LogVprintf] Failed to flush log file. GetLastError()=%u\n

What it means

In SuperPMI's shared Logger::LogVprintf, after a log line is written with WriteFile the logger calls FlushFileBuffers(s_logFile) to force it to disk. If the flush fails the logger only emits this WARNING with GetLastError() and continues; logging is best-effort and never throws. It signals that subsequent log lines may be buffered or lost.

Source

Thrown at src/coreclr/tools/superpmi/superpmi-shared/logging.cpp:294

        const char* timeStr = "";
#endif // TARGET_UNIX

        const char logEntryFmtStr[] = "%s - %s [%s:%d] - %s - %s\r\n";
        size_t logEntryBuffSize = sizeof(logEntryFmtStr) + strlen(timeStr) + strlen(function) + strlen(file) + /* line number */ 10 +
                                  strlen(logLevelStr) + strlen(fullMsg);

        char* logEntry = new char[logEntryBuffSize];
        sprintf_s(logEntry, logEntryBuffSize, logEntryFmtStr, timeStr, function, file, line, logLevelStr, fullMsg);
        size_t logEntryLen = strlen(logEntry);

        DWORD bytesWritten;

        if (!WriteFile(s_logFile, logEntry, (DWORD)logEntryLen, &bytesWritten, nullptr))
            fprintf(stderr, "WARNING: [Logger::LogVprintf] Failed to write to log file. GetLastError()=%u\n",
                    GetLastError());

        if (!FlushFileBuffers(s_logFile))
            fprintf(stderr, "WARNING: [Logger::LogVprintf] Failed to flush log file. GetLastError()=%u\n",
                    GetLastError());

        delete[] logEntry;

#ifndef TARGET_UNIX
        free((void*)timeStr);
#endif // !TARGET_UNIX
    }

#ifndef TARGET_UNIX
CleanUp:
#endif // !TARGET_UNIX

    minipal_mutex_leave(&s_critSec);
    delete[] fullMsg;
}

//

View on GitHub (pinned to 60108ba66e)

Solutions

Point SuperPMIShimLogPath at a local disk with ample free space.
Free disk space / raise quota on the log volume and restart the SuperPMI run.
Disable or move antivirus/backup that locks the log file during the run.
If warnings are non-fatal, ignore them but re-run on a healthy disk to get complete logs.

Defensive patterns

Strategy: validation

Validate before calling

// Before a SuperPMI run, check the log path is healthy.
#include <sys/statvfs.h>
static bool LogPathHealthy(const char* dir) {
    struct statvfs s;
    if (statvfs(dir, &s) != 0) return false;
    unsigned long freeMb = (s.f_bavail * (s.f_frsize ?: 1)) >> 20;
    return freeMb > 100; // require >100MB free
}

Prevention

Point SuperPMIShimLogPath at a local disk with generous free space.
Monitor disk usage during long SuperPMI runs; .mch files grow large.
Exclude the log directory from AV/backup scans that lock handles.
Treat the WARNING as non-fatal but re-run on a healthy disk for complete logs.

When it happens

Trigger: The SuperPMI log file lives on a volume that fails FlushFileBuffers: disk full/quota hit, removable/network drive disconnected, file truncated/deleted underneath the handle, or the handle became invalid after an I/O error. Windows reports the reason via GetLastError.

Common situations: Logging to a network share that dropped; /tmp filled by a large .mch method-context file while SuperPMIShimLogPath points at the same volume; antivirus/backup locking the log handle; running on a CI agent whose workspace is cleaned mid-run.

Related errors

AI-assisted analysis of dotnet/runtime@60108ba66e (2026-08-10). Data as JSON: /api/errors/5152b69175e7c59d. Report an issue: GitHub.