Problem reading from createdump child_read_pipe: %s (%d)\n

Error message

Problem reading from createdump child_read_pipe: %s (%d)\n

What it means

In the createdump child process (PalCreateDump.cpp:253-260), after the parent signals readiness via the pipe, the child performs `read(child_read_pipe, &buffer, 1)`. If bytesRead != 1 (the parent did not write the expected single byte — e.g. it died, closed early, or the write failed) the child prints this with the errno and exits. It indicates the parent/child coordination handshake broke down.

Source

Thrown at src/coreclr/nativeaot/Runtime/unix/PalCreateDump.cpp:258

        {
            close(pipe_descs[i]);
        }
        return false;
    }
    else if (childpid == 0)
    {
        close(parent_read_pipe);
        close(parent_write_pipe);

        // Wait for prctl(PR_SET_PTRACER, childpid) in parent
        char buffer;
        int bytesRead;
        while((bytesRead = read(child_read_pipe, &buffer, 1)) < 0 && errno == EINTR);
        close(child_read_pipe);

        if (bytesRead != 1)
        {
            fprintf(stderr, "Problem reading from createdump child_read_pipe: %s (%d)\n", strerror(errno), errno);
            close(child_write_pipe);
            exit(-1);
        }

        // Only dup the child's stderr if there is error buffer
        if (errorMessageBuffer != nullptr)
        {
            dup2(child_write_pipe, STDERR_FILENO);
        }
        // Execute the createdump program
        if (execv(argv[0], (char* const *)argv) == -1)
        {
            if (errno == ENOENT)
            {
                fprintf(stderr, "DOTNET_DbgEnableMiniDump is set and the createdump binary does not exist: %s\n", argv[0]);
            }
            else
            {

View on GitHub (pinned to 290d5ab72c)

Solutions

Check the parent-side logs/stderr for the preceding failure (often the prctl or write error).
On Linux, ensure ptrace scope permits the operation (e.g. `/proc/sys/kernel/yama/ptrace_scope`).
Update to a current .NET build where the handshake is hardened.
If reproducible, capture strace of the parent to see why the byte was never written.

Example fix

# before: createdump spawned but parent died / blocked by ptrace_scope=3
# after: allow ptrace
sudo sysctl kernel.yama.ptrace_scope=2

Defensive patterns

Strategy: retry

Validate before calling

// In the parent, ensure the write completes; surface the real failure.
ssize_t w;
do { w = write(parent_write_pipe, "S", 1); } while (w < 0 && errno == EINTR);
if (w != 1) { fprintf(stderr, "parent write to createdump pipe failed: %s\n", strerror(errno)); }

Try / catch

// Around the createdump spawn in the parent; treat EINTR as retry, others as fatal.
for (;;) {
  rc = spawn_createdump(...);
  if (rc == 0) break;
  if (errno != EINTR) { log_and_disable_minidump(); break; }
}

Prevention

Relax ptrace_scope if Yama LSM blocks the parent/child ptrace handshake.
Keep .NET updated; the pipe handshake has been hardened over releases.
Capture stderr from the parent alongside createdump to see the originating failure.

When it happens

Trigger: The parent process crashing or being killed between fork and the `write(parent_write_pipe,...)`; the parent failing the prctl/ptracer step and closing the pipe; a race where the pipe is closed before the byte is written.

Common situations: The crashing process exits before createdump is fully primed; system restrictions on ptrace (Yama LSM) killing the parent; resource exhaustion closing file descriptors.

Related errors

AI-assisted analysis of dotnet/runtime@290d5ab72c (2026-08-06). Data as JSON: /api/errors/05a42149ff1d8e4a. Report an issue: GitHub.