clockworklabs/SpacetimeDB · error

ERROR: Failed to read from BytesSource: %d\n

Error message

ERROR: Failed to read from BytesSource: %d\n

What it means

Runtime FFI error inside ConsumeBytes' unknown-length read loop: FFI::bytes_source_read returned a status other than 0 (OK) or -1 (EXHAUSTED) while draining the BytesSource that carries reducer/HTTP arguments across the Wasm boundary. The message is logged and the loop breaks, so the caller receives a truncated buffer and the failure usually resurfaces as a BSATN deserialization error downstream.

Source

Thrown at crates/bindings-cpp/src/internal/Module.cpp:417

    auto ret = FFI::bytes_source_remaining_length(source, &remaining_len);
    if (ret != 0) {
        // If we can't get the length, fall back to incremental reading
        // This shouldn't happen with current host implementation
        constexpr size_t CHUNK_SIZE = 1024;
        iter_buf.reserve(CHUNK_SIZE);
        
        while (true) {
            size_t chunk_size = CHUNK_SIZE;
            size_t old_size = iter_buf.size();
            iter_buf.resize(old_size + chunk_size);
            
            ret = FFI::bytes_source_read(source, iter_buf.data() + old_size, &chunk_size);
            iter_buf.resize(old_size + chunk_size);  // Resize to actual bytes read
            
            if (ret == -1) {  // EXHAUSTED
                break;
            } else if (ret != 0) {  // Error
                fprintf(stderr, "ERROR: Failed to read from BytesSource: %d\n", ret);
                break;
            }
        }
        return iter_buf.release();
    }
    
    // Reserve exact size needed (often no-op since pool buffer is 64 KiB)
    iter_buf.resize(remaining_len);  // Resize to exact size BEFORE reading
    
    // Read all bytes - should complete in one call since we have capacity
    size_t bytes_read = 0;
    while (bytes_read < remaining_len) {
        size_t chunk_size = remaining_len - bytes_read;
        ret = FFI::bytes_source_read(source, iter_buf.data() + bytes_read, &chunk_size);
        bytes_read += chunk_size;
        
        if (ret == -1) {  // EXHAUSTED
            break;

View on GitHub (pinned to 6dee26c6ef)

Solutions

1. Rebuild and republish the module with the same spacetimedb toolchain/CLI version used by the server so the FFI ABI matches
2. Check the server/host logs for the underlying read failure at the same timestamp - the module-side message only mirrors it
3. If it reproduces with a specific large argument, reduce the payload or split the call to rule out size-related truncation
4. If persistent, report with the returned status code (the %d value) to SpacetimeDB maintainers - nonzero non-minus-one statuses are host-side faults

When it happens

Trigger: The host-side byte stream faults mid-read while the module drains reducer arguments of unknown length; ABI/host version mismatch between the module binary and the spacetimedb host; corrupted or closed BytesSource passed across the FFI boundary.

Common situations: Publishing a module built with an older bindings-cpp ABI against a newer server (or vice versa); intermittent host I/O failures; truncation of large argument payloads at the Wasm boundary.

Related errors

• unable to decode args
• cannot deserialize refs without a typespace
• Invalid JSON: failed to parse string
• wrong number of elements
• too many elements

AI-assisted analysis of clockworklabs/SpacetimeDB@6dee26c6ef (2026-08-20). Data as JSON: /api/errors/8d84f59781c32441. Report an issue: GitHub.