hi, we encount a problem that blocked at the `plain_types_UnderlayMessageDataWriter_write` function about 2 minute.
it will occur when we do the diff host transport, whatever the Udp and Tcp,
after the remote peer leave（be killed）, continue to using `plain_types_UnderlayMessageDataWriter_write` to write msg, it will block.
rtidds version: 6.1.1
our code:
plain_types_UnderlayMessage m;
int64_t message_byte_size = message::ByteSize(msg);
DDS_UInt8Seq_ensure_length(&m.data, message_byte_size, message_byte_size);
RETURN_VAL_IF(
!message::SerializeToArray(msg, DDS_UInt8Seq_get_reference(&m.data, 0),
message_byte_size),
false);
{
std::lock_guard<std::mutex> lg(lock_);
if (data_writer_ != NULL) {
DDS_InstanceHandle_t instance_handle = DDS_HANDLE_NIL;
DDS_ReturnCode_t retcode = plain_types_UnderlayMessageDataWriter_write(
data_writer_, &m, &instance_handle);
if (retcode != DDS_RETCODE_OK) {
AERROR << "data_writer failed to write";
return false;
}
}
}
Best Regards.
Tang from xiaomi Company
and i have doing the same test in the same host with shm transport, it will not occur.
hi, we use the Asynchronous PublishMode to write
Hi,
Could you kindly contact me at zklim@rti.com?
There could be many reasons for blocking, we need to know more details about the situation to help you.
Thank you