Gabriel39 commented on code in PR #70693:
URL: https://github.com/apache/doris/pull/70693#discussion_r4624787236
##########
be/src/exec/sink/vrow_distribution.cpp:
##########
@@ -136,13 +136,16 @@ Status VRowDistribution::automatic_create_partition() {
} else {
master_addr = _vpartition->get_master_address();
}
+ // Remote partition callbacks must use the same external policy as
their write channels.
int time_out = _state->execution_timeout() * 1000;
RETURN_IF_ERROR(ThriftRpcHelper::rpc<FrontendServiceClient>(
master_addr->hostname, master_addr->port,
[&request, &result](FrontendServiceConnection& client) {
client->createPartition(result, request);
},
- time_out));
+ time_out,
+ _remote_endpoint ?
ExecEnv::GetInstance()->remote_frontend_client_cache()
Review Comment:
Fixed in f842904af888d3aa3fb8ba32fa8f8ca5e6eca0a5. TabletSinkHashPartitioner
now initializes the callback policy from the existing optional remote master
address; no new Thrift field or FE change is needed. A native test covers
local/remote metadata and cloning, and a second test executes the exchange's
real createPartition RPC against a plaintext frontend while local private TLS
is enabled without local credentials. Both tests fail before the fix and pass
afterward; all 32 related ASAN tests pass.
The default NonConcurrent regression now runs with two local TLS BEs and one
remote plaintext BE. Fresh ASAN BE packaging, output generation and normal
replay pass (six result blocks, five expected refusals, nine helper tests, zero
skips). Full remote auto-partition INSERT SELECT also encounters existing
metadata/replica ID problems with TLS disabled, so I kept those unrelated
changes outside this transport PR and isolated the callback with the real-wire
test.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]