[GH-ISSUE #358] UT MgmtdClientTest::testRetryUnknownAddrs failed #219

Open
opened 2026-10-26 09:47:16 +00:00 by bsakaguchi · 1 comment
Owner

Originally created by @songqing on GitHub (Oct 29, 2025).
Original GitHub issue: https://github.com/deepseek-ai/3FS/issues/358

[----------] 1 test from MgmtdClientTest
[ RUN      ] MgmtdClientTest.testRetryUnknownAddrs
[2025-10-29T10:19:55.193599116+00:00 test_client:70442 MgmtdClient.cc:509 ERROR] MgmtdClient: probePrimary NodeId(0) failed: RPC::ConnectFailed(2014)
./tests/client/TestMgmtdClient.cc:545: Failure
Expected equality of these values:
  _actual
    Which is: { ("GetPrimaryMgmtd", "TCP://192.168.0.3:8800"), ("GetPrimaryMgmtd", "TCP://192.168.0.4:8888"), ("GetRoutingInfo", "TCP://192.168.0.4:8888") }
  _expected
    Which is: { ("GetPrimaryMgmtd", "TCP://192.168.0.4:8888"), ("GetRoutingInfo", "TCP://192.168.0.4:8888") }
[2025-10-29T10:19:55.193769060+00:00 test_client:70442 MgmtdClient.cc:329 ERROR] MgmtdClient: ClientSessionPayload not set, can't extend ClientSession
[2025-10-29T10:19:55.194080616+00:00 test_client:70442 MgmtdClient.cc:203 ERROR] [MgmtdClientOp ExtendClientSession No.2] failed: InvalidArg(3) ClientSessionPayload not set. latency: 0ns
[2025-10-29T10:19:55.194103974+00:00 test_client:70442 MgmtdClient.cc:313 ERROR] MgmtdClient: AutoExtendClientSession failed: InvalidArg(3) ClientSessionPayload not set
[  FAILED  ] MgmtdClientTest.testRetryUnknownAddrs (3 ms)
Originally created by @songqing on GitHub (Oct 29, 2025). Original GitHub issue: https://github.com/deepseek-ai/3FS/issues/358 ``` [----------] 1 test from MgmtdClientTest [ RUN ] MgmtdClientTest.testRetryUnknownAddrs [2025-10-29T10:19:55.193599116+00:00 test_client:70442 MgmtdClient.cc:509 ERROR] MgmtdClient: probePrimary NodeId(0) failed: RPC::ConnectFailed(2014) ./tests/client/TestMgmtdClient.cc:545: Failure Expected equality of these values: _actual Which is: { ("GetPrimaryMgmtd", "TCP://192.168.0.3:8800"), ("GetPrimaryMgmtd", "TCP://192.168.0.4:8888"), ("GetRoutingInfo", "TCP://192.168.0.4:8888") } _expected Which is: { ("GetPrimaryMgmtd", "TCP://192.168.0.4:8888"), ("GetRoutingInfo", "TCP://192.168.0.4:8888") } [2025-10-29T10:19:55.193769060+00:00 test_client:70442 MgmtdClient.cc:329 ERROR] MgmtdClient: ClientSessionPayload not set, can't extend ClientSession [2025-10-29T10:19:55.194080616+00:00 test_client:70442 MgmtdClient.cc:203 ERROR] [MgmtdClientOp ExtendClientSession No.2] failed: InvalidArg(3) ClientSessionPayload not set. latency: 0ns [2025-10-29T10:19:55.194103974+00:00 test_client:70442 MgmtdClient.cc:313 ERROR] MgmtdClient: AutoExtendClientSession failed: InvalidArg(3) ClientSessionPayload not set [ FAILED ] MgmtdClientTest.testRetryUnknownAddrs (3 ms) ```
Author
Owner

@songqing commented on GitHub (Oct 31, 2025):

As mgmtd_server has 2 addresses and add is visited first but failed, then visit adj and success.
So, visitRecords has 2 records not 1.
config.set_mgmtd_server_addresses({adj, add});

[2025-10-31T03:32:40.911042408+00:00 test_client:303689 MgmtdClient.cc:222 INFO] MgmtdClient: started
[2025-10-31T03:32:40.911207605+00:00 test_client:303689 MgmtdClient.cc:388 DEBUG] MgmtdClient: receive RefreshItem force=false
[2025-10-31T03:32:40.911240390+00:00 test_client:303689 MgmtdClient.cc:390 INFO] [MgmtdClientOp RefreshRoutingInfo force=false No.1] start execution. from:
[2025-10-31T03:32:40.911399533+00:00 test_client:303689 MgmtdClient.cc:424 INFO] MgmtdClient: start probe for connecting ...
[2025-10-31T03:32:40.911430657+00:00 test_client:303689 MgmtdClient.cc:542 INFO] MgmtdClient: start probe from unknown address TCP://192.168.0.3:8800
[2025-10-31T03:32:40.911471949+00:00 test_client:303689 MgmtdClient.cc:504 INFO] MgmtdClient: probePrimary NodeId(0) via TCP://192.168.0.3:8800 ...
[2025-10-31T03:32:40.911532893+00:00 test_client:303689 MgmtdClient.cc:509 ERROR] MgmtdClient: probePrimary NodeId(0) failed: RPC::ConnectFailed(2014)
[2025-10-31T03:32:40.911592071+00:00 test_client:303689 MgmtdClient.cc:162 INFO] MgmtdClient: mark node as Skipped NodeId(0) ["TCP://192.168.0.3:8800"]
[2025-10-31T03:32:40.911635187+00:00 test_client:303689 MgmtdClient.cc:542 INFO] MgmtdClient: start probe from unknown address TCP://192.168.0.4:8888
[2025-10-31T03:32:40.911654049+00:00 test_client:303689 MgmtdClient.cc:504 INFO] MgmtdClient: probePrimary NodeId(0) via TCP://192.168.0.4:8888 ...
[2025-10-31T03:32:40.911693934+00:00 test_client:303689 MgmtdClient.cc:585 INFO] MgmtdClient: found new mgmtd NodeId(3) TCP://192.168.0.4:8888
[2025-10-31T03:32:40.911715384+00:00 test_client:303689 MgmtdClient.cc:532 INFO] MgmtdClient: probePrimary succeeded at NodeId(3)
[2025-10-31T03:32:40.911735521+00:00 test_client:303689 MgmtdClient.cc:428 INFO] MgmtdClient: found new primary NodeId(3)
[2025-10-31T03:32:40.911756808+00:00 test_client:303689 MgmtdClient.cc:181 DEBUG] MgmtdClient: execute RefreshRoutingInfo via NodeId(3) TCP://192.168.0.4:8888
[2025-10-31T03:32:40.911809169+00:00 test_client:303689 MgmtdClient.cc:183 INFO] MgmtdClient: RefreshRoutingInfo finished
[2025-10-31T03:32:40.911839242+00:00 test_client:303689 MgmtdClient.cc:612 INFO] MgmtdClient: get new routing info version 2
<!-- gh-comment-id:3471191751 --> @songqing commented on GitHub (Oct 31, 2025): As mgmtd_server has 2 addresses and add is visited first but failed, then visit adj and success. So, visitRecords has 2 records not 1. `config.set_mgmtd_server_addresses({adj, add});` ``` [2025-10-31T03:32:40.911042408+00:00 test_client:303689 MgmtdClient.cc:222 INFO] MgmtdClient: started [2025-10-31T03:32:40.911207605+00:00 test_client:303689 MgmtdClient.cc:388 DEBUG] MgmtdClient: receive RefreshItem force=false [2025-10-31T03:32:40.911240390+00:00 test_client:303689 MgmtdClient.cc:390 INFO] [MgmtdClientOp RefreshRoutingInfo force=false No.1] start execution. from: [2025-10-31T03:32:40.911399533+00:00 test_client:303689 MgmtdClient.cc:424 INFO] MgmtdClient: start probe for connecting ... [2025-10-31T03:32:40.911430657+00:00 test_client:303689 MgmtdClient.cc:542 INFO] MgmtdClient: start probe from unknown address TCP://192.168.0.3:8800 [2025-10-31T03:32:40.911471949+00:00 test_client:303689 MgmtdClient.cc:504 INFO] MgmtdClient: probePrimary NodeId(0) via TCP://192.168.0.3:8800 ... [2025-10-31T03:32:40.911532893+00:00 test_client:303689 MgmtdClient.cc:509 ERROR] MgmtdClient: probePrimary NodeId(0) failed: RPC::ConnectFailed(2014) [2025-10-31T03:32:40.911592071+00:00 test_client:303689 MgmtdClient.cc:162 INFO] MgmtdClient: mark node as Skipped NodeId(0) ["TCP://192.168.0.3:8800"] [2025-10-31T03:32:40.911635187+00:00 test_client:303689 MgmtdClient.cc:542 INFO] MgmtdClient: start probe from unknown address TCP://192.168.0.4:8888 [2025-10-31T03:32:40.911654049+00:00 test_client:303689 MgmtdClient.cc:504 INFO] MgmtdClient: probePrimary NodeId(0) via TCP://192.168.0.4:8888 ... [2025-10-31T03:32:40.911693934+00:00 test_client:303689 MgmtdClient.cc:585 INFO] MgmtdClient: found new mgmtd NodeId(3) TCP://192.168.0.4:8888 [2025-10-31T03:32:40.911715384+00:00 test_client:303689 MgmtdClient.cc:532 INFO] MgmtdClient: probePrimary succeeded at NodeId(3) [2025-10-31T03:32:40.911735521+00:00 test_client:303689 MgmtdClient.cc:428 INFO] MgmtdClient: found new primary NodeId(3) [2025-10-31T03:32:40.911756808+00:00 test_client:303689 MgmtdClient.cc:181 DEBUG] MgmtdClient: execute RefreshRoutingInfo via NodeId(3) TCP://192.168.0.4:8888 [2025-10-31T03:32:40.911809169+00:00 test_client:303689 MgmtdClient.cc:183 INFO] MgmtdClient: RefreshRoutingInfo finished [2025-10-31T03:32:40.911839242+00:00 test_client:303689 MgmtdClient.cc:612 INFO] MgmtdClient: get new routing info version 2 ```
Sign in to join this conversation.
No labels
pull-request
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/3FS#219
No description provided.