Flaky test: tests::signer::v0::multiple_miners_empty_sortition

[obycode]

Sample logs here.

On quick investigation, it appears that both miners are saying that they will not extend their tenures prior to the timeout:

INFO [1738255479.028965] [testnet/stacks-node/src/nakamoto_node/relayer.rs:1411] [relayer-http://127.0.0.1:47211] Relayer: will not extend tenure -- we won sortition b6f053c9e6d02533582c288890da9a1f3ed17ddf, but the highest valid sortition is 2675cd0f6486580ff3870e7ef4c213ac22a3dee4

INFO [1738255479.580336] [testnet/stacks-node/src/nakamoto_node/relayer.rs:1390] [relayer-http://127.0.0.1:11691] Relayer: Did not win the last sortition that commits to our Stacks fork. Cannot continue tenure.

[kantai]

Sort of duplicate issue (#5777).

Logs are here.

My logs seem to indicate a different failure:

2025-01-30T16:08:43.5655808Z thread 'tests::signer::v0::multiple_miners_empty_sortition' panicked at testnet/stacks-node/src/tests/signer/v0.rs:11748:9:
2025-01-30T16:08:43.5659720Z assertion failed: !cur_empty_sortition.was_sortition
2025-01-30T16:08:43.5660276Z stack backtrace:

[obycode]

Duplicate (#5777)

Oops, I missed that one.

[kantai]

In my logs, at least, the issue seems to be that the test thinks the last bitcoin block has been processed when it queries for the latest sortition (in the logs, the latest bitcoin block is 234, but 234 hasn't been processed yet). That has a simple-ish fix:

diff --git a/testnet/stacks-node/src/tests/signer/v0.rs b/testnet/stacks-node/src/tests/signer/v0.rs
index feb870143f..5aca9c618a 100644
--- a/testnet/stacks-node/src/tests/signer/v0.rs
+++ b/testnet/stacks-node/src/tests/signer/v0.rs
@@ -11733,13 +11733,17 @@ fn multiple_miners_empty_sortition() {
         // lets mine a btc flash block
         let rl2_commits_before = rl2_commits.load(Ordering::SeqCst);
         let rl1_commits_before = rl1_commits.load(Ordering::SeqCst);
+        let info_before = get_chain_info(&conf);
+
         signer_test
             .running_nodes
             .btc_regtest_controller
             .build_next_block(2);
 
         wait_for(60, || {
-            Ok(rl2_commits.load(Ordering::SeqCst) > rl2_commits_before
+            let info = get_chain_info(&conf);
+            Ok(info.burn_block_height >= 2 + info_before.burn_block_height
+                && rl2_commits.load(Ordering::SeqCst) > rl2_commits_before
                 && rl1_commits.load(Ordering::SeqCst) > rl1_commits_before)
         })
         .unwrap();

[kantai]

Okay, I'm pretty sure I know why your logs' flake happens: the "flash block" occurs across a reward cycle boundary: the miner can't extend from one reward cycle to the next.

[kantai]

I think this is resolved now (by #5784)

[blockstack-devops]

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.