The timeout is due to outstanding txrequests with python peers, which have the same timeout (60s) as the mempool sync timeout.
I explained this in more detail in #31721 (comment) and also mentioned there how to reproduce it.
Fix this by disconnecting the python peers after they send their txns, they aren’t needed after this point anyway because the main goal of the test is the sync between the 4 full nodes.
Fixes #31721
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.
For details see: https://corecheck.dev/bitcoin/bitcoin/pulls/31751.
See the guideline for information on the review process.
| Type | Reviewers |
|---|---|
| ACK | instagibbs, marcofleon, glozow, achow101 |
If your review is incorrectly listed, please react with 👎 to this comment and the bot will ignore it on the next update.
143@@ -144,6 +144,10 @@ def run_test(self):
144 for tx in transactions_to_presend[i]:
145 peer.send_and_ping(msg_tx(tx))
146
147+ # Disconnect python peers to clear outstanding orphan requests with them, avoiding timeouts.
0 # Disconnect python peers to clear outstanding orphan requests with them, avoiding timeouts.
1 # We are only interested in the syncing behavior between real nodes.
ACK 215d8b0a2ccf3a7275ea483339c58da5e0085e83
I suggested a comment change to leave more breadcrumbs for readers, but the fix seems correct
The timeout is due to outstanding txrequests with
python peers. Fix this by disconnecting these peers
after they send their txns, they aren't needed after
this point anyway.
ACK 152a2dcdefa6ec744db5b106d5c0a8c5b8aca416
iiuc, when we disconnect, this clears any 60s requests for parents that node1 has with the initial setup peers. So now there’s no chance the sync_mempools() timeout is hit before the parent request timeout.
