Built-in stratum server serves coinbase-only jobs (mempool txs never included) #53
Labels
No labels
do-not-open-till-X-Day
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
SubGeniusFinance/dobbscoin-source#53
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Version: dobbscoin-qt 0.14.1 (protocol 70005), Windows 10
What happens
Every job from the built-in stratum server logs
0 branch, meaning the template it sends to miners contains only the coinbase. This happens even while the mempool holds a valid transaction that pays well above the minimum fee.Expected
The stratum template should include mempool transactions.
getblocktemplateon the same node at the same moment does list the pending tx, so only the stratum job builder drops them.Evidence (2026-11-14)
ce91a7ba322b848d493a2436c6556745565ace40a925a07c35f43dd194cb64a9(23 inputs, 4,208 bytes, fee 0.00025094, about 0.00006/kB) sat in the local mempool from 20:26 to 21:08 UTC (getmininginfo→pooledtx: 1).debug.logduring that window shows0 branch, e.g.:stratum: job 00003c8e (prev eaf331a9782a93ad.., nbits 1d267f9b, ntime 6af8cd8b, 0 branch)getmininginforeportedcurrentblocktx: 0the whole time.getblocktemplateat the same time returned the tx intransactions.Impact
Solo miners using the wallet's stratum server lose all fee revenue, and their blocks never help confirm pending transactions.
Likely area
The stratum job builder probably makes its own template, or reuses a cached one from before the mempool had entries, instead of calling the same
CreateNewBlockpath asgetblocktemplate. It may also be building the merkle branch from coinbase-only data.Traced this on the pool's node log. Its block templates were coinbase-only (
CreateNewBlock(): total size 1000) right up to 21:08:08 UTC, when they grew to 5208 bytes, exactly 1000 plus this 4,208-byte transaction, 54 seconds after the manualsendrawtransaction. At 21:02:21 it did include another, smaller transaction. So the pool includes mempool transactions; its node simply never received this one until the re-send. The0 branchlines come from the wallet's pool client logging the jobs the pool sends. This is the propagation problem in #54, seen from the pool's side, so I'll close it as a duplicate of #54.