fix: fix the hasrate loss when running cm on 2.9.5 alpha#852
Conversation
| ar_mining_io_tests, | ||
| ar_mining_worker_tests, | ||
| ar_poller_tests, | ||
| ar_reject_chunks_tests, |
There was a problem hiding this comment.
Is it normal to remove this test from the suite?
There was a problem hiding this comment.
It wsa a duplicate registration. Also exists in the long running tests section
| ar_test_node:post_chunk(main, ar_serialize:jsonify(FirstProof3)) | ||
| ), | ||
| ar_test_node:mine(peer1), | ||
| assert_wait_until_height(main, 1), |
There was a problem hiding this comment.
it turns out we ma not need it. I added it because the loop afterwards is checking the status of the tx on th main node but the block had been mined on the peer1 node so without the wait it's possible the block doesn't get gossipped to main before the time out occurs.
That said, adding the wait did not ddress the core issue. I can remove it if needed.
| @@ -525,8 +525,7 @@ handle_task({compute_h2_for_peer, Candidate, _ExtraArgs}, State) -> | |||
| ar_mining_stats:h1_received_from_peer(Peer, length(H1List)), | |||
| %% Mark second recall range as missing, just like if we didn't have the recall range | |||
There was a problem hiding this comment.
These comments are outdated, right?
JamesPiechota
force-pushed
the
jp/fix-cm-hashrate-20250819
branch
from
ae8a3e4 to
fcae200
Compare
| %% After we marked the whole second recall range as missing, we can cache the H1 list. | ||
| %% During this process, we also reset the chunk2_missing flag to false for the entries | ||
| %% we have H1 for. | ||
| %% First we mark the whole second recall range as missing |
There was a problem hiding this comment.
I am confused, do we mark the second or the first recall range as missing?
There was a problem hiding this comment.
Sigh. first recall range. I'd done the comment update before having coffee and incorrectly said "second"
Ive updated it again!
Also fixes the cache_limit_exceeded error spam
Since we're posting chunks to main we should wait on main to process the recently mined block
JamesPiechota
force-pushed
the
jp/fix-cm-hashrate-20250819
branch
from
fcae200 to
58d1fae
Compare
3 participants
Also fixes the cache_limit_exceeded error spam