I'm running a bitcoin full node (v0.20.0) with an electrumx server (version 1.13.0) on top and I noticed that electrumx sometimes is veryyy slow to update itself with new transactions. What I am doing is: create a new signed transaction, broadcast it, make a CPFP transaction and broadcast it too. Then, I try to see if electrumx can detect the child transaction by querying electrumx for the scriptPubKeys of the parent transaction and see if anything is spending it. Around 2/3 of the times, electrumx updates quickly and can detect the child transaction but around 1/3 of the times it takes many minutes, like 15 minutes for it to detect and answer this child transaction. I have done this experiment like 9 times. I don't see any issues with my bitcoin node, its mempool contains the child transaction and when I run
bitcoin-cli getrawmempool true | jq '."parent_txid"'
I see the child txid in the "spentby" field.
Here is my electrumx.conf configuration:
DB_DIRECTORY = /media/pedro/Portable_2TB/Bitcoin/electrumx/data
DAEMON_URL = http://user:***@127.0.0.1
COIN = "BitcoinSegwit"
NET = "mainnet"
ALLOW_ROOT=1
PEER_DISCOVERY= self
SERVICES = rpc://localhost:8000,tcp://localhost:50001
INITIAL_CONCURRENT=500
COST_SOFT_LIMIT=10000000
COST_HARD_LIMIT=10000000
BANDWIDTH_UNIT_COST=1
NOTE: my bitcoin node gives a weird answer to `bitcoin-cli getchaintips` but looks good otherwise. Here is the answer:
>bitcoin-cli getchaintips
[
{
"height": 644185,
"hash": "00000000000000000002fc463486fa70cc80cecdb0ebdfd8f7e2f056cbeb929c",
"branchlen": 0,
"status": "active"
},
{
"height": 638519,
"hash": "00000000000000000011398b899fb881c61bd6e8670270fbc8e225c11fd95d26",
"branchlen": 1,
"status": "valid-fork"
}
]
I tried restarting the node and running 'bitcoin-cli reconsiderblock 00000000000000000011398b899fb881c61bd6e8670270fbc8e225c11fd95d26' but it didn't solve, could this be related? It's working fine otherwise it seems.