ZMQ died silentlySee original GitHub issue
ZMQ silently died again without signal about disconnecting with node. Last signal I got on my node was from
2021-08-02 10:27:36,964 INFO f.a.e.b.b.zmq.ZMQActor - connected to tcp://192.168.2.7:29000
then I used it for a while and on 6th Aug node didn’t signal about ZMQ disruptions while Watchdog indicated eclipsing.
2021-08-06 12:39:20,703 WARN f.a.e.b.w.BlockchainWatchdog - blockcypher.com: we are 10 blocks late: we may be eclipsed from the bitcoin network 2021-08-06 12:54:10,281 WARN f.a.e.b.w.BlockchainWatchdog - blockstream.info: we are 13 blocks late: we may be eclipsed from the bitcoin network 2021-08-06 12:54:10,998 WARN f.a.e.b.w.BlockchainWatchdog - mempool.space: we are 13 blocks late: we may be eclipsed from the bitcoin network 2021-08-06 12:54:11,860 WARN f.a.e.b.w.BlockchainWatchdog - bitcoinheaders.net: we are 9 blocks late: we may be eclipsed from the bitcoin network 2021-08-06 12:54:20,448 WARN f.a.e.b.w.BlockchainWatchdog - blockcypher.com: we are 13 blocks late: we may be eclipsed from the bitcoin network
Restarting BITCOIND didn’t help that’s why I think ZMQ actor died.
- Created a year ago
- Comments:35 (35 by maintainers)
Top GitHub Comments
That’s also a good point, we only need an event and we’re only using the block hash (and just for logging purposes). We initially scanned the actual block, but that changed a while ago. I’ll prototype migrating to block hash to save bandwidth.
Btw, is block itself needed or do we only need a new block event (in which case block hash would suffice)?
ZmqWatcher and I don’t see stuff like scanning block contents, rather everything seems to be done by calling bitcoind api. Am I missing something?
If not then switching to block hashes can save traffic for setups with remote bitcoind.