Block Rebroadcasting not occurring automatically - causing invalid blocks across the network

[aramsdale]

Some UDP block broadcasts appear to only be on the local ledger, and not broadcasted to the wider network. Additionally, when this occurs, the node never appears to 'recover' or rebroadcast the blocks in question. A local explorer search shows the block information, but taking that same hash to another node shows 'Invalid Block' or Not Found.

This appears to be more common on high-volume nodes, and potentially saturated routers, where the UDP broadcast may get dropped.

The current workaround is to perform a rebroadcast.

This block from 24 hours prior still shows on nanode's node:

But the same hash is invalid on nano.org explorer (500):

Steps to reproduce the issue:

1. Force the UDP broadcast to not be received by the larger network, either by transaction flooding, outage, etc
2. Check local ledger for hash, will show. Then check hash on outside ledger, won't appear.
3. Report it to Austin in #support that you're missing a transaction! :D

[argakiig]

This issue has not been observed on the network since multiple improvements were added to rebroadcast, will close this ticket after discussion with @aramsdale. If the issue resumes please reopen this issue