Łukasz Magiera
Łukasz Magiera
Digging through the data there a few things I see: * We should give LAN address dials lower timeout * Most dials to these addresses just sit around taking dial...
So I just realized that there may be another problem: Example: * Start mass-dialing peers (basically any dht action) * hit a set of peers which will 'clog up' the...
> However, moving the dial timeout may make things worse: a single peer with a ton of addresses could clog the dialer forever. Not really as we allow only 8...
The default timeout is in libp2p-transport: https://github.com/libp2p/go-libp2p-transport/blob/c45cca89b916558b5aff9cb46d2d903b3cc78e3a/transport.go#L18
Note that this should be optional as it may impact DHT performance quite a bit
> SYN and RST storms This was reported in libp2p-swarm some time ago: https://github.com/libp2p/go-libp2p/issues/1550
So go-ipfs exposes libp2p streams via `ipfs p2p` for quite a while already, but it's not really useful for browsers since it requires use of native sockets. There was some...
(AFAIK it should be possible to implement on the extension level, cc @lidel)
This Go issue seems related: https://github.com/golang/go/issues/23037
Jenkins is broken in `ipfs-shipyard` for go projects, I'll try to work around it tomorrow if time permits