You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When a peer has been discovered and appears in the remotes of the "discoveree", it stays there for a long time (I can't tell if "forever", or just a very long time), well after the peer has closed/crashed/ended.
The suggested method of using last_used to filter out peers discovered some time ago doesn't work:
It seems last_used field (not called last_seen) is only updated when data transfer between peers takes place, not when that peer is discovered, or "re-discovered". It returns None in discovered peers in my tests, and that seems to match the docs description also.
So, I think we'd either need some other method of expiring peers (mDNS has a TTL setting when registering a service), or this field would need to be redefined, and updated each time the peer is discovered/confirmed-alive. But I'd argue that of the TTL expires and we try to confirm-alive and the peer is not alive, then it should be removed from the remotes - and updating the last_used would not be useful.
The text was updated successfully, but these errors were encountered:
When a peer has been discovered and appears in the remotes of the "discoveree", it stays there for a long time (I can't tell if "forever", or just a very long time), well after the peer has closed/crashed/ended.
The suggested method of using last_used to filter out peers discovered some time ago doesn't work:
Docs
https://docs.rs/iroh-net/latest/iroh_net/endpoint/struct.RemoteInfo.html
It seems
last_used
field (not calledlast_seen
) is only updated when data transfer between peers takes place, not when that peer is discovered, or "re-discovered". It returnsNone
in discovered peers in my tests, and that seems to match the docs description also.So, I think we'd either need some other method of expiring peers (mDNS has a TTL setting when registering a service), or this field would need to be redefined, and updated each time the peer is discovered/confirmed-alive. But I'd argue that of the TTL expires and we try to confirm-alive and the peer is not alive, then it should be removed from the remotes - and updating the
last_used
would not be useful.The text was updated successfully, but these errors were encountered: