When a peer is resumed from file there is no way to know which of its peers are still active or still have the same connection info. First, the restored peer pings a set of peers with the most promising connection statistics and waits for their responses if any. If none it keeps pinging peers further down the list until it runs out of options. If it runs out it attempts to bootstrap itself again. If however the first set of peers are still connected then it goes about querying normally. Only replacing peers as necessary.
When a peer is resumed from file there is no way to know which of its peers are still active or still have the same connection info. First, the restored peer pings a set of peers with the most promising connection statistics and waits for their responses if any. If none it keeps pinging peers further down the list until it runs out of options. If it runs out it attempts to bootstrap itself again. If however the first set of peers are still connected then it goes about querying normally. Only replacing peers as necessary.