On Wed, Oct 24, 2018 at 02:17:52PM +0200, Martijn van Duren wrote: > A couple of machines at my $DAYJOB produce some weird behaviour when > using ftp(1). The connection seems to stall and ftp hangs indefinitely. > This mostly happens in the cron check for syspatch.
Send a netstat -an output. In what state is the TCP connction? Is it only a pf problem? Does pfctl -d fix the problem after it has occured? Does pfctl -d prevent the problem when you start the ftp program? Does it happen after TCP has transfered some data or at the beginning? How does your pf rule look like that creates the state. Do you have a default block rule? Does a "pass no state" fix it? A wild guess would be that some TCP reset has killed the pf state but not the TCP socket. Do you see any reset packets in tcpdump? bluhm
