If vsmserver decides to abort the XML-RPC connection, for whatever reason, then thinlinc-login is currently very unable to handle this scenario: > 2024-01-08T10:06:47: ssh[E]: Traceback (most recent call last): > 2024-01-08T10:06:47: ssh[E]: File "/usr/bin/thinlinc-login", line 223, in <module> > 2024-01-08T10:06:47: ssh[E]: IIIi1111iiIi1 ( ) > 2024-01-08T10:06:47: ssh[E]: File "/usr/bin/thinlinc-login", line 200, in IIIi1111iiIi1 > 2024-01-08T10:06:47: ssh[E]: IIi1I1I1i . sendall ( Iii1I1I1 ) > 2024-01-08T10:06:47: ssh[E]: BrokenPipeError: [Errno 32] Broken pipe > 2024-01-08T10:06:47: ssh[E]: sys:1: ResourceWarning: unclosed <socket.socket fd=4, family=1, type=1, proto=0, laddr=b'\x000d642', raddr=/var/run/thinlinc/master/1000> In this case, I triggered the issue by sending an excessive amount of data. Crashing also means we don't report anything sane back to the client, which causes it to misbehave as well (bug 8284). We should probably report an HTTP error code in these cases.
Note that this also triggered the ResourceWarning from bug 8283.