Hi again:
The trouble is, the manual does not say what the predefined value for such
connections are.
The default value is 300 seconds. You're right, the manual shall mention it.
Something is not quite right. To the best of my knowledge (I am no Emacs guru),
I haven't configured tramp-connection-properties anywhere. However, its value
is reported as:
Its value is (("/sshfs:" "direct-async-process" t))
Original value was nil
If I do customize-variable, tramp-connection-properties, then it states "CHANGED
outside Customize".
In any case, 300 seconds are 5 minutes, but it only takes a few seconds to
timeout. These are the corresponding log entries in *Messages*:
Tramp: Timeout session /sudo:root@my-hostname:
Cleaning up the recentf list...
File /sudo:root@my-hostname:/etc/prometheus/prometheus.yml removed from the
recentf list
Cleaning up the recentf list...done (1 removed)
[...]
Tramp: Timeout session /sudo:root@my-hostname:
Cleaning up the recentf list...
File /sudo:root@my-hostname:/etc/prometheus/prometheus.yml removed from the
recentf list
Cleaning up the recentf list...done (1 removed)
More such messages keep appearing, even though I haven't modified the remote
file (any further) for a while.
Furthermore, most timeout messages do not appear in any of the Tramp debug
logs. This is the only one I have seen with some delay:
-----8<-----8<-----8<-----
21:22:39.639096 tramp-get-connection-property (7) # process-name nil; cache
used: nil
21:22:39.639407 tramp-get-connection-property (7) # locked nil; cache used: nil
21:22:39.640428 tramp-timeout-session (3) # Timeout session
/sudo:root@my-hostname:
21:22:39.641258 tramp-flush-connection-properties (7) # *tramp/sudo
root@my-hostname* (inline-decompress inline-compress remote-decoding
remote-encoding local-decoding local-encoding device scripts connected
remote-tty chunksize remote-shell session-timeout last-cmd-time password-vector
temp-file)
21:22:39.641684 tramp-get-connection-property (7) # prompt nil; cache used: nil
21:22:39.641957 tramp-process-sentinel (5) # Sentinel called: ‘#<process
*tramp/sudo root@my-hostname*>’ ‘killed
’
21:22:39.642321 tramp-flush-connection-properties (7) # *tramp/sudo
root@my-hostname* nil
21:22:39.642837 tramp-get-file-property (8) # / file-truename nil; inhibit: 10;
cache used: nil; cached at: nil
21:22:39.643383 tramp-flush-directory-properties (8) # /
21:22:39.644433 tramp-get-connection-property (7) # process-buffer nil; cache
used: t
21:22:39.644962 tramp-get-file-property (8) # / file-truename nil; inhibit: 10;
cache used: nil; cached at: nil
21:22:39.645245 tramp-flush-directory-properties (8) # /
21:22:39.645573 tramp-flush-connection-properties (7) # (tramp-file-name sudo
root nil my-hostname nil nil nil) (hg git bzr groups-string groups-integer
gid-string gid-integer uid-string uid-integer id stat readlink ~
case-insensitive perl perl-cwd-realpath perl-file-spec file-exists remote-shell
pipe-buf remote-path test locale uname process-buffer null-device)
21:22:39.648682 tramp-get-connection-property (7) # case-insensitive undef;
cache used: nil
21:22:39.649173 tramp-get-connection-property (7) # ~ nil; cache used: nil
-----8<-----8<-----8<-----
Note also the wrong new-line character after "‘killed".
The lack of timely connection timeout messages in the debug log makes debugging
harder, because the *Messages* buffer has no timestamps, so it is hard to see
what is causing trouble when.
Best regards,
rdiez
