0.4.0-beta2 stops downloading. Restart of SABnzbd needed.

[jcfp]

All primary servers, of which some were not functioning (no account, or trying out SSL which did not work on xlned.com). So the non-functioning primary server could cause the block? Interesting.

This is indeed reproducible by adding a bogus server to an otherwise completely working setup. For testing, I added example.org port 119 as a standard server (user/pass empty, 120s timeout, no ssl, 1 connection); there's no news server at that address and any attempt to connect will just time out. After doing this and restarting the program, sabnzbd stopped downloading completely; even though it was able to establish a connection to the other, working servers, nothing was ever downloaded. Instead, the logs show a repetitive cycle of connecting to all servers, timing out on the bogus one after 120 seconds (= timeout setting for all my servers) with the ValueError: need more than 1 value to unpack while downloading absolutely nothing. Restarting sabnzbd with the bogus server removed fixed the problem, although it did seem to take longer to pick up speed initially. Console logging and netstat output attached.

[nzb_leecher]

How did you add the changes ? Via editing the ini file or via one of the interfaces?

[jcfp]

How did you add the changes ? Via editing the ini file or via one of the interfaces?

Added the bogus server via the Default interface, then restarted the program. Removal was done manually by editing out the offending server from the ini file after shutting down sabnzbd.

[sander]

I recognize the "ValueError: need more than 1 value to unpack" from my logging.

And, even better: I can reproduce this, indeed by adding a non-functional FQDN: the downloads stops!

I now also see this on stdout (stderr?):

Code: Select all

17/Apr/2008:20:41:28 HTTP INFO Page handler: "The path '/sabnzbd/m/static/PlotKi
t/Plotkit.js' was not found."
Traceback (most recent call last):
  File "/home/sander/SABnzbd-0.4.0Beta2/cherrypy/_cphttptools.py", line 105, in
_run
    self.main()
  File "/home/sander/SABnzbd-0.4.0Beta2/cherrypy/_cphttptools.py", line 246, in
main
    page_handler, object_path, virtual_path = self.mapPathToObject(path)
  File "/home/sander/SABnzbd-0.4.0Beta2/cherrypy/_cphttptools.py", line 312, in
mapPathToObject
    raise cherrypy.NotFound(objectpath)
NotFound: 404

Not sure I it's caused by the non-functional newsserver.

[sander]

How did you add the changes ? Via editing the ini file or via one of the interfaces?

I've added it via the web interface. BTW: it takes a lot of time before the web interface returns with the new settings after changing them.

[jcfp]

The problem described in my post (Reply #15 on: April 17, 2008) still holds exactly as before on beta5, except that the python error is gone. When some server is unreachable, nothing is ever downloaded from any server and the log shows a repetitive cycle of connection attempts and subsequent failures/timeouts .

[shypike]

BTW: it takes a lot of time before the web interface returns with the new settings after changing them.

That's because of the soft restart. Especially when you use a large memory cache it will be flushed to disk.
Someday we will fix that soft restart.

[shypike]

When some server is unreachable, nothing is ever downloaded from any server and the log shows a repetitive cycle of connection attempts and subsequent failures/timeouts .

We didn't do anything about this problem yet.