Crash - consitent error about plexconnect HOT 20 CLOSED

What is client "19"? Why is it requesting "get_log.cgi"? Do you have some other server processes running? Other clients in the network that might get confused?

from plexconnect.

"Client 19" is secret code for "Client 192.168.0.123" with a line break inserted after 80 characters ;)

from plexconnect.

Ha. That is cool. Somehow the \n made my brain reset. :-D
So 192.168.0.123 is your aTV? I wonder why it requests those strange paths...

from plexconnect.

no... My ATV is is not 192.168.0.123.. it infact it is my XBMC box... which in theory has nothing to do with this. the XBMC box is the one serving the content SMB share.

I have no clue why it would be probing the plex server box.

Why would it cause plexconnect to stop working. Is there more info I can give you guys?

from plexconnect.

Don't know... this issue was reported ealier in the Plex forums.
So far we were unable to figure out the root cause.
Your setup - with XBMC box - might give some hints... once my brain get's up and running again :-)

from plexconnect.

LOL. I'll see what other clues I can give you....

It's either the fact that the media for my plex server instance does not reside on the local machine, or my xbmc box is doing something funny,

I've restarted plexconnect and not made any requests form my ATV, and the XBMC box .123 has not made any strange calls.

So I think it must have something to do with my plex media files not residing in the same server...

anyways.... 'll fool around a bit..

PS... this thin you have created here is revolutionary for people with ATV3 !!! My hat is off!!

from plexconnect.

I've looked into this a little further...

I don't think the crash has anything to do with the random requests from my xbmc box.

I am running the plex server from my laptop (it's just for testing at the moment). It is only connected to the rest via wireless 802.11n. I am sure the crash comes after connection is lost (network disconnect or something). Even though plexconnect and plex media server is on the same machine. plexconnect detects the media server on 192.168.0.6 (the IP for my laptop). This is correct, but the arp/routing table will change the moment the wireless disconnects.

I have two questions and some suggestions.

1. Does plexconnect keep an open socket connction to the plex server, or does it connect upon each request?
2. Is there some re-connect funtionality should it loose conection?

So my suggestions would be as follows.

1. First test if plex media server can be reached via 127.0.0.1 this would bypass any dependency on network.
2. Dont try and keep any sockets open for long period of time. They wont stay there..rather connect, do, disconnect. or verify the socket is still valid on each request, and gracefully handle disconnects.

That's my two cents worth.

Regards

from plexconnect.

Port to PMS is re-newed everytime it is needed, I guess. Handled in WebServer/XMLConverter.
DNS Port (incoming) is always open and the same. We don't want to loose requests.
DNS Port (outgoing) for forwarding is also open all the time. I did try to open it only on demand, but it didn't solve those DNS issues either.

from plexconnect.

What about changing the following in DNSServer.py on line no. 208 / 209

change from:
except socket.timeout:
pass

change to:
except socket.error:
pass

At least then the DNS server will continue to run rather than shutting down when [Errno 10054] is thrown?

from plexconnect.

yes, we could do something like that - though I would print at least a "warning" when there were a buffer overrun. On the other hand, that only hides - not solves the issue...

from plexconnect.

Hi guys. Just a stupid question. Is it the dns server(plexconnect) throwing the error or the connection to the upstream dns?

from plexconnect.

first msg:
)) File "C:\plexconnect\DNSServer.py", line 166, in Run
)) data, addr = DNS.recvfrom(1024)
This seems to be the aTV request into PlexConnect/DNSServer.

from plexconnect.

Looking at the code in DNSServer.py.

I agree that the
except socket.timeout: should change. I have changed it in mine for testing purposes.

At least this will capture the error. and stop it from crashing... Will it also catch any timeout errors?

will test.

from plexconnect.

put these lines in... Think this is the best option right now.
except socket.timeout:
pass
except socket.error as e:
dprint(name, 1, "WARNING! DNS error({0}): {1}".format(e.errno, e.strerror))

from plexconnect.

never coded in Python before...sorry...hope it is right.

from plexconnect.

It seems right to me to catch the error. You can use nice formatting in these comments if you want:

except socket.timeout:
    pass
except socket.error as e:
    dprint(__name__, 1, "WARNING! DNS error({0}): {1}".format(e.errno, e.strerror))

from plexconnect.

@elan - That would work nicely.

Catching the error resolved the problem locally for me.

from plexconnect.

Will make the comments pretty next time!

Implemented the catch and seems to be working. No crash.

from plexconnect.

)) except socket.timeout: should change
the timeout is by design - at least on the aTV incoming side. It will stay there without action ('pass').
I will add the except socket error + warning...

from plexconnect.

done.

from plexconnect.