[Spread-users] "Connected closed by Spread" on receiving

[John Schultz]

> By the way, after SP_multicast returns with success isn't the sender able to close the connection without problems anyway?

Not necessarily: 

http://ia600609.us.archive.org/22/items/TheUltimateSo_lingerPageOrWhyIsMyTcpNotReliable/the-ultimate-so_linger-page-or-why-is-my-tcp-not-reliable.html

Indeed, from reading that it might be that a peer calling close (i.e. - SP_disconnect) while data to receive is pending in its OS, then the OS may immediately send an RST to the server, which would cause it to report the error we saw in the logs.  

So, what I think might be happening here is that your sender sends his msg no problem, the daemon sends that msg to the receiver, but for some unknown reason around that time the receiver closes its socket while there is still data for it to receive in its OS, which causes its OS to immediately send an RST, which causes the server to get the error we saw in your logs.  In this theory, the underlying issue is your receiver client calling close somehow on its socket for some reason.

Maybe if you tell us more about your Spread version and OS environments that might shed some more light?

Cheers!

-----
John Lane Schultz
Spread Concepts LLC
Phn: 301 830 8100
Cell: 443 838 2200

On Apr 20, 2012, at 3:10 PM, Lisa Vitolo wrote:

Yes, after the sender sends the first message it waits for input from the user, while SP_disconnect is called only when it terminates.
By the way, after SP_multicast returns with success isn't the sender able to close the connection without problems anyway?

About the receiving mailbox, it's closed only when a SIGINT signal arrives.

Thanks for your patience :)

Il giorno 20 aprile 2012 20:56, John Schultz <jschultz at spreadconcepts.com> ha scritto:
That is strange.  The daemon's OS is claiming that the receiving client's socket was closed abruptly for some reason.

Are you 100% sure that the mailbox you are receiving on couldn't be closed somehow (e.g. - by the sender calling SP_disconnect)?

Cheers!

-----
John Lane Schultz
Spread Concepts LLC
Phn: 301 830 8100
Cell: 443 838 2200

On Apr 20, 2012, at 2:35 PM, Lisa Vitolo wrote:

Sorry! Gmail failed me and sent the message privately :D I'll repeat it here.

Thanks for the suggestion. It prints more useful debug messages now, but I still can't figure out what they really mean.

- The receiver connects:
Sess_accept: set sndbuf/rcvbuf to 204800
Setting TCP_NODELAY on socket 9
Sess_recv_client_auth: Client requested NULL type authentication
Sess_session_authorized: Accepting from 127.0.0.1 with private name get on mailbox 9
Sess_read: Message has type field 0x80010080
Sess_read: queueing message of type 8 with len 0 to the protocol <-- is "len 0" here the issue?

- The sender connects:
Sess_accept: set sndbuf/rcvbuf to 204800
Setting TCP_NODELAY on socket 10
Sess_recv_client_auth: Client requested NULL type authentication
Sess_session_authorized: Accepting from 127.0.0.1 with private name send on mailbox 10

- The sender sends the message:
Sess_read: Message has type field 0x80000081
Sess_read: queueing message of type 2 with len 0 to the protocol
Sess_read: Message has type field 0x800000a0
Sess_read: queueing message of type 16 with len 0 to the protocol
Sess_read: failed receiving header on session 9: ret -1: error: Connection reset by peer
Sess_kill: killing session get ( mailbox 9 )

Cheers,
Lisa

Il giorno 20 aprile 2012 20:17, John Schultz <jschultz at spreadconcepts.com> ha scritto:

What does the daemon output around the time your receiver is disconnected?

If nothing, then I would add the SESSION flag to your daemon's DebugFlags in its configuration file and re-run to try to see why the daemon is closing your receiver.

Cheers!

-----
John Lane Schultz
Spread Concepts LLC
Phn: 301 830 8100
Cell: 443 838 2200


--
They say a little knowledge is a dangerous thing, but it's not one half so bad as a lot of ignorance.

_______________________________________________
Spread-users mailing list
Spread-users at lists.spread.org
http://lists.spread.org/mailman/listinfo/spread-users


_______________________________________________
Spread-users mailing list
Spread-users at lists.spread.org
http://lists.spread.org/mailman/listinfo/spread-users




-- 
They say a little knowledge is a dangerous thing, but it's not one half so bad as a lot of ignorance.
_______________________________________________
Spread-users mailing list
Spread-users at lists.spread.org
http://lists.spread.org/mailman/listinfo/spread-users

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3805 bytes
Desc: not available
Url : http://lists.spread.org/pipermail/spread-users/attachments/20120420/72888f65/attachment.bin