camel-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Chandler <a...@riftware.com>
Subject Re: Mina blows up
Date Thu, 08 Apr 2010 21:33:16 GMT
Well I fixed it I think - our routes weren't expecting to return a
response and it ultimately seems to have turned out to the default of
sync=true , not sure why the error message in question was what it was
but essentially when work after mina was done and transitioned through a
splitter etc apparenlty the messages coming out of the splitter got
returned as a response to the mina queue which choked on it because
there were non - serializable items and actually fairly large items too.


On Thu, 2010-04-08 at 14:14 -0500, Andrew Chandler wrote:

> Is there anything wrong with this definition:
> from("direct:rcv").to("mina:tcp://" + pingResponseHost +":" +
> pingResponsePort + "?timeout=300000&transferExchange=false"/*+
> "?textline=true"*/);
> 
> My expectation is that  the textline is ignored (commented out) and
> that I have set a tcp object serialized connection outbound with a
> timeout of 300 seconds.   I can tell you that I'm seeing this error
> within 30 seconds : 
> 
> 
> 
> 
> 
>    ERROR Failed delivery for exchangeId:
> 128c966e-499b-4383-82b1-bf6badd9d504. Exhausted after delivery
> attempt: 1 caught: org.apache.camel.CamelExchangeException: No
> response received from remote server:
> mina://tcp://localhost:5556?timeout=300000&transferExchange=false.
> Exchange[Message: 3,439]      
> org.apache.camel.CamelExchangeException: No response received from
> remote server:
> mina://tcp://localhost:5556?timeout=300000&transferExchange=false.
> Exchange[Message: 3,439]                                          
> 
> 
> On Thu, 2010-04-08 at 13:52 -0500, Andrew Chandler wrote:
> 
> > Ok further weirdness - since I'm using tcp and I don't really care
> > if my objects go as text I just wanted a lightweight socket protocol
> > I took the textline bit out of my uri's (should be serialized
> > objects now )   So in watching it worked for a couple of seconds and
> > then it blew up, however when it blew up it was because it was
> > trying tserialize some object that shouldn't have been present in
> > the message body presented to mina (might have been in a header).
> > My understanding of the doc was that only the body should ever get
> > serialized by default (could have been happening with textline too
> > for all I know).     So I added the transferExchange=false
> > explicitly and am about to test that.   I'll let you know what I
> > find.
> > 
> > 
> > On Thu, 2010-04-08 at 13:05 -0500, Andrew Chandler wrote:
> > 
> > > Ok I'll look at that but what puzzles me is if each message is
> > > sent and is only ever  an int,IPADDRESS  structure how could it
> > > exceed 1024?   Or is the buffer length really ALL messages that
> > > are pending for work?
> > > 
> > > 
> > > 
> > > On Thu, 2010-04-08 at 19:59 +0200, Claus Ibsen wrote: 
> > > 
> > > > Hi
> > > > 
> > > > Have you looked at the encorder/decoder max length options.
> > > > http://camel.apache.org/mina.html
> > > > 
> > > > Could be that Mina has a default of one of them at 1024.
> > > > 
> > > > 
> > > > 
> > > > On Thu, Apr 8, 2010 at 7:14 PM, Andrew Chandler <andy@riftware.com>
wrote:
> > > > > Hi there folks - hoping you can help me from gaining a permanent
dent in
> > > > > my end from hitting the wall.    We are using Camel for many things
but
> > > > > we have setup a dedicate route to a simple standalone process - both
> > > > > ends set up a send and a receive route (different ports)  - The route
> > > > > works until it doesn't.     We are putting in simple text messages
in
> > > > > the form of    "9,10.9.9.9"    That is an entire message so there
> > > > > shouldn't be any reason for the error we are getting except that
we are
> > > > > multithreaded on the both ends.    We are using camel 2.2, and mina-core
> > > > > 1.1.7 ,   Even if you don't know how to guide me to fixing the problem
> > > > > if anybody is able to succesffuly decode the blasted hex dump into
> > > > > meaningful text so that I can understand how the "line is too long"
is
> > > > > occuring it would be wonderful.
> > > > >
> > > > >
> > > > > routes are setup as follows:  First the simple standalone process
sets
> > > > > up route to send responses when work is done to the main process:
> > > > >                        from("direct:rcv").to(
> > > > >                                "mina:tcp://"
> > > > >                                        +
> > > > > responder.this.pingResponseHost
> > > > >                                        + ":"
> > > > >                                        +
> > > > > responder.this.pingResponsePort
> > > > >                                        + "?textline=true&sync=true");
> > > > >
> > > > >
> > > > > and a snippet from the more comlicated  route
> > > > > getFormatOutgoingPingProcessor()).to(
> > > > >                "mina:tcp://" + daemonAddress +
> > > > > "?textline=true&sync=true");
> > > > >
> > > > >
> > > > > The following message was generated on the Main process sending to
> > > > > standalone process side.   My assumption below is that we're somehow
> > > > > overrunnign a buffer with a Message = 10.3.1.12,13   ???????   (Mind
you
> > > > > there were lots of those messages being sent quickly but still)
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > 26931  [Camel thread 23: Threads] ERROR
> > > > > camel.processor.DefaultErrorHandler     - Failed delivery for
> > > > > exchangeId: 5081d54b-5824-4598-94cb-f202559e6e84. Exhausted after
> > > > > delivery attempt: 1 caught: org.apache.camel.CamelExchangeException:
> > > > > Response Handler had an exception. Exchange[Message: 10.3.1.12,13].
> > > > > Caused by: [org.apache.mina.filter.codec.ProtocolDecoderException
-
> > > > > org.apache.mina.common.BufferDataException: Line is too long: 1472
> > > > > (Hexdump: 34 20 36 35 20 36 44 20 35 33 20 36 35 20 37 32 20 37 36
20 36
> > > > > 39 20 36 33 20 36 35 20 37 33 20 33 44 20 33 30 20 32 43 20 32 30
20 37
> > > > > 33 20 37 39 20 37 33 20 37 34 20 36 35 20 36 44 20 35 35 20 37 30
20 35
> > > > > 34 20 36 39 20 36 44 20 36 35 20 33 44 20 33 30 20 32 43 20 32 30
20 37
> > > > > 34 20 37 39 20 37 30 20 36 35 20 33 44 20 36 45 20 37 35 20 36 43
20 36
> > > > > 43 20 35 44 20 32 43 20 32 30 20 36 45 20 36 33 20 36 44 20 35 34
20 36
> > > > > 31 20 37 33 20 36 42 20 34 39 20 36 34 20 33 44 20 36 32 20 33 38
20 36
> > > > > 34 20 33 32 20 33 30 20 33 35 20 36 35 20 33 32 20 32 44 20 36 32
20 33
> > > > > 39 20 33 31 20 36 31 20 32 44 20 33 34 20 36 34 20 33 36 20 33 31
20 32
> > > > > 44 20 33 38 20 33 39 20 33 34 20 33 33 20 32 44 20 33 33 20 33 33
20 33
> > > > > 37 20 33 35 20 36 32 20 33 32 20 33 32 20 36 31 20 36 32 20 33 35
20 33
> > > > > 30 20 33 33 20 32 45 20 33 31 20 33 31 20 32 43 20 32 30 20 36 39
20 36
> > > > > 33 20 36 44 20 37 30 20 35 30 20 36 39 20 36 45 20 36 37 20 35 33
20 37
> > > > > 34 20 36 31 20 37 34 20 37 35 20 37 33 20 34 36 20 36 43 20 36 31
20 36
> > > > > 37 20 33 44 20 34 44 20 36 35 20 37 34 20 36 35 20 37 32 20 36 35
20 36
> > > > > 34 20 34 32 20 36 46 20 36 46 20 36 43 20 36 35 20 36 31 20 36 45
20 32
> > > > > 38 20 37 34 20 37 32 20 37 35 20 36 35 20 32 39 20 32 43 20 32 30
20 36
> > > > > 39 20 37 30 20 33 44 20 33 31 20 33 30 20 32 45 20 33 33 20 32 45
20 33
> > > > > 31 20 32 45 20 33 31 20 33 32 20 37 44 20 30 41 29 5D 0A)]. Processed
by
> > > > > failure processor:
> > > > > Pipeline[Channel[com.visionael.ncm.agent.AgentRouteBuilder
> > > > > $SimpleExceptionProcessor@1d82752b],
> > > > > Channel[Stop]]
> > > > >
> > > > 
> > > > 
> > > > 
> > > 
> > > 
> > 
> > 
> 
> 



Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message