|
Home > Archive > Apache Directory Project > November 2005 > [MINA] Error on StreamIOHandler
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
[MINA] Error on StreamIOHandler
|
|
| Jose Alberto Fernandez 2005-11-24, 5:46 pm |
| Hi, we are getting the following stack trace when one is not constantly
typing on a telnet session connected using StreamIOHandler:
java.io.IOException: Pipe broken
at java.io.PipedInputStream.read(PipedInputStream.java:255)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:223)
at java.io.PipedInputStream.read(PipedInputStream.java:305)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:229)
at
sun.nio.cs.StreamDecoder$CharsetSD.readBytes(StreamDecoder.java:411)
at
sun.nio.cs.StreamDecoder$CharsetSD.implRead(StreamDecoder.java:453)
at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:183)
at java.io.InputStreamReader.read(InputStreamReader.java:167)
at java.io.BufferedReader.fill(BufferedReader.java:136)
at java.io.BufferedReader.readLine(BufferedReader.java:299)
at java.io.BufferedReader.readLine(BufferedReader.java:362)
at
com.cellectivity.poker.clienthandler.telnet.PokerTelnetHandler$Worker.ru
n(PokerTelnetHandler.java:340)
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecuto
r.java:650)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.ja
va:675)
at java.lang.Thread.run(Thread.java:595)
This is MINA 0.8.1. We have only one connection and this occurs every
time we do not use the telnet session for a little while,
less than a minute. Has anyone seen this error before? Any way to solve
it?
Jose Alberto
| |
| Irving, Dave 2005-11-24, 5:47 pm |
| This is just a guess off the top of my head (so apologies if its wildly
inaccurate), but I wonder whether it has anything to do with processor
threads being shut down if they dont do any work?
When an attempt is made to read from a PipedInputStream, a check is made
every second to see whether the last thread to write to the associated
PipedOutputStream is still alive.
From PipedInputStream:
int trials = 2;
while (in < 0) {
if (closedByWriter) {
/* closed by writer, return EOF */
return -1;
}
if ((writeSide != null) && (!writeSide.isAlive()) && (--trials <
0)) {
throw new IOException("Pipe broken");
}
/* might be a writer waiting */
notifyAll();
try {
wait(1000);
} catch (InterruptedException ex) {
throw new java.io.InterruptedIOException();
}
}
So, if mina shuts down the thread which caused the last write after
inactivity (which I think it might do), this could be the cause.
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:11
To: Apache Directory Developers List
Subject: [MINA] Error on StreamIOHandler
Hi, we are getting the following stack trace when one is not constantly
typing on a telnet session connected using StreamIOHandler:
java.io.IOException: Pipe broken
at java.io.PipedInputStream.read(PipedInputStream.java:255)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:223)
at java.io.PipedInputStream.read(PipedInputStream.java:305)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:229)
at
sun.nio.cs.StreamDecoder$CharsetSD.readBytes(StreamDecoder.java:411)
at
sun.nio.cs.StreamDecoder$CharsetSD.implRead(StreamDecoder.java:453)
at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:183)
at java.io.InputStreamReader.read(InputStreamReader.java:167)
at java.io.BufferedReader.fill(BufferedReader.java:136)
at java.io.BufferedReader.readLine(BufferedReader.java:299)
at java.io.BufferedReader.readLine(BufferedReader.java:362)
at
com.cellectivity.poker.clienthandler.telnet.PokerTelnetHandler$Worker.ru
n(PokerTelnetHandler.java:340)
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecuto
r.java:650)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.ja
va:675)
at java.lang.Thread.run(Thread.java:595)
This is MINA 0.8.1. We have only one connection and this occurs every
time we do not use the telnet session for a little while,
less than a minute. Has anyone seen this error before? Any way to solve
it?
Jose Alberto
This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copiesand inform the sender. Thank you.
| |
| Jose Alberto Fernandez 2005-11-24, 5:47 pm |
| Thanks, not sure if this is my case. As a matter of fact, the stack
trace that I printed comes as part of the Telnet output just before the
connection is closed.
This are some abstracts of the code loop of my worker thread:
protected void processStreamIo(IoSession i_ioSession, InputStream
i_inputStream, OutputStream i_outputStream) {
try {
EXECUTOR_SERVICE.execute(new Worker(i_inputStream,
i_outputStream));
} catch (IOException e) {
e.printStackTrace();
i_ioSession.close();
}
}
public Worker(final InputStream i_in, final OutputStream i_out)
throws IOException {
in = new BufferedReader(new InputStreamReader( i_in ) );
out = new PrintWriter(new BufferedWriter( new
OutputStreamWriter( i_out ) ), true);
}
public void run() {
try {
while (true) {
final String command = in.readLine();
if (processCommand(command.trim())) {
break;
}
}
} catch (IOException e) {
e.printStackTrace(out);
} finally {
out.close();
}
}
As you can see the stack trace is send to the telnet client before
closing the connection.
________________________________
From: Irving, Dave [mailto:dave.irving-iKsOTpgdUR76V6G2DxALlg@public.gmane.org]
Sent: 24 November 2005 15:26
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
This is just a guess off the top of my head (so apologies if its wildly
inaccurate), but I wonder whether it has anything to do with processor
threads being shut down if they dont do any work?
When an attempt is made to read from a PipedInputStream, a check is made
every second to see whether the last thread to write to the associated
PipedOutputStream is still alive.
From PipedInputStream:
int trials = 2;
while (in < 0) {
if (closedByWriter) {
/* closed by writer, return EOF */
return -1;
}
if ((writeSide != null) && (!writeSide.isAlive()) && (--trials <
0)) {
throw new IOException("Pipe broken");
}
/* might be a writer waiting */
notifyAll();
try {
wait(1000);
} catch (InterruptedException ex) {
throw new java.io.InterruptedIOException();
}
}
So, if mina shuts down the thread which caused the last write after
inactivity (which I think it might do), this could be the cause.
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:11
To: Apache Directory Developers List
Subject: [MINA] Error on StreamIOHandler
Hi, we are getting the following stack trace when one is not constantly
typing on a telnet session connected using StreamIOHandler:
java.io.IOException: Pipe broken
at java.io.PipedInputStream.read(PipedInputStream.java:255)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:223)
at java.io.PipedInputStream.read(PipedInputStream.java:305)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:229)
at
sun.nio.cs.StreamDecoder$CharsetSD.readBytes(StreamDecoder.java:411)
at
sun.nio.cs.StreamDecoder$CharsetSD.implRead(StreamDecoder.java:453)
at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:183)
at java.io.InputStreamReader.read(InputStreamReader.java:167)
at java.io.BufferedReader.fill(BufferedReader.java:136)
at java.io.BufferedReader.readLine(BufferedReader.java:299)
at java.io.BufferedReader.readLine(BufferedReader.java:362)
at
com.cellectivity.poker.clienthandler.telnet.PokerTelnetHandler$Worker.ru
n(PokerTelnetHandler.java:340)
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecuto
r.java:650)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.ja
va:675)
at java.lang.Thread.run(Thread.java:595)
This is MINA 0.8.1. We have only one connection and this occurs every
time we do not use the telnet session for a little while,
less than a minute. Has anyone seen this error before? Any way to solve
it?
Jose Alberto
This e-mail and any attachment is for authorised use by the intended
recipient(s) only. It may contain proprietary material, confidential
information and/or be subject to legal privilege. It should not be
copied, disclosed to, retained or used by, any other party. If you are
not an intended recipient then please promptly delete this e-mail and
any attachment and all copies and inform the sender. Thank you.
| |
| Irving, Dave 2005-11-24, 5:47 pm |
| Expanding a bit on this:
ThreadPoolFilter has a default keep alive time of one minute.
If no work comes in to the filter in this time (e.g because you haven't
typed anything for a while), a worker thread could shut down.
So, in your example, the following could happen:
1) You write some data over telnet
2) A processor thread writes the data to the StreamIoHandlers
PipedOutputStream
3) Your PipedInputStream reads the data. At this point, the processor
thread is marked as the last writer (PipedInputStrea#writeSide).
4) You dont send any data for a while (so you are currently blocked in
final String command = in.readLine(); in your code).
5) The processor thread idles out (default keep-alive time is one
minute)
6) The PipedInputStream detects that the last writer thread is no longer
alive and throws a "Pipe Broken" IOException
7) Your code writes this back to the client (in your catch block)
8) Your code closes the connection (your finally block)
I think thats probably whats happening.
To test my theory, try increasing the idle time out of the
ThreadPoolFilter - you shouldn't get the exception until this time has
elapsed (or decrease it, and observe the failure quicker).
So maybe we need to either use a custom piped stream in StreamIOHandler
, or be careful with processor thread keep alives when using VM pipe?
Dave
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:51
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
Thanks, not sure if this is my case. As a matter of fact, the stack
trace that I printed comes as part of the Telnet output just before the
connection is closed.
This are some abstracts of the code loop of my worker thread:
protected void processStreamIo(IoSession i_ioSession, InputStream
i_inputStream, OutputStream i_outputStream) {
try {
EXECUTOR_SERVICE.execute(new Worker(i_inputStream,
i_outputStream));
} catch (IOException e) {
e.printStackTrace();
i_ioSession.close();
}
}
public Worker(final InputStream i_in, final OutputStream i_out)
throws IOException {
in = new BufferedReader(new InputStreamReader( i_in ) );
out = new PrintWriter(new BufferedWriter( new
OutputStreamWriter( i_out ) ), true);
}
public void run() {
try {
while (true) {
final String command = in.readLine();
if (processCommand(command.trim())) {
break;
}
}
} catch (IOException e) {
e.printStackTrace(out);
} finally {
out.close();
}
}
As you can see the stack trace is send to the telnet client before
closing the connection.
________________________________
From: Irving, Dave [mailto:dave.irving-iKsOTpgdUR76V6G2DxALlg@public.gmane.org]
Sent: 24 November 2005 15:26
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
This is just a guess off the top of my head (so apologies if its wildly
inaccurate), but I wonder whether it has anything to do with processor
threads being shut down if they dont do any work?
When an attempt is made to read from a PipedInputStream, a check is made
every second to see whether the last thread to write to the associated
PipedOutputStream is still alive.
From PipedInputStream:
int trials = 2;
while (in < 0) {
if (closedByWriter) {
/* closed by writer, return EOF */
return -1;
}
if ((writeSide != null) && (!writeSide.isAlive()) && (--trials <
0)) {
throw new IOException("Pipe broken");
}
/* might be a writer waiting */
notifyAll();
try {
wait(1000);
} catch (InterruptedException ex) {
throw new java.io.InterruptedIOException();
}
}
So, if mina shuts down the thread which caused the last write after
inactivity (which I think it might do), this could be the cause.
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:11
To: Apache Directory Developers List
Subject: [MINA] Error on StreamIOHandler
Hi, we are getting the following stack trace when one is not constantly
typing on a telnet session connected using StreamIOHandler:
java.io.IOException: Pipe broken
at java.io.PipedInputStream.read(PipedInputStream.java:255)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:223)
at java.io.PipedInputStream.read(PipedInputStream.java:305)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:229)
at
sun.nio.cs.StreamDecoder$CharsetSD.readBytes(StreamDecoder.java:411)
at
sun.nio.cs.StreamDecoder$CharsetSD.implRead(StreamDecoder.java:453)
at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:183)
at java.io.InputStreamReader.read(InputStreamReader.java:167)
at java.io.BufferedReader.fill(BufferedReader.java:136)
at java.io.BufferedReader.readLine(BufferedReader.java:299)
at java.io.BufferedReader.readLine(BufferedReader.java:362)
at
com.cellectivity.poker.clienthandler.telnet.PokerTelnetHandler$Worker.ru
n(PokerTelnetHandler.java:340)
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecuto
r.java:650)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.ja
va:675)
at java.lang.Thread.run(Thread.java:595)
This is MINA 0.8.1. We have only one connection and this occurs every
time we do not use the telnet session for a little while,
less than a minute. Has anyone seen this error before? Any way to solve
it?
Jose Alberto
This e-mail and any attachment is for authorised use by the intended
recipient(s) only. It may contain proprietary material, confidential
information and/or be subject to legal privilege. It should not be
copied, disclosed to, retained or used by, any other party. If you are
not an intended recipient then please promptly delete this e-mail and
any attachment and all copies and inform the sender. Thank you.
| |
| Jose Alberto Fernandez 2005-11-24, 5:47 pm |
| Thanks Dave, now I understand the issue.
Will see if you guys come with a smart solution before mocking with the
thread parameters since we use several protocols and they all share the
same threadfilter in the IoAcceptor.
Jose Alberto
________________________________
From: Irving, Dave [mailto:dave.irving-iKsOTpgdUR76V6G2DxALlg@public.gmane.org]
Sent: 24 November 2005 16:05
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
Expanding a bit on this:
ThreadPoolFilter has a default keep alive time of one minute.
If no work comes in to the filter in this time (e.g because you haven't
typed anything for a while), a worker thread could shut down.
So, in your example, the following could happen:
1) You write some data over telnet
2) A processor thread writes the data to the StreamIoHandlers
PipedOutputStream
3) Your PipedInputStream reads the data. At this point, the processor
thread is marked as the last writer (PipedInputStrea#writeSide).
4) You dont send any data for a while (so you are currently blocked in
final String command = in.readLine(); in your code).
5) The processor thread idles out (default keep-alive time is one
minute)
6) The PipedInputStream detects that the last writer thread is no longer
alive and throws a "Pipe Broken" IOException
7) Your code writes this back to the client (in your catch block)
8) Your code closes the connection (your finally block)
I think thats probably whats happening.
To test my theory, try increasing the idle time out of the
ThreadPoolFilter - you shouldn't get the exception until this time has
elapsed (or decrease it, and observe the failure quicker).
So maybe we need to either use a custom piped stream in StreamIOHandler
, or be careful with processor thread keep alives when using VM pipe?
Dave
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:51
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
Thanks, not sure if this is my case. As a matter of fact, the stack
trace that I printed comes as part of the Telnet output just before the
connection is closed.
This are some abstracts of the code loop of my worker thread:
protected void processStreamIo(IoSession i_ioSession, InputStream
i_inputStream, OutputStream i_outputStream) {
try {
EXECUTOR_SERVICE.execute(new Worker(i_inputStream,
i_outputStream));
} catch (IOException e) {
e.printStackTrace();
i_ioSession.close();
}
}
public Worker(final InputStream i_in, final OutputStream i_out)
throws IOException {
in = new BufferedReader(new InputStreamReader( i_in ) );
out = new PrintWriter(new BufferedWriter( new
OutputStreamWriter( i_out ) ), true);
}
public void run() {
try {
while (true) {
final String command = in.readLine();
if (processCommand(command.trim())) {
break;
}
}
} catch (IOException e) {
e.printStackTrace(out);
} finally {
out.close();
}
}
As you can see the stack trace is send to the telnet client before
closing the connection.
________________________________
From: Irving, Dave [mailto:dave.irving-iKsOTpgdUR76V6G2DxALlg@public.gmane.org]
Sent: 24 November 2005 15:26
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
This is just a guess off the top of my head (so apologies if its wildly
inaccurate), but I wonder whether it has anything to do with processor
threads being shut down if they dont do any work?
When an attempt is made to read from a PipedInputStream, a check is made
every second to see whether the last thread to write to the associated
PipedOutputStream is still alive.
From PipedInputStream:
int trials = 2;
while (in < 0) {
if (closedByWriter) {
/* closed by writer, return EOF */
return -1;
}
if ((writeSide != null) && (!writeSide.isAlive()) && (--trials <
0)) {
throw new IOException("Pipe broken");
}
/* might be a writer waiting */
notifyAll();
try {
wait(1000);
} catch (InterruptedException ex) {
throw new java.io.InterruptedIOException();
}
}
So, if mina shuts down the thread which caused the last write after
inactivity (which I think it might do), this could be the cause.
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:11
To: Apache Directory Developers List
Subject: [MINA] Error on StreamIOHandler
Hi, we are getting the following stack trace when one is not constantly
typing on a telnet session connected using StreamIOHandler:
java.io.IOException: Pipe broken
at java.io.PipedInputStream.read(PipedInputStream.java:255)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:223)
at java.io.PipedInputStream.read(PipedInputStream.java:305)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:229)
at
sun.nio.cs.StreamDecoder$CharsetSD.readBytes(StreamDecoder.java:411)
at
sun.nio.cs.StreamDecoder$CharsetSD.implRead(StreamDecoder.java:453)
at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:183)
at java.io.InputStreamReader.read(InputStreamReader.java:167)
at java.io.BufferedReader.fill(BufferedReader.java:136)
at java.io.BufferedReader.readLine(BufferedReader.java:299)
at java.io.BufferedReader.readLine(BufferedReader.java:362)
at
com.cellectivity.poker.clienthandler.telnet.PokerTelnetHandler$Worker.ru
n(PokerTelnetHandler.java:340)
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecuto
r.java:650)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.ja
va:675)
at java.lang.Thread.run(Thread.java:595)
This is MINA 0.8.1. We have only one connection and this occurs every
time we do not use the telnet session for a little while,
less than a minute. Has anyone seen this error before? Any way to solve
it?
Jose Alberto
This e-mail and any attachment is for authorised use by the intended
recipient(s) only. It may contain proprietary material, confidential
information and/or be subject to legal privilege. It should not be
copied, disclosed to, retained or used by, any other party. If you are
not an intended recipient then please promptly delete this e-mail and
any attachment and all copies and inform the sender. Thank you.
| |
| Irving, Dave 2005-11-24, 5:47 pm |
| I can confirm that I've now reproduced this problem. Changing the
ThreadPoolFilter keep-alive time changes the amount of time required
before the exception is obvserved - so I think we got to the root cause
(PipedInputStream doesn't like producer threads dying).
Possible solutions which spring to mind:
1) Mina could use its own "MinaPipedInput/OutputStream" or something.
2) A ThreadPoolFilter Worker thread could somehow know that it was
logically still required even though it hadn't been active for a while
(maybe a session attribute?). StreamIoHandler would then set this
attribute on a write or something like that (and remove it onIdle /
close whatever). This is quite a bit more hacky than (1).
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 16:53
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
Thanks Dave, now I understand the issue.
Will see if you guys come with a smart solution before mocking with the
thread parameters since we use several protocols and they all share the
same threadfilter in the IoAcceptor.
Jose Alberto
________________________________
From: Irving, Dave [mailto:dave.irving-iKsOTpgdUR76V6G2DxALlg@public.gmane.org]
Sent: 24 November 2005 16:05
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
Expanding a bit on this:
ThreadPoolFilter has a default keep alive time of one minute.
If no work comes in to the filter in this time (e.g because you haven't
typed anything for a while), a worker thread could shut down.
So, in your example, the following could happen:
1) You write some data over telnet
2) A processor thread writes the data to the StreamIoHandlers
PipedOutputStream
3) Your PipedInputStream reads the data. At this point, the processor
thread is marked as the last writer (PipedInputStrea#writeSide).
4) You dont send any data for a while (so you are currently blocked in
final String command = in.readLine(); in your code).
5) The processor thread idles out (default keep-alive time is one
minute)
6) The PipedInputStream detects that the last writer thread is no longer
alive and throws a "Pipe Broken" IOException
7) Your code writes this back to the client (in your catch block)
8) Your code closes the connection (your finally block)
I think thats probably whats happening.
To test my theory, try increasing the idle time out of the
ThreadPoolFilter - you shouldn't get the exception until this time has
elapsed (or decrease it, and observe the failure quicker).
So maybe we need to either use a custom piped stream in StreamIOHandler
, or be careful with processor thread keep alives when using VM pipe?
Dave
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:51
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
Thanks, not sure if this is my case. As a matter of fact, the stack
trace that I printed comes as part of the Telnet output just before the
connection is closed.
This are some abstracts of the code loop of my worker thread:
protected void processStreamIo(IoSession i_ioSession, InputStream
i_inputStream, OutputStream i_outputStream) {
try {
EXECUTOR_SERVICE.execute(new Worker(i_inputStream,
i_outputStream));
} catch (IOException e) {
e.printStackTrace();
i_ioSession.close();
}
}
public Worker(final InputStream i_in, final OutputStream i_out)
throws IOException {
in = new BufferedReader(new InputStreamReader( i_in ) );
out = new PrintWriter(new BufferedWriter( new
OutputStreamWriter( i_out ) ), true);
}
public void run() {
try {
while (true) {
final String command = in.readLine();
if (processCommand(command.trim())) {
break;
}
}
} catch (IOException e) {
e.printStackTrace(out);
} finally {
out.close();
}
}
As you can see the stack trace is send to the telnet client before
closing the connection.
________________________________
From: Irving, Dave [mailto:dave.irving-iKsOTpgdUR76V6G2DxALlg@public.gmane.org]
Sent: 24 November 2005 15:26
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
This is just a guess off the top of my head (so apologies if its wildly
inaccurate), but I wonder whether it has anything to do with processor
threads being shut down if they dont do any work?
When an attempt is made to read from a PipedInputStream, a check is made
every second to see whether the last thread to write to the associated
PipedOutputStream is still alive.
From PipedInputStream:
int trials = 2;
while (in < 0) {
if (closedByWriter) {
/* closed by writer, return EOF */
return -1;
}
if ((writeSide != null) && (!writeSide.isAlive()) && (--trials <
0)) {
throw new IOException("Pipe broken");
}
/* might be a writer waiting */
notifyAll();
try {
wait(1000);
} catch (InterruptedException ex) {
throw new java.io.InterruptedIOException();
}
}
So, if mina shuts down the thread which caused the last write after
inactivity (which I think it might do), this could be the cause.
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:11
To: Apache Directory Developers List
Subject: [MINA] Error on StreamIOHandler
Hi, we are getting the following stack trace when one is not constantly
typing on a telnet session connected using StreamIOHandler:
java.io.IOException: Pipe broken
at java.io.PipedInputStream.read(PipedInputStream.java:255)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:223)
at java.io.PipedInputStream.read(PipedInputStream.java:305)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:229)
at
sun.nio.cs.StreamDecoder$CharsetSD.readBytes(StreamDecoder.java:411)
at
sun.nio.cs.StreamDecoder$CharsetSD.implRead(StreamDecoder.java:453)
at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:183)
at java.io.InputStreamReader.read(InputStreamReader.java:167)
at java.io.BufferedReader.fill(BufferedReader.java:136)
at java.io.BufferedReader.readLine(BufferedReader.java:299)
at java.io.BufferedReader.readLine(BufferedReader.java:362)
at
com.cellectivity.poker.clienthandler.telnet.PokerTelnetHandler$Worker.ru
n(PokerTelnetHandler.java:340)
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecuto
r.java:650)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.ja
va:675)
at java.lang.Thread.run(Thread.java:595)
This is MINA 0.8.1. We have only one connection and this occurs every
time we do not use the telnet session for a little while,
less than a minute. Has anyone seen this error before? Any way to solve
it?
Jose Alberto
This e-mail and any attachment is for authorised use by the intended
recipient(s) only. It may contain proprietary material, confidential
information and/or be subject to legal privilege. It should not be
copied, disclosed to, retained or used by, any other party. If you are
not an intended recipient then please promptly delete this e-mail and
any attachment and all copies and inform the sender. Thank you.
| |
| Irving, Dave 2005-11-25, 2:45 am |
| DIRMINA-129 created to cover this issue.
________________________________
From: Irving, Dave [mailto:dave.irving-iKsOTpgdUR76V6G2DxALlg@public.gmane.org]
Sent: 24 November 2005 17:55
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
I can confirm that I've now reproduced this problem. Changing the
ThreadPoolFilter keep-alive time changes the amount of time required
before the exception is obvserved - so I think we got to the root cause
(PipedInputStream doesn't like producer threads dying).
Possible solutions which spring to mind:
1) Mina could use its own "MinaPipedInput/OutputStream" or something.
2) A ThreadPoolFilter Worker thread could somehow know that it was
logically still required even though it hadn't been active for a while
(maybe a session attribute?). StreamIoHandler would then set this
attribute on a write or something like that (and remove it onIdle /
close whatever). This is quite a bit more hacky than (1).
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 16:53
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
Thanks Dave, now I understand the issue.
Will see if you guys come with a smart solution before mocking with the
thread parameters since we use several protocols and they all share the
same threadfilter in the IoAcceptor.
Jose Alberto
________________________________
From: Irving, Dave [mailto:dave.irving-iKsOTpgdUR76V6G2DxALlg@public.gmane.org]
Sent: 24 November 2005 16:05
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
Expanding a bit on this:
ThreadPoolFilter has a default keep alive time of one minute.
If no work comes in to the filter in this time (e.g because you haven't
typed anything for a while), a worker thread could shut down.
So, in your example, the following could happen:
1) You write some data over telnet
2) A processor thread writes the data to the StreamIoHandlers
PipedOutputStream
3) Your PipedInputStream reads the data. At this point, the processor
thread is marked as the last writer (PipedInputStrea#writeSide).
4) You dont send any data for a while (so you are currently blocked in
final String command = in.readLine(); in your code).
5) The processor thread idles out (default keep-alive time is one
minute)
6) The PipedInputStream detects that the last writer thread is no longer
alive and throws a "Pipe Broken" IOException
7) Your code writes this back to the client (in your catch block)
8) Your code closes the connection (your finally block)
I think thats probably whats happening.
To test my theory, try increasing the idle time out of the
ThreadPoolFilter - you shouldn't get the exception until this time has
elapsed (or decrease it, and observe the failure quicker).
So maybe we need to either use a custom piped stream in StreamIOHandler
, or be careful with processor thread keep alives when using VM pipe?
Dave
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:51
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
Thanks, not sure if this is my case. As a matter of fact, the stack
trace that I printed comes as part of the Telnet output just before the
connection is closed.
This are some abstracts of the code loop of my worker thread:
protected void processStreamIo(IoSession i_ioSession, InputStream
i_inputStream, OutputStream i_outputStream) {
try {
EXECUTOR_SERVICE.execute(new Worker(i_inputStream,
i_outputStream));
} catch (IOException e) {
e.printStackTrace();
i_ioSession.close();
}
}
public Worker(final InputStream i_in, final OutputStream i_out)
throws IOException {
in = new BufferedReader(new InputStreamReader( i_in ) );
out = new PrintWriter(new BufferedWriter( new
OutputStreamWriter( i_out ) ), true);
}
public void run() {
try {
while (true) {
final String command = in.readLine();
if (processCommand(command.trim())) {
break;
}
}
} catch (IOException e) {
e.printStackTrace(out);
} finally {
out.close();
}
}
As you can see the stack trace is send to the telnet client before
closing the connection.
________________________________
From: Irving, Dave [mailto:dave.irving-iKsOTpgdUR76V6G2DxALlg@public.gmane.org]
Sent: 24 November 2005 15:26
To: Apache Directory Developers List
Subject: RE: [MINA] Error on StreamIOHandler
This is just a guess off the top of my head (so apologies if its wildly
inaccurate), but I wonder whether it has anything to do with processor
threads being shut down if they dont do any work?
When an attempt is made to read from a PipedInputStream, a check is made
every second to see whether the last thread to write to the associated
PipedOutputStream is still alive.
From PipedInputStream:
int trials = 2;
while (in < 0) {
if (closedByWriter) {
/* closed by writer, return EOF */
return -1;
}
if ((writeSide != null) && (!writeSide.isAlive()) && (--trials <
0)) {
throw new IOException("Pipe broken");
}
/* might be a writer waiting */
notifyAll();
try {
wait(1000);
} catch (InterruptedException ex) {
throw new java.io.InterruptedIOException();
}
}
So, if mina shuts down the thread which caused the last write after
inactivity (which I think it might do), this could be the cause.
Dave
________________________________
From: Jose Alberto Fernandez [mailto:jalberto-FQMVDHFwfny6lNtOUNzE6A@public.gmane.org]
Sent: 24 November 2005 15:11
To: Apache Directory Developers List
Subject: [MINA] Error on StreamIOHandler
Hi, we are getting the following stack trace when one is not constantly
typing on a telnet session connected using StreamIOHandler:
java.io.IOException: Pipe broken
at java.io.PipedInputStream.read(PipedInputStream.java:255)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:223)
at java.io.PipedInputStream.read(PipedInputStream.java:305)
at
org.apache.mina.io.handler.StreamIoHandler$PipedInputStream.read(StreamI
oHandler.java:229)
at
sun.nio.cs.StreamDecoder$CharsetSD.readBytes(StreamDecoder.java:411)
at
sun.nio.cs.StreamDecoder$CharsetSD.implRead(StreamDecoder.java:453)
at sun.nio.cs.StreamDecoder.read(StreamDecoder.java:183)
at java.io.InputStreamReader.read(InputStreamReader.java:167)
at java.io.BufferedReader.fill(BufferedReader.java:136)
at java.io.BufferedReader.readLine(BufferedReader.java:299)
at java.io.BufferedReader.readLine(BufferedReader.java:362)
at
com.cellectivity.poker.clienthandler.telnet.PokerTelnetHandler$Worker.ru
n(PokerTelnetHandler.java:340)
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecuto
r.java:650)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.ja
va:675)
at java.lang.Thread.run(Thread.java:595)
This is MINA 0.8.1. We have only one connection and this occurs every
time we do not use the telnet session for a little while,
less than a minute. Has anyone seen this error before? Any way to solve
it?
Jose Alberto
This e-mail and any attachment is for authorised use by the intended
recipient(s) only. It may contain proprietary material, confidential
information and/or be subject to legal privilege. It should not be
copied, disclosed to, retained or used by, any other party. If you are
not an intended recipient then please promptly delete this e-mail and
any attachment and all copies and inform the sender. Thank you.
| |
| Alex Cruise 2005-11-25, 5:45 pm |
| On a related note, I'm naively considering a move to a semi-synchronous
architecture for my project and have been wondering whether MINA and
specifically StreamIOHandler are a good fit.
Specifically, I have a lot of code that expects to use InputStreams, but
parts of my code also spend a lot of time waiting for external systems
to do their business, during which time I'd prefer not to have useful
threads blocked waiting for them.
I was thinking I could extend StreamIOHandler with a "streaming" flag.
If the streaming flag were set, likely because someone had asked for an
InputStream and not closed it yet, I would fire any received data into
the Piped*Stream pairs, and if not, I would take a more event-driven
approach to handling the block.
Does this sound reasonable?
Thanks,
-0xe1a
| |
| Trustin Lee 2005-11-28, 2:45 am |
| Hi Alex,
2005/11/26, Alex Cruise <alex-a0FWtchhtz9DPfheJLI6IQ@public.gmane.org>:
>
> On a related note, I'm naively considering a move to a semi-synchronous
> architecture for my project and have been wondering whether MINA and
> specifically StreamIOHandler are a good fit.
>
> Specifically, I have a lot of code that expects to use InputStreams, but
> parts of my code also spend a lot of time waiting for external systems
> to do their business, during which time I'd prefer not to have useful
> threads blocked waiting for them.
>
> I was thinking I could extend StreamIOHandler with a "streaming" flag.
> If the streaming flag were set, likely because someone had asked for an
> InputStream and not closed it yet, I would fire any received data into
> the Piped*Stream pairs, and if not, I would take a more event-driven
> approach to handling the block.
>
> Does this sound reasonable?
Are you saying that it would be great if you can switch the mode
(semi-synchronous stream I/O and original MINA way)? If so, this is a
wonderful idea.
Thanks,
Trustin
--
what we call human nature is actually human habit
--
http://gleamynode.net/
| |
| Alex Cruise 2005-11-28, 5:54 pm |
| Trustin Lee wrote:
> Are you saying that it would be great if you can switch the mode
> (semi-synchronous stream I/O and original MINA way)? If so, this is a
> wonderful idea.
That is exactly what I'd *like* to do, but I lack sufficient depth in
asynchronous architecture to know whether it's feasible. I was hoping
for some opinions. 
BTW, it's conventional wisdom that apart from handling keepalives
efficiently, asynchronous I/O isn't strongly beneficial for Java servlet
systems, since each thread is either 100% busy doing "servlet business"
or 100% idle in the pool.
However, I think there are several "waity" areas that would benefit from
async. You shouldn't have to block a thread waiting for a slow backend
in a proxy-like system, or when you're waiting for an authentication
response from a heavily loaded LDAP server. It probably makes sense for
slow clients. It might even make sense to implement semi-sync database
drivers someday!
In the long run, I'd like to employ a heuristic approach that would
switch, based on observed latency, back and forth between blocking and
non-blocking I/O on each connection (e.g. go async when latency is >
100ms?). I like how MINA's modular architecture makes it fairly easy
to experiment with different approaches. I hope it won't add too much
latency.
-0xe1a
--
This message has been scanned for viruses and dangerous content by
MailScanner, and is believed to be clean.
|
|
|
|
|