Hi,
I think we've already got a lot of issues on the list regarding OOM
errors. In reality, every server needs to be able to manage I/O
efficiently to avoid high memory utilization, which ultimately results
in an OOM.
What can be done? MINA Connectors and Acceptors need to have the Read
and Wr
[
https://issues.apache.org/jira/browse/DIRMINA-372?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12491528
]
peter royal commented on DIRMINA-372:
-
eero: decent idea. patches welcome :)
> Generics support for IoHandler
>
On Apr 24, 2007, at 11:34 AM, mat wrote:
I believe the protocol codec filter is to encode and decode message
under
each session and I believe it uses IOService threads. My question
is: what
if I try to boardcast the message to all sessions connected. The
encode part
will be executed a lot of
On Apr 23, 2007, at 4:09 PM, Pérez Paz Luis Alberto wrote:
1) I disabled the default thread model setting and I added a
newCachedThreadPool in the I/O Layer:
cfg.setThreadModel(ThreadModel.MANUAL);
IoAcceptor acceptor = new SocketAcceptor( IO_THREADS,
Executors.newCachedThreadPool());
On Apr 24, 2007, at 7:13 AM, mat wrote:
Hi folks,
Out of direct buffer memory occured again. I already setup
ByteBuffer.setUseDirectBuffers(false); // Direct buffers perform bad.
ByteBuffer.setAllocator(new SimpleByteBufferAllocator()); //
Disables
pooling.
is data arriving faster tha
On Apr 24, 2007, at 12:18 AM, Brigham Stevens wrote:
I meant to say implement Authentication ("Login", not logging) in the
filterchain... This would be before the ExecutorFilter (which is
last) and
will be making DB requests... Those requests could potentially
block the
IOProcessor thread bec
On Apr 22, 2007, at 8:36 AM, mat wrote:
How mina does if it can't handle incoming events. Drop them? It is
possible,
right?
it doesn't drop anything by default.. You'd have to implement code to
do that..
-pete
--
[EMAIL PROTECTED] - http://fotap.org/~osi
smime.p7s
Description: S/M
On Apr 21, 2007, at 8:38 AM, mat wrote:
I implemented my protocol encoding in the procotol filter and I
used one
session one object method. I just wonder if I try to boardbast the
message
to N clients, which means the encoding part will be called N times?
If so,
should I move my encoding par
On Apr 20, 2007, at 9:18 PM, mat wrote:
I am doing my testing on server. I open 150 concurrent sessions.
However, my
server CPU useage reaches 20% and memory is increasing 300M
(increasing).
Also I got a Exception: Java.lang.OutOfMemoryError: Direct buffer
memory.
at java.nio.Bits.reserveMem
On Apr 19, 2007, at 5:26 AM, Rob Butler wrote:
I've been contemplating building a completely event driven non-
blocking server using Mina. The server will have to listen to
multiple ports and handle both UDP and TCP. I'd like to use a
single thread for the entire application (all acceptors)
Hi,
We built OpenMBean based JMX layer on top of our Mina based gateway for
management and monitoring, we don't see any significant performance degrade
so far.
Best regards,
Jian
On 4/24/07, Brigham Stevens <[EMAIL PROTECTED]> wrote:
Does enabling JMX cause any significant performance hit?
Does enabling JMX cause any significant performance hit?
I am new to JMX, but just tried the Console,and it really looks like an easy
way to remote control the server.
Thanks in advance!!
[
https://issues.apache.org/jira/browse/DIRMINA-372?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12491446
]
Eero Nevalainen commented on DIRMINA-372:
-
Julien: then you leave type unspecified and use Object like befor
Hi folks,
I believe the protocol codec filter is to encode and decode message under
each session and I believe it uses IOService threads. My question is: what
if I try to boardcast the message to all sessions connected. The encode part
will be executed a lot of times. However, if I encode once an
We're close to importing the code now that the software grant for AsyncWeb
has been
received and filed.
For now I recommend discussing it on this ML with a [AsyncWeb] prefix. We
should
ask infrastructure for [EMAIL PROTECTED] ML lists. So until
everything
is setup we should allow ourselves to d
On 4/24/07, Kumaran Arul <[EMAIL PROTECTED]> wrote:
Hi
I am using MINA 1.1.0 and works well in Windows environment when I use 200
concurrent requests. Takes 0.45ms per request.
But when I deployed my application to the HP UX machine it runs fast for first
100 requests and the CPU gets to 10
***
Click here to view our e-mail legal notice:
http://www.mxit.co.za/pdfs/mxit_legal.pdf or call: +27 21 888 7000
***
Mat,
It could be that your application is not proce
Hi MINA community,
I'm programming a server with mina 1.0.3, in order to attend concurrent clients
I'm trying to have a good Executors usage.
I would appreciate your opinion about my implementation:
1) I disabled the default thread model setting and I added a
newCachedThreadPool in the I/O
I believe NIO buffer does use direct buffer memory, right? Actually i got
Out of direct buffer memory exception. BTW, my physical memory is 4G.
2007/4/24, Niklas Therning <[EMAIL PROTECTED]>:
mat wrote:
> Hi folks,
>
> Out of direct buffer memory exception occured again. I already did
> By
mat wrote:
> Hi folks,
>
> Out of direct buffer memory exception occured again. I already did
> ByteBuffer.setUseDirectBuffers(false); // Direct buffers perform
> bad.
> ByteBuffer.setAllocator(new SimpleByteBufferAllocator()); //
> Disables
> pooling.
>
> Why still get this exception? S
As I knew, Direct buffer memory is not collected by GC. Am I right?
2007/4/24, Mark Webb <[EMAIL PROTECTED]>:
depending on your machine, I would increase the MaxDirectMemorySize
parameter to maybe 512M to start with.
--
..Cheers
Mark
On 4/24/07, mat <[EMAIL PROTECTED]> wrote:
> Hi folks,
>
>
depending on your machine, I would increase the MaxDirectMemorySize
parameter to maybe 512M to start with.
--
..Cheers
Mark
On 4/24/07, mat <[EMAIL PROTECTED]> wrote:
Hi folks,
Out of direct buffer memory exception occured again. I already did
ByteBuffer.setUseDirectBuffers(false); //
[
https://issues.apache.org/jira/browse/DIRMINA-372?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12491294
]
Niklas Therning commented on DIRMINA-372:
-
I've thought of something similar: provide an IoHandler implement
Hi folks,
Out of direct buffer memory exception occured again. I already did
ByteBuffer.setUseDirectBuffers(false); // Direct buffers perform bad.
ByteBuffer.setAllocator(new SimpleByteBufferAllocator()); // Disables
pooling.
Why still get this exception? Should I still
increasing max
Hi folks,
Out of direct buffer memory occured again. I already setup
ByteBuffer.setUseDirectBuffers(false); // Direct buffers perform bad.
ByteBuffer.setAllocator(new SimpleByteBufferAllocator()); // Disables
pooling.
[
https://issues.apache.org/jira/browse/DIRMINA-372?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12491282
]
Julien Vermillard commented on DIRMINA-372:
---
why not, but if some codec produce Object of different types
Generics support for IoHandler
--
Key: DIRMINA-372
URL: https://issues.apache.org/jira/browse/DIRMINA-372
Project: MINA
Issue Type: Improvement
Affects Versions: 1.1.0
Reporter: Eero Nevalainen
Did you put:
ByteBuffer.setUseDirectBuffers(false); // Direct buffers perform bad.
ByteBuffer.setAllocator(new SimpleByteBufferAllocator()); // Disables
pooling.
It helps.
2007/4/24, Kumaran Arul <[EMAIL PROTECTED]>:
Hi
I am using MINA 1.1.0 and works well in Windows environment when I us
On 4/24/07, Uday Prakash (DHL CZ) <[EMAIL PROTECTED]> wrote:
Hi,
I also realized that the IoSession ifc, has a method isConnected(), this
returns if the current session is connected to a remote peer. This is
pretty useful and helpful.
That's exactly what I was referring to when I said "no reli
Hi,
I also realized that the IoSession ifc, has a method isConnected(), this
returns if the current session is connected to a remote peer. This is
pretty useful and helpful.
Thank you for all your pointers,
-Uday
-Original Message-
From: Vinod Panicker [mailto:[EMAIL PROTECTED]
Sent:
I am using this:
IoSession s = con.getIoSession();
if(s.isConnected() && con.alive) {
s.write(msg);
}
Also my app sends Pings every 2 minutes, and I've set the idle timer to 2
minutes in sessionOpened, using session.setIdleTime, and use the idle
On 4/24/07, Uday Prakash (DHL CZ) <[EMAIL PROTECTED]> wrote:
Hi,
I was looking thru some documentation online and realized that there is
no standard way [well, at least I didn't find any, may be I didn't look
hard enough or in the right place] to check if a tcp/ip connection is
alive between the
On 4/17/07, James Im <[EMAIL PROTECTED]> wrote:
>* Can someone help me to subscribe to AsyncWeb's mailing list? It seem
>that there is no response after I sent email to
>[EMAIL PROTECTED]
I think that the list is not working correctly for the moment.
I'm sorry but I need your help here... where
Hi,
I was looking thru some documentation online and realized that there is
no standard way [well, at least I didn't find any, may be I didn't look
hard enough or in the right place] to check if a tcp/ip connection is
alive between the server and client.
The reason I am posting is, to figure out
Does this change affect the 1.0 branch either? Then we need to merge
this change there too.
Trustin
On 4/22/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
Author: proyal
Date: Sat Apr 21 13:06:05 2007
New Revision: 531099
URL: http://svn.apache.org/viewvc?view=rev&rev=531099
Log:
Fix exampl
Woops I had a typo..
# > Thanks for your answer.
# > If I implemented logging in the filterchain, doesn't that have a
# > chance of blocking IO on the server until it completes, because the
# > filters are running in the IoProcessor thread?
#
I meant to say implement Authentication ("Login
36 matches
Mail list logo