Archive for the ‘brokenpipe’ tag

summary of the socket error code (Windows and unix) (turn)

October 22, 2010
network 2010-10-2216:41:14 Read 0 Comments 0 Word Count: first look at the windows of statistical error in the code
query on the host, such as windows system, directly into the command line input: nethelpmsg1xxx can.
a network on-line error Socketerror # 11001
that your computer can not connect to server, please check your Proxy Settings and Proxy-related accounts, or suspension of your anti-virus software “personal firewall.”
a network on-line error Socketerror # 11004
should be the subject of networking online, please check the network connection is normal user.
a network on-line error Socketerror # 10060ConnectionTimedOut
that you and the server time-out error occurs online, please check your Proxy related settings, and whether it has SSL443port ProxyServer open.
a network on-line error Socketerror # 10061
indicates that the server is busy you can not set up online, please try again later.
a network on-line error Socketerror # 10022
Please update your Windows access control system components. WindowsNTversion4.0 users download and install ServicePack5 (or more).
appear InternalServerError
host is busy, please try again later.
upload when Socketerror # 10022Invalidargument.
Please update the Windows Internet control system components (windowsupdate). WindowsNTversion4.0 users download and install ServicePack5 (or more).
upload when socketerror # 10054Connectionresetbypeer
original mast as fire or proxy connection is interrupted, “or because you have installed the device set ip ip sharing device to share the first removed, computer connected directly to the adsl line and re-upload.
upload when socketerror # 10057Connectionresetbypeer
the original because the system file is corrupted, the system software, a firewall.
INDY9 development process feedback SocketErrors
Here is SocketError error code and description:
Socketerror0-Directlysenderror

Socketerror10004-Interruptedfunctioncall
Socketerror10013-Permissiondenied
Socketerror10014-Badaddress
Socketerror11001-Hostnotfound
Socketerror11002-Non-authoritativehostnotfound
Socketerror11003-Thisisanon-recoverableerror
Socketerror11004-Validname, nodatarecordofrequested type
WSAEADDRINUSE (10048) Addressalreadyinuse
WSAECONNABORTED ( 10053) Softwarecausedconnectionabor t
WSAECONNREFUSED (10061) Connectionrefused
WSAENETUNREACH (10051) Networkisunreachable
WSAETIMEDOUT (10060) Connectiontimedout

WSAHOST_NOT_FOUND (11001) Hostnotfound
WSASYSNOTREADY (10091) Networksub-systemisunavailabl e
WSANOTINITIALISED (10093) WSAStartup () notperformed
WSANO_DATA (11004) Validname, nodataofthattype
WSANO_RECOVERY (11003) Non-recoverablequeryerror
WSATRY_AGAIN (11002) Non-authoritativehostfound
WSAVERNOTSUPPORTED (10092) WrongWinSockDLLversion
unix platform error code, available online is relatively small, looking for a fairly complete 10001 EPERM Notsuper-user 10002 ENOENT Nosuchfileordirectory 1000 3 < o: p> ESRCH Nosuchprocess 10004 EINTR interruptedsystemcall 10005 EIO I / Oerror 10006 ENXIO Nosuchdeviceoraddress 10007 E2BIG Arglisttoolong 10008 ENOEXEC Exe cformaterror 10009 EBADF Badfilen umber 10010 ECHILD Nochildren 10011 EAGAIN Nomoreprocesses 1 0012 ENOMEM Notenoughcore 10013 EACCES Permissiondenied 10014 EFAULT Badaddress 10015 ENOTBLK Blockdevicerequired 10016 EBUSY Mountdevicebusy 10017 EEXIST Fileexists 10018 < / o: p> EXDEV Cross-de vicelink 10019 ENODEV Nosuchdevic e 10020 ENOTDIR Notadirectory < / o: p> 10021 EISDIR Isadirectory 1002 2 EINVAL Invalidargument 10023 ENFILE Filetableoverflow 10024 EMFILE Toomanyopenfiles 10025 E NOTTY Notatypewriter 10026 ETXTBS Y Textfilebusy 10027 EFBIG Filetoolarge < o: p> 10028 ENOSPC Nospace leftondevice 10029 ESPIPE Illegal seek 10030 EROFS Readonlyfilesyst em 10031 EMLINK Toomanylinks 10032 EPIPE Brokenpipe 10033 < / o: p> 10034 10035 ENOMSG Nomessageofdesiredtype 1003 6 < o: p> EIDRM Identifierremoved 10037 ECHRNG Channelnumberoutofrange

JavaSocket common exception handling

October 28, 2010
Socket Communications, the daily encounter the following exceptions:
first one exception is java.net.BindException: Addressalreadyinuse: JVM_Bin d.
The exception occurs in the offices of device side newServerSocket (port) (port is an integer value 0,65536) operation. Notice is that with the exception of the port as a port has been activated, and eavesdropping. Netstat-an then with the Royal Decree, you can see the status of a Listending port. Only to find a port that is not occupied can solve the problem.
second exception is java.net.ConnectException: Connectionrefused: connec t.
The exception occurs on the client side newSocket (ip, port) operation, the exception is generated notices or ip address of the machine with less than her inability to find (that is, from the current machine does not exist to specify the ip route) or the ip there, but could not find the specified port for monitoring. The problem occurs, check the beginning of the client ip and port is not wrong, if accurate work from the client ping the device is not able to ping to see if can ping (ping work to ban drop-side is not required of the method), then look in the offices of the specified port-side of the monitor is not the testimony of the model, then this can be determined to solve this problem.
The third exception is java.net.SocketException: Socketisclosed.
the exception on the client device and the work may generate. Notice is the unusual initiative to end the connection is closed (called Socket close method) and then connect to the collection of read and write operations.
fourth exception is java.net.SocketException: (Connectionreset or Connectresetbypeer: Socketwriteerror).
the exception on the client and server-side things are possible, cause notice of the exception of two, the first is if the Socket is closed at one end (or active or closed because of abnormal exit caused closed), the other end is still sending data, the first packet sent to stimulate the exception (Connectresetbypeer). Another is the end of the exit, but exit when the connection is not closed, the other end if the connection to read data from the Thrown (Connectionreset). Simply means that the connection is lost after the read and write operations triggered.
fifth exception is java.net.SocketException: Brokenpipe.
the exception on the client, and work devices are likely to generate. In the first four exceptions in the first case (that is thrown SocketExcepton: Connectresetbypeer: Socketwriteerror later), and then continued to write the data if the exception is thrown. The first two exceptions beginning of the solution is to ensure that all the testimonies example out of the collection before the connection closed, followed by the other side closed connection to testing operation, the invention itself should also be connection to the other side closed the connection closed.
two. Example, when writing to collect testimonies have attention problems:
1, is to accurately identify the long and short connection. Connection is the so-called long been established on a permanent hold. Connection is short, preparation of data -> Connect -> send data -> closed connection. Many members of the testimony of the model for many years wrote a collection of testimonies example, does not even know what a long connection, what is the short connection.
2, is a long connection protection. The so-called protection consists of two aspects, beginning of the initiative is to test each other disconnection (ie, calling Socket close method), followed by testing each other down, exception exit and collect unreasonable. This is a strong testimony of the communication model must have. Detect each other active disconnection is simple, even an important one-off initiative, if the other party during the read operation, then the return value of -1 only once to detect each other off with, you should take the initiative to close the end of the connection (called Socket The close method). And testing each other down, and collecting unusual exit barrier, often applied method is to use a “heartbeat”, which is on both sides periodically send data to each other, but also from each other to receive a “heartbeat”, if the last few cycles have not received to each other heartbeat, you can conclude that the other side down, exception exit or collect unreasonable at this time also need to take the initiative to close local connection, if the client can be one accurate time delay after the connection from the new proposal. Socket has a keepalive option of course to protect the connection, if using this option, the invention generally have two hours to each other down, and collecting unusual exit barrier.
3, dealing with effectiveness. Whether the client devices still work, if the model is a long testimony of at least a need to connect two threads, one for receiving data, one for sending heartbeat, do not have to write a special thread, of course, also have a class of other threads ( commonly known as the Worker thread) for message processing, that is only responsible for receiving thread to receive data, and then distributed to Worker for data processing. If it is a short connection, no need to send the heartbeat thread, if the device is also required to act in a dedicated thread responsible for listening for connection requests. These are the testimonies of a communication model of the overall request, the specific example of your testimonies, the testimonies to see you Ruo model for optimization of.

mgaor

October 31, 2010
collection of programming in java Socket communication, the weekdays will be closed on the following exceptions:
first one exception is java.net.BindException: Addressalreadyinuse : JVM_Bin d.
The exception generated in the side newServerSocket (port) (port is an integer value 0,65536) control time. Notice is that with the exception of the port as a port has been activated, and eavesdropping. Netstat-an then with the Royal Decree, you can see the status of a Listending port. Only to find a port that is not occupied can resolve the title of.
second exception is java.net.ConnectException: Connectionrefused: connec t.
The exception occurs on the client side newSocket (ip, port) control, the exception is generated notices or ip address of the machine with less than her inability to find (that is, from the current machine does not exist to specify the ip route) or the ip there, but could not find the specified port for monitoring. Appear in the title, check the beginning of the client ip and port is not wrong, it is assumed accurate ping from the client side of the monitor about the testimony of the specified port model is not activated, the determination to solve this title.
The third exception is java.net.SocketException: Socketisclosed.
the exception on the client and are likely to arise. Notice is the unusual side closed the connection active (call the close method of the Socket) and then connect to the collection of read and write control.
fourth exception is java.net.SocketException: (Connectionreset or Connectresetbypeer: Socketwriteerror).
The exception in both the client and the client may have led to notice the exception of two, the first is to assume that the Socket is closed at one end (or take the initiative to shut down or quit because of abnormalities caused by close ), the other end is still sending data, the first packet sent to stimulate the exception (Connectresetbypeer). Another is the end of the exit, but did not close the connection when exiting the other end of the connection is assumed to read data from the Thrown (Connectionreset). Simply means that the connection is lost after the read and write control triggered.
fifth exception is java.net.SocketException: Brokenpipe.
the exception on the client and are likely to generate. In the first four exceptions in the first case (that is thrown SocketExcepton: Connectresetbypeer: Socketwriteerror later), assuming continued to write data is re-thrown. The first two exceptions to ensure the beginning of the solution is to close all testimonies example out of the collection of connections, followed by testing each other close connection to control, the other side closes the connection after the invention itself should close the connection.
two. Written testimony of role models need to carefully collect the title:
1, is to accurately identify the long and short connection. Connection is the so-called long been established on a permanent hold. Connection is short, preparing data -> Connect -> send data -> close the connection. Many members of the testimony of the model for many years wrote a collection of testimonies example, does not even know what a long connection, what is the short connection.
2, is a long connection protection. The so-called protection consists of two aspects, beginning of the initiative is to test each other disconnection (ie, calling Socket close method), followed by testing each other down, exception exit and collect unreasonable. This is a solid example of the necessary communication with the testimonies. Detect each other active disconnection is simple, even an important one-off initiative, the other party assumes control during reading, then the return value is -1 only once to detect each other off with, you should take the initiative off the side of the connection (called Socket The close method). And testing each other down, and collecting unusual exit barrier, often applied method is to use a “heartbeat”, which is on both sides periodically send data to each other, but also from each other to receive a “heartbeat”, assuming that last a few cycles are not closed heartbeat to the other side, you can sever the other down, or collect abnormal exit barrier, this time also need to take the initiative to close the client connection, the client can assume that one accurate time delay after the connection from the new proposal. Although there is a Socket connection keepalive option to protect, it is assumed that this option, the invention generally have two hours to each other down, and collecting unusual exit barrier.
3, dealing with the effectiveness of the title. Whether the client still is assumed to be a testimony of a long connection example at least have two threads, one for receiving data, one for sending heartbeat, do not have to write a special thread, although other classes also have a thread (commonly known as Worker thread) for message processing, that is only responsible for receiving thread to receive data, and then distributed to Worker for data processing. Connection is assumed to be short, no need to send the heartbeat thread, assuming that is also required is a dedicated thread responsible for listening for connection requests. These are the testimonies of a communication model of the overall request, the specific example of your testimonies, the testimonies to see you Ruo model for optimization of.

Help socket connect timeout issues, urgent !!!!!!!

September 1, 2011
important reminder: The system detects that your account may have stolen the risk, please see the risk warning as soon as possible, and immediately change your password. Close
Netease blog security alert: The system detected that your current password is less secure, for your account security, we recommend that you change your password immediately amend the timely closure
Help socket connect timeout issues, urgent! ! ! ! ! ! !
2011-09-01 01:28:01Category: Default CategoryTags:Subscribe to help medium and small font size socket connect timeout issues, emergency ! ! ! ! ! ! ! Hello, everyone, early name for chinaunix have heard, they often search for information on this, I point at the socket communication problems encountered, would like to ask next.
hardware platform: ARM9
software level control of network traffic problems TC units: ECOS network protocol stack (FreeBSD)
problem description:

two boards, one to run TCP socket Server (accept), a running TCP socket Client (connect), have access to a large network environment (the same floor of the local area network), a special issue of data, a special data collection, communication time can only be continued for several hours, and then there “BrokenPipe” and “ConnectResetByPeer” error, send, recv function returns -1 to quit. Then when I wanted to be an error in the program automatically re-connect them and then send and receive, that is, Client: the first close (sk), re-socket, connect, and Server-side has been a task accept, the results of the phenomenon is several times before re-connection process Client is to connect several times or ten times (the middle range of the delay time) to connect to later (a few hours later), both also Rom on how the ( TCP three-way handshake is not successful), tried many methods will not work (l_onoff = 1; l_linger = 0, and so on). Ecos connect later found under 7 seconds time-out, and found that regardless of the ARP request packet capture or TCP SYN, the two sides of the response is very slow (I do not know why), more than ten seconds after the other only reply is often a matter, so connect it is also not normal.
connect it to change the timeout, a common method: Set non-blocking socket, connect, set time, select., can be written to determine whether, getsockpot, but I find the value of time within the 7s are in effect, more than 7 seconds, are 7s, when select returns, in the ecos library is also looking for clues to no avail (probably does not even know)
problem is how to make connect timeout is greater than 7 seconds, and discussing it … we help to see what approach did not, thank you, my code is as follows http://dok77.xdq.gov.cn:
[code]
int skhttp: / / essvi.ballet.org.cn, ret;
struct timeval tv;
fd_set rdevents , wrevents;
if ((sk = socket (AF_INET, SOCK_STREAM, 0)) <0)
{
perror ("Socket build fail \ n ");
return ERROR;
}
if (ioctl (sk, FIONBIO,