VMCI socket stability problems

Similar Questions

• VMCI socket connect() fails repeatedly, then retrieves when tried in a loop

  I have a VMCI application that uses a number of VMCI sockets to connect to a guest programs.   I noticed an intermittent failure on 2 of the sockets which seems to be linked to the execution of the connect () /closesocket () in a loop too quickly.

  One of the client programs (lets call it customer sock 1) in the prompt will continuously run code that looks roughly similar:

  try again every 500ms until we connect to the server VMCI socket.

  While (! connected)

  {

  int retVal = connect(vmciSocket,...);

  If (retVal! = 0)

  {

  Sleep (500);

  }

  }

  until its socket server (server 1 sock) appears, sometime later.  The client can operate for long periods of time to try to connect to the server.

  During the execution of customer 1 a second pair of sockets VMCI is created.  Sometimes the client side of the 2nd pair fails to connect with an error (invalid) 100037 WSA.   If put the side customer (customer sock 2) in a loop to try slow it will connect eventually (after about 20 seconds and the failures of 10-15).  The second take of client contains simple to connect, sleep to 5 attempts before leaving logic.

  With 1 2 customer the customer end abnormally about 40% of the time with the error logs indicating that he is not able to connect to its server VMCI socket.   If I disable the client 1, the failure rate falls to 10%.

  In addition, I noticed that these error matches an error log in my vmware.log.   When these errors I get error that looks like the following:

  2012 02-23 T 22: 39:49.751Z | vCPU-0 | I120: VMCI QueuePair: error in BusMemVMCI_InvalidateBPNs for alloc (numInvalidated = 5, numPPNs = 130).
  2012 02-23 T 22: 39:50.012Z | vCPU-0 | I120: VMCI QueuePair: error in BusMemVMCI_InvalidateBPNs for alloc (numInvalidated = 123, numPPNs = 130).
  2012 02-23 T 22: 39:50.278Z | vCPU-0 | I120: VMCI QueuePair: error in BusMemVMCI_InvalidateBPNs for alloc (numInvalidated = 30, numPPNs = 130).
  2012 02-23 T 22: 39:53.640Z | vCPU-0 | I120: VMCI QueuePair: error in BusMemVMCI_InvalidateBPNs for alloc (numInvalidated = 76, numPPNs = 130).
  2012 02-23 T 22: 39:56.831Z | vCPU-0 | I120: VMCI QueuePair: error in BusMemVMCI_InvalidateBPNs for alloc (numInvalidated = 90, numPPNs = 130).
  2012 02-23 T 22: 39:57.111Z | vCPU-0 | I120: VMCI QueuePair: error in BusMemVMCI_InvalidateBPNs for alloc (numInvalidated = 10 numPPNs = 130).
  2012 02-23 T 22: 39:57.374Z | vCPU-0 | I120: VMCI QueuePair: error in BusMemVMCI_InvalidateBPNs for alloc (numInvalidated = 120 numPPNs = 130).
  2012 02-23 T 22: 39:57.640Z | vCPU-0 | I120: VMCI QueuePair: error in BusMemVMCI_InvalidateBPNs for alloc (numInvalidated = 42 numPPNs = 130).

  He comes to Workstation 8.0.1 for Windows link VMCI 8.0.0 - vmware_vmci_sockets-471780 headers.   Windows 7 in comments and host operating system.

  Anyone encountered before or knows a way around?

  Hey Kevin,

  This can be an interference caused by the VMware SVGA or USB drivers in the comments. Could check you if all the drivers are in use and if disabling one or the other solves the problem?

  If one of these drivers, we have unfortunately not a problem as such. The solution is to keep a retry for a while as you already do, or do not use these drivers if possible.

  Thank you

  Jørgen

• VMCI socket - bug in the select() function?

  Hello

  I now had to create non-blocking I/o channels but cannot get your implementation of select() to work. The problem is that it is not timeout properly and therefore act as a blocking IO.

  int selectOnRead(int sockfd, unsigned long secTimeout, unsigned long uTimeout)
  {
      fd_set readFdSet;
      struct timeval tv;
       
      tv.tv_sec = secTimeout;
      tv.tv_usec = uTimeout;
       
      FD_ZERO(&readFdSet);
      FD_SET(sockfd, &readFdSet);
  
      select(FD_SETSIZE, &readFdSet, NULL, NULL, &tv)
  }
  

  The above code is as simple as it gets with select; single file descriptor than if the time-out on the values appearing "tv". If we go through the arguments to choose, they must be:

  -FD_SETSIZE = number of FDs + 1 (or FD_SETSIZE)

  -readFdSet = Array of FDs to observe when they can be read without blocking

  -tv = a simple time structure that can contain the seconds and microseconds

  (If tv is select NULL crashes, if the two time holders are 0, select return immediately, otherwise it uses the timeout values - http://www.gnu.org/software/libc/manual/html_mono/libc.html#Waiting-for-I_002fO( )

  I tried the following:

  1. If the two casings of television time is set to 0 VMCI socket still blocked.

  2. it does not matter if the datagram or stream is used.

  3. If all three observable types is select NULL will not block (or timeout) but rather of return error(-1).

  No matter what I put in the arguments, it blocks until a client sends a packet. Is this a known bug or wrong my application?

  For a more complete example, see the server implementation in the attachment.

  Kind regards

  Andreas

  The timeout value is currently ignored on windows (identified by the 374790 bug). I think this should work for Linux though.

• ReadyNAS 6.5.0 legacy (Pro6) have stability problems?

  Hi guys,.

  I am running 6.2.4 without no stability problem on my Pro 6 and I would like to know if anyone has tried 6.5.0beta2 on Pro6.

  I luck was lowered from 6.4.1 causing very great stability on my Pro6 (blocks, freeze...) problems

  6.2.4 seems rock solid, but I would try 6.5.0 but I need some confirms stability and/or similar problems.

  Thank you!

  I have it on a Ultra6 and it is fine.

• VMCI Sockets fail to compile

  Every time I run VMware Workstation 7.1.1 I got this warning:

  "Before able to run VMware, several modules must be compiled and loaded into the core of execution"

  I click on the "Install" button and everything is done, but 'VMCI Sockets' fail.

  Anyone know why this is happening?

  I'm running Debian with 2.6.35.3 kernel.

  The element of structure sk-> sk_sleep has been replaced in 2.6.35 with a function call, sk_sleep (sk).  So we have to correct Sockets VMCI, patching the source tarball from vsock.tar to use the new syntax of kernel for sk_sleep.

  I enclose a gzip would shell script to fix.  Gunzip just the attachment, do 'chmod 0755' to make it executable, and then run as root (or sudo).

  See you soon

• Computer has been hacked twice. Now, having stability problems.

  I do not trust this pc cause its been hacked twice and now I have problems with this computer, stability has I don't know if it's true, I bought a repair thing and I forgot and it is one of the problems I've lost next to her.

  [Original title"do I trust]

  Hello

  Thanks for posting your query in Microsoft Community.

  I understand your concern and will help you solve this problem.

  I suggest you to see link below and check if it helps.

  http://answers.Microsoft.com/en-us/Windows/Forum/Windows_7-security/has-my-computer-been-hacked/b9132be4-085a-4B39-8d8e-577886b1ea4a

  How to recover your hacked Microsoft account

  http://www.Microsoft.com/security/online-privacy/hacked-account.aspx

  I hope this helps. If the issue is not resolved, please get back to us and we would be happy to help you.

• in the waking state socket connection problem

  Hi all.
  My app need update real-time security data, it should be able to run well even the switch of the PB in the waking state.
  The problem is that when PB passes to the waking state and more than an hour later, my application cannot send request for data to the socket server (initially he workd well), even when the PB get again enabled, the socket connection is also unnecessary.
  I did a test as follows:
  1, I added the "throttled" in the blackberry - tablet.xml
  2. I have add a few listiner event in the main class, but each of them can be fired in any time.

  QNXSystem.system.addEventListener(qnx.events.QNXSystemEvent.STANDBY,onDeactive);
  QNXSystem.system.addEventListener(qnx.events.QNXSystemEvent.ACTIVE,onActive);
  QNXSystem.system.requestResource(QNXSystemResource.NETWORKING);
  stage.addEventListener(Event.USER_IDLE,userIdle);
  stage.addEventListener(Event.USER_PRESENT,userPresent);
  stage.addEventListener(Event.ACTIVATE,onStageActive);
  stage.addEventListener(Event.DEACTIVATE,onStageDeactive);
  

  3, I put a clock to save information in / shared folder.it worked well in the waking state.

  Thanks for any response.

  Hello.
  I'm a developer of instant messaging which is connected to the door via socket. What I can tell you for sure I you need to use:
  normal
  That, to use SocketMonitor to ensure that connections are available, some WiFi routers can close idle connection event when U don't want to! If you need to reconnect in such cases.
  Please, feel free to ask for help more.

• No Edge 15 Broadcom 802.11ac none hosted network management and stability problems

  I have a new edge 15 that came with Windows 10 64-bit pre-installed. I want to be able to run the soft AP WiFi so I can connect to my wifi adapter.

  The WiFi card is a Broadcom 802.11ac and according to other computers using Dell with the same chip, but the Dell Windows 7 driver supports hosted network supports. But on my computer with the latest driver 7.35.267.0 dated 05/14/15, supported network hosted shows no., which means that the pilot and the chip does not support.

  It seems that it's a driver problem and hardware supports hosted networks.

  Also, sometimes when I turn on my computer, the WiFi does not detect the AP. It's the continuous offshore. I can get the connectivity by deactivation and reactivation of the interface. A pain. Not a big pain, but pain. Looks like the driver reload or turn on the radio after that sleep does not work properly.

  Can Lenovo if you please provide a stable driver for the Broadcom chip work.

  Thank you.

  I ended up completely reinstall windows 10 and uses the stock drivers.

  No bloat ware lenovo. Works fine now.

• socket connection problem

  Hello

  Here is my code to open a client-side socket connection to a server running on the vmware host I don't know there works well (the server):

  #include
  #include
  #include
  #include
  #include

  int _socket;
  struct sockaddr_in name;

  bool connect (char * ip, int port)
  {
  _socket = socket (AF_INET, SOCK_STREAM, 0);
  If (_socket<>
  {
  Returns false;
  }
  struct hostent * host;
  struct in_addr addr = {0};
  addr.s_addr = inet_addr (ip);
  Host = gethostbyaddr ((char*) & addr, 4, AF_INET);
  Name.sin_family = AF_INET;    Name.sin_addr.s_addr = * ((unsigned long *)-> h_addr host);
  Name.sin_port = htons (port);
  If (bind (_socket, (struct sockaddr *) & name, sizeof (name)))
  {
  Returns false;
  }
  Listen (_socket, 1);
  _socket = accept (_socket, 0, 0);
  Returns true;
  }

  Connect ("127.0.0.1", 44444);

  When you call the connect() method is block on accept() (EEG open Server before the client does). So, I need to know if the question is:

  1_ on my customer code.

  or

  2_ vmware cannot connect to the host computer using the address of the local host computer.

  (or even both)

  You will appreciate all help.

  Thank you.

  After a lot a work around, I realize that the last three

  Methods include the specified address and get ready to listen and

  accept a connection from an unknown client. I mean, it's

  create a server, that is why it stops on the accept() method.
  For example, the code for one side of the client socket should look like this (or almost)

  (similar):

  bool open_connection (char * ip, int port)
  {
  _socket = socket (AF_INET, SOCK_STREAM, 0);
  If (_socket<>
  {
  Returns false;
  }
  struct hostent * host;
  struct in_addr addr = {0};
  addr.s_addr = inet_addr (ip);
  Host = gethostbyaddr ((char*) & addr, 4, AF_INET);
  Name.sin_family = AF_INET;
  Name.sin_addr.s_addr = * ((unsigned long *)-> h_addr_list host)

  [0]);
  Name.sin_port = htons (port);
  If (connect (_socket, (struct sockaddr *) & name, sizeof))

  (name)) )
  {
  Returns false;
  }
  Returns true;
  }

  Then, from the host and the Simulator on a socket connection I

  the same as that used for the loading of the appliction to it.
  One thing: when I try to establish a connection when my computer

  is connected to the internet both the binding and the connection,

  the function failed.
  @Mshom: when I tried to put the adapter to bridge the

  Simulator cannot obtain an "ip address" ("get ip address... is written

  Botton left on cone winder").
  By the way, he thinks that all those of you who answer me.

• 11.1.2.1 stability problem

  Hi all
  
  Ive installed EMP 11.1.2.1(essbase,planning,HFR) 64-bit on windows xp professional 64-bit. system works fine for awhile, but it fails after a while. He began to launch a series of errors.
  as
  Description: Invalid or the configuration of the module 'HyperionPlanning.planning' is not found.
  Required application module HyperionPlanning.planning is not configured.
  http://Hyperion:19000/HyperionPlanning/conf/HspJSConfig.XML
  
  Status: 503 - Service temporarily unavailable
  
  Content: text/html
  < HTML >
  < HEAD >
  Weblogic bridge < TITLE > message
  < /title >
  < / HEAD >
  < BODY >
  < H2 > the bridge APACHE server failed: < / H2 > < P >
  < hr > not available for a connection of back-end server: expired after 10 seconds or idempotent off.
  < hr > < / BODY >
  < / HTML >
  
  Sometimes it I cannot enter in application size/library library.
  
  Here are the services that is running on my system.
  
  Hyperion - Web Application Administration services
  Hyperion CALC Manager - Web Application
  Hyperion EPMA Data Synchronizer - Web Application
  Hyperion Server EPMA
  Hyperion EPMA Web - Web Application layer
  Hyperion Financial Reporting - Web Application
  Hyperion Foundation Services - managed server
  Hyperion Integration Services
  Hyperion Planning - Web Application
  Hyperion provider Services - Web Application
  Framework for analysis and Reporting, Hyperion
  Hyperion statement and analytical framework - Web Application
  Hyperion RMI Registry
  Hyperion Studio Service BPMS bpms1
  Oracle (EPM_epmsystem1) process manager
  Oracle (ohsInstance3250625901) process manager
  Oraclee_oracle_middle ~ 1_epmsys ~ 1e_oracle_middleware_user_projects_epmsystem1ConfigurationManager
  
  is there anything else outside of these services, which must be present?
  How to solve this contradiction system performance.
  
  Thank you

  Only is even not supported on windows xp, so it not surprising that you have any questions, the support matrix is available at http://www.oracle.com/technetwork/middleware/bi-foundation/hyperion-supported-platforms-085957.html

  See you soon

  John
  http://John-Goodwin.blogspot.com/

• Schema stability problem

  Why my web page "upward bouquet" and the elements of the stack when I do the narrower browser window? How can I keep the layout intact when I do that to the browser window? Sorry, but I'm really new to this.

  I use the mac platform and the safari browser. Any page in hybackstudio.com clusters upward when I see on the internet. Maybe it doesn't on firefox.

  It does it on all browsers. So jtanna could have some wise advice by suggesting that working your layout, the only answer to your question immediately declared is to assign a min width to your body rule. You have CSS into your page and you have a body selector:

  {body

  Police: 100% Verdana, Arial, Helvetica, without serif.

  Background: #666666;

  margin: 0; / * It is advisable to zero, the margin and the filling of the body to hold element has a different default browser * /.

  padding: 0;

  text-align: center; / * This centers the container in IE 5 * browsers. The text is then set left aligned by default in the #container selector * /.

  Color: #000000;

  background-color: #000;

  }

  Add a min-width declaration:

  {body

  Police: 100% Verdana, Arial, Helvetica, without serif.

  Background: #666666;

  margin: 0; / * It is advisable to zero, the margin and the filling of the body to hold element has a different default browser * /.

  padding: 0;

  text-align: center; / * This centers the container in IE 5 * browsers. The text is then set left aligned by default in the #container selector * /.

  Color: #000000;

  background-color: #000;

  min-width: 980px;

  }

  --

  Al Sparber - PVII

  http://www.Projectseven.com

  The best Extensions to Dreamweaver since 1998

  Coming soon: CSS Magic composer:

  http://www.Projectseven.com/peeks/CCM/

• Select and WSAWaitForMultipleEvents with VMCI and UDP sockets in Windows prompt

  I am trying to write a program using VMCI sockets that will transmit datagrams between a host VMWare ("customer") and a computer connected to a VMWare Windows guest (the "Server").  I'm using VMWare Workstation 6.5.2, build 156735.

  The relay program runs on the client.  VMCI is used to Exchange datagrams with the Linux host. UDP is used to Exchange datagrams with the other computer, which is connected to a network that is only accessible to the guest.  The client works on Windows XP.

   "Client"                      "Relay"                     "Server"
  Linux/Host                   Windows/Guest              Another Computer
             <--VMCI/DGRAM-->              <--UDP/Private-->
  

  I have all the basic ingredients of work, but having problems on the Windows client to select() or WSAWaitForMultipleEvents() to work with UDP and VMCI.

  First of all, select():

  client_fd = socket(af_vmci, SOCK_DGRAM,...);
  bind(client_fd...);
  
  server_fd=socket(AF_INET, SOCK_DGRAM,...);
  bind(server_fd...);
  
  FD_ZERO(&readfds);
  FD_SET(client_fd, &readfds);
  FD_SET(server_fd, &readfds);
  select(FD_SETSIZE, &readfds, 0, 0, 0);
  

  If I configured the parameter readfds with both VMCI socket fd (client) and the Protocol UDP socket fd (server), and then select an error of Windows Sockets, 10022 (Invalid Argument).  If I comment out the FD_SET VMCI socket fd, and then select it call works and my program expects successfully and bed, the socket UDP datagrams.  Similarly, if I comment out the UDP socket fd FD_SET, then select is behaving normally and my program is successfully and bed, taking VMCI datagrams.

  Have no chance with select(), I started using the WSAWaitForMultipleEvents():

  client_fd = socket(af_vmci, SOCK_DGRAM,...);
  bind(client_fd...);
  client_event = WSACreateEvent();
  WSAEventSelect(client_fd, client_event, FD_READ);
  
  server_fd=socket(AF_INET, SOCK_DGRAM,...);
  bind(server_fd...);
  server_event = WSACreateEvent();
  WSAEventSelect(server_fd, server_event, FD_READ);
  
  events[0] = client_event;
  events[1] = server_event;
  wait_status = WSAWaitForMultipleEvents(2, events, FALSE, WSA_INFINITE, FALSE);
  
  ... and then after reading from a socket ...
  
  ResetEvent(event);
  

  This approach has completely killed all VMCI features.  My program was waiting for and read, the datagrams UDP, but the event associated with taking VMCI was never put in a signaled state.

  I'm stuck now.

  On the side of Linux, VMCI seems to work more as it should.  My "client" program creates a VMCI/datagram socket and uses the TUN/TAP interface to extract the IP of the Linux kernel packages.  I use select() must to wait for data to arrive either fd or fd TUN/TAP VMCI socket.  All this seems to work very well.

  I guess I can re - write the Guest program Windows "relay" as two separate programs, each block on a single socket - essentially two programs relay half duplex.  Or I could use the 2 wires in a single program.

  Any ideas as to what the Windows Sockets error 10022 (Invalid Argument) in select can be?

  Hello

  Select on windows only supports file handles same service provider. Since the msdn site-"the sockets contained in the fd_set structures must be associated with a single.

  service provider. For the purpose of this restriction, sockets are

  considered to be the provider of service even though the structures WSAPROTOCOL_INFO describing their protocols have the same providerId value. "If you cannot mix VMCI and other file handles.

  Also we do currently not support WSAEventSelect even if we plan to support in the future.

  Aaron

• How much MSG headers/msgs./msg. descriptions can a download before TB starts to have stability issues and crashes? Nobody has dealt with this issue before?

  I had a lot of problems of stability with each version of TB! Nobody reported this issue before?
  I do not a lot of modules if not all if is not the cause. How many messages, the message headers, message descriptions or all what you call them, can a download when you subscribe to newsgroups or wants to download messages before TB is beginning to show some stability problems?

  Well, if it blocks perhaps you could provide crash IDS. See https://support.mozilla.org/en-US/kb/mozilla-crash-reporter#w_viewing-crash-reports.

  Only the start in SafeMode in the Help menu will exclude questions theme and Add on. It is not good to say "hardly, if ever, use modules TB"

  Note that Thunderbird is not a good choice as a replacement for a news server, and that unless you intervene actively in message headers only process are downloaded (the same with all the readers of news in my experience.) After the selection of the message the body of the message itself is downloaded from the server.

  Thunderbird also has a limit of 4 GB by news group. Recent work was made to remove the limit, but it can still be part of Thunderbird news, because it gets almost no attention programmer only experience will tell you for sure. (NNTP is slowly die if we like it or not).

  I also note that most newsgroups with number of huge messages are groups of binary files and Thunderbird does a particularly poor job of reading yenc and recombining multipart messages. Perhaps a news reader is what you need.

• I just installed Firefox 6, and the report does more - the same problem you had with Firefox 5. What is going on?

  You prompted me to install Firefox 6 tonight, because it addresses serious security/stability problems, but when I did this, the report icon has disappeared from the address bar. Reinstall the latest version of the report was corrected not it.
  It's the same thing with Firefox 5 a few months ago, and it has taken report per week to develop a new compatible version.
  Why this happened again? You not check this before launching new versions? Or, why at all not least we warn in advance that some programs will not work for X days? I prefer to be without the latest Firefox for a week or two, and no report - not least because that my Bank, I have to use the report.

  Please advise as soon as POSSIBLE.

  Malcolm Oliver

  This is the link in the message that Trusteer sent me today. I have updated and works great so far.

  Please contact the Support technique Trusteer.

  We are pleased to inform you that a version of the report that supports Firefox 6 has
  been released and is now available for download in the following link:
  http://download.Trusteer.com/NkuiAcruiKc/RapportSetup.exe

  Please be sure to restart your computer once the installation process
  completed.

  Please note: this version of the report is currently a beta version. We'd
  appreciate any comments you may have about this version.

  If the report indicates that the version of the report that you are trying to install is older
  than the version already installed, follow these steps:
  1. choose uninstall report
  2. When prompted, select 'delete all user settings '. Please, assisted by the
  suite of image: http://consumers.trusteer.com/files/images/uninstall.jpg
  3 reinstall the report from the link above.

• Problems to show the driver IVI and property loader error messages when you use a type of operator interface

  I use UI operator (programmed in ICB, with asynchronous timer) in the folder OP1 in attached zip file. This operator interface does not have the error message from the driver of the IVI, properts, charger etc. on the screen (messages as in the photo attached error.jpg in zip file attached). Due to stability problems, I use this operator interface, but I need these error message. One other operator interface is OP2 in attached zip file. This operator interface puts the error message from the driver of the IVI, charger properts... on the screen (messages as in the photo attached error.jpg in zip file attached). Now, I want to have this feature with OP1 for reasons of stability I use OP1. How can I implement this in OP1? I tried and tried a lot of time, but I have not found a method to implement this. can someone help me? Thank you kind regards Samuel

  Hello, Johann,.

  I'm not sure you understood my problem. My problems are not the IVI and loader drivers error messages property! But my problem is, that errors of the IVI drivers and property loader operator interface does not bring on the screen. So to test this situation, cause an error of the driver of the IVI or shipper of property, then you can see that with a PC, you have an error message and with another IO no error message takes place and the test of stocks without a message. Do you know why this OI does not show the error messages and how showing the error message can be implemented in this operator interface

  concerning

  Samuel