Have you tried to tune maxKeepAliveRequests and disableProxyCaching in tomcat? If you are running a firewall with a fixup protocol for HTTP (i.e. it does stateful packet inspection), have you tried turning off the fixup protocol and
>>> [EMAIL PROTECTED] 12/17/04 11:23AM >>> We are seeing this when we are running on the loopback address. The network is definitely not misconfigured for that IP. To add to the complexity, we are sending attachments with the message. The timeout only occurs on attachment requests of significant size. What could cause this problem? John Walker On Fri, 17 Dec 2004 08:40:39 -0600, Yu Feng <[EMAIL PROTECTED]> wrote: > Thanks for the reply! > > You are right that tcpmon won't support HTTPS. > > I have proven the timeout issue is related to the network configuration to > that "bad" workstation: I set up a proxy server in a "good" workstation and > re-directed all Axis HTTPS requests to the "good" from the "bad", and all > queries worked. > > -----Original Message----- > From: Greg Michalopoulos [mailto:[EMAIL PROTECTED] > Sent: Thursday, December 16, 2004 10:12 AM > To: [EMAIL PROTECTED] > Subject: RE: Question on timeout > > If you are using https then tcpmon wont be much help because the > transmission will be encrypted before it leaves the client/server. If you > change your request to http then the server will reject the request. > > -----Original Message----- > From: Yu Feng [mailto:[EMAIL PROTECTED] > Sent: Thursday, December 16, 2004 11:06 AM > To: [EMAIL PROTECTED] > Subject: RE: Question on timeout > > Two questions on TCPMon. > > If I originally wants to connect to > https://www.xyz.com/Axis/DownloadService, when I use TcpMon, is this how I > should use: set listening port 5050, target host www.xyz.com and target port > 443, and change my request to http://localhost:5050/Axis/DownloadService? > > Second question, I noticed something different when I use Axis1.1 or Axis1.2 > to start TcpMon: > > With above setting (assuming it's correct), with Axis 1.1, I can see connect > to www.xyz.com in Request panel; however, with Axis 1.2, it says connecting > to www.xyz.com:5050, is this a bug in Axis 1.2? It really should be > www.xyz.com:443 if port is shown as well. > > Yu > > -----Original Message----- > From: Elaine Nance [mailto:[EMAIL PROTECTED] > Sent: Tuesday, December 14, 2004 5:37 PM > To: [EMAIL PROTECTED] > Subject: Re: Question on timeout > > Maybe the timeout settings? or set up TCPMonitor and see whats going back > and forth. > > Yu Feng wrote: > > > Thanks for help! > > > > The computer where this connection problem occurs is our application > server, > > so I haven't got a chance to replace network card. However, it runs > > all other non-Axis applications ok that require Internet connection. > > > > The exact SocketTimeoutException problem also happened in the > > computers > that > > I said "almost always receives response" -- but very rarely and I > > couldn't recreate it if I aim to. So somehow I think that application > > server > computer > > just had worse network configuration that couldn't survive a query > > most time. I am wondering if there're some software aspect approach I > > can look into? > > > > I once heard about TCP_NODELAY setup in registry, but apparently > > there're > no > > such setting in any computers here. > > > > Yu Feng > > > > -----Original Message----- > > From: Elaine Nance [mailto:[EMAIL PROTECTED] > > Sent: Tuesday, December 14, 2004 2:44 PM > > To: [EMAIL PROTECTED] > > Subject: Re: Question on timeout > > > > > > I would bet that the Network Interface Card (NIC) is not working > > properly, if it is an issue only for one computer. Or that the patch > > cord for the computer is bad, for example if someone tripped over it, > > and so on. > > > > Yu Feng wrote: > > > > > >>>Hi, > >>> > >>>I have been bothered by a time-out issue for quite a few days and > >>>wonder if I can get some help from the mailing list. > >>> > >>>We have a Axis 1.1 client application that connects to a remote Web > >>>Service written also in Axis 1.1. The application almost always > >>>receives response in all computers excepts one. In that one > >>>particular computer, most time it gets SocketTimeoutException as reported > widely in Internet: > >>> > >>>AxisFault > >>>faultCode: > {http://schemas.xmlsoap.org/soap/envelope/}Server.userException > >>>faultSubcode: > >>>faultString: java.net.SocketTimeoutException: Read timed out > >>>faultActor: > >>>faultNode: > >>>faultDetail: > >>>{http://xml.apache.org/axis/}stackTrace: java.net.SocketTimeoutException: > >>>Read timed out > >>>at java.net.SocketInputStream.socketRead0(Native Method) at > >>>java.net.SocketInputStream.read(Unknown Source) ... > >>> > >>>The only answer I found so far is to increase the timeout value (more > than > >>>60 seconds) of the binding stub. That didn't work for us. > >>> > >>>However, the application gets response sometimes (about 5% of all the > >>>times we tried) without any change. > >>> > >>>This convinced me that this might be a computer configuration issue. > >>>It has same general configuration as another computer that worked -- > >>>Windows 2000, dynamic IP, on the same network. > >>> > >>>Somebody would have some clue? > >>> > >>>Thanks! > >>>Yu Feng > > > > > > -- > > <~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > | Computers are useless. They can only give you answers. > > | -- Pablo Picasso -- > > <~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > > > > > > > > > > > -- > <~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > | Computers are useless. They can only give you answers. > | -- Pablo Picasso -- > <~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > >
