| Subject | Re: [firebird-support] connection forcibly closed by host |
|---|---|
| Author | Helen Borrie |
| Post date | 2004-12-06T23:13:58Z |
At 09:20 PM 6/12/2004 +0000, you wrote:
something happens that the peer at the other end of the connection has
stopped sending responses to your client, or your client has stopped
sending messages to the server. It's your classic
needle-in-a-haystack. The server might have crashed, or been shut
down; the client application itself might have crashed; a router might
have failed; a cable might have been joggled loose; a network card might
have an intermittent fault. On Windows networks, where TCP/IP and Windows
shared file/device services are sharing the wire, it can just be that
someone started a big print job or someone with a misconfigured Internet
gateway connection went online.
If the server is running at the time the blackout occurs, it would log the
network errors. If the server log doesn't show anything then lean towards
abnormal server shutdowns being the cause.
If all the sites are using the same application and it's only happening on
one site, you can eliminate an application fault as the cause. At the
other end of the scale, if you can discover whether the problem occurs only
when a particular user is using the application, it might be easier to
track down the cause of the network fault and see whether replacing the NIC
and patch cable cures the problem.
If the site is in a damp area or near the sea, then (speaking from bitter
experience) it can be a good (read "time-saving and money-saving") policy
to treat NICs and patch cables as consumables and replace them every six
months.
./hb
>I'm using Firebird 1.0.0.796 on W2K. The clients are intermittentlyIt's a socket error, 10054 "Connection reset by peer". It occurs when
>getting the following error:
>
>Unable to complete network to host image1.
>Error writing data to connection.
>An existing connection was forcibly closed by the remote host.
>
>My database path looks like this:
>
>image1:d:\transdoc\auroraimaging.gdb
>
>There is no indication of any problems in interbase.log or the event log.
>
>Anyone have any ideas what might be causing this? This application is
>installed at a couple of dozen sites and only one of the is giving me
>this error.
something happens that the peer at the other end of the connection has
stopped sending responses to your client, or your client has stopped
sending messages to the server. It's your classic
needle-in-a-haystack. The server might have crashed, or been shut
down; the client application itself might have crashed; a router might
have failed; a cable might have been joggled loose; a network card might
have an intermittent fault. On Windows networks, where TCP/IP and Windows
shared file/device services are sharing the wire, it can just be that
someone started a big print job or someone with a misconfigured Internet
gateway connection went online.
If the server is running at the time the blackout occurs, it would log the
network errors. If the server log doesn't show anything then lean towards
abnormal server shutdowns being the cause.
If all the sites are using the same application and it's only happening on
one site, you can eliminate an application fault as the cause. At the
other end of the scale, if you can discover whether the problem occurs only
when a particular user is using the application, it might be easier to
track down the cause of the network fault and see whether replacing the NIC
and patch cable cures the problem.
If the site is in a damp area or near the sea, then (speaking from bitter
experience) it can be a good (read "time-saving and money-saving") policy
to treat NICs and patch cables as consumables and replace them every six
months.
./hb