What does it mean ?????

[Franck Y]

Hello everyone !

My server's name is "constellation"

I don t know waht does this thing mean.....like the "get_peer_addr "

Can you excplain me thk you

Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
lib/util_sock.c:get_peer_addr(1150)
Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
was Transport endpoint is not connected

Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
lib/util_sock.c:write_socket_data(430)
Oct 20 13:41:09 constellation smbd[3927]:   write_socket_data: write
failure. Error =
Connection reset by peer
Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
lib/util_sock.c:write_socket(455)
Oct 20 13:41:09 constellation smbd[3927]:   write_socket: Error
writing 4 bytes to socket 5: ERRNO = Connection reset by peer
Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
lib/util_sock.c:send_smb(647)
Oct 20 13:41:09 constellation smbd[3927]:   Error writing 4 bytes to
client. -1. (Connection reset by peer)
Oct 20 14:13:09 constellation smbd[4254]: [2005/10/20 14:13:09, 0]
lib/util_sock.c:get_peer_addr(1150)
Oct 20 14:13:09 constellation smbd[4254]:   getpeername failed. Error
was Transport endpoint is not connected
Oct 20 14:13:09 constellation smbd[4254]: [2005/10/20 14:13:09, 0]
lib/util_sock.c:write_socket_data(430)
Oct 20 14:13:09 constellation smbd[4254]:   write_socket_data: write
failure. Error = Connection reset by peer
Oct 20 14:13:09 constellation smbd[4254]: [2005/10/20 14:13:09, 0]
lib/util_sock.c:write_socket(455)
Oct 20 14:13:09 constellation smbd[4254]:   write_socket: Error
writing 4 bytes to socket 24: ERRNO = Connection reset by peer
Oct 20 14:13:09 constellation smbd[4254]: [2005/10/20 14:13:09, 0]
lib/util_sock.c:send_smb(647)
Oct 20 14:13:09 constellation smbd[4254]:   Error writing 4 bytes to
client. -1. (Connection reset by peer)
--
Franck

[Rex Dieter]

Franck Y wrote:
> I don t know waht does this thing mean.....like the "get_peer_addr "
>
> Can you excplain me thk you
>
> Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
> lib/util_sock.c:get_peer_addr(1150)
> Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
> was Transport endpoint is not connected

getpeername failed => failed DNS/hostname lookup.

-- Rex

[Philip Prindeville]

Rex Dieter wrote:
> Franck Y wrote:
> > I don t know waht does this thing mean.....like the "get_peer_addr "
> >
> > Can you excplain me thk you
> >
> > Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
> > lib/util_sock.c:get_peer_addr(1150)
> > Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
> > was Transport endpoint is not connected
>
> getpeername failed => failed DNS/hostname lookup.
>
> -- Rex

No, that would be gethostbyaddr().  getpeername() is a system call
that looks for the address of the other side (the remove side) of an
association or socket pair.

Of course, if you're using connectionless sockets (i.e. datagrams),
then it won't return an endpoint...  For that, you'd have to use
recvfrom() and note the endpoint from individual requests.

-Philip

[Aly Dharshi]

I believe that you are running samba, which allows for FC to be a Windows file 
server. You can turn this off by typing: service smb stop, and it should be 
shutdow if that is what you want, if that is NOT what you want then the 
responses previously supplied should suffice.

Philip Prindeville wrote:
> Rex Dieter wrote:
> > Franck Y wrote:
> > > I don t know waht does this thing mean.....like the "get_peer_addr "
> > >
> > > Can you excplain me thk you
> > >
> > > Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
> > > lib/util_sock.c:get_peer_addr(1150)
> > > Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
> > > was Transport endpoint is not connected
> >
> > getpeername failed => failed DNS/hostname lookup.
> >
> > -- Rex
>
> No, that would be gethostbyaddr().  getpeername() is a system call
> that looks for the address of the other side (the remove side) of an
> association or socket pair.
>
> Of course, if you're using connectionless sockets (i.e. datagrams),
> then it won't return an endpoint...  For that, you'd have to use
> recvfrom() and note the endpoint from individual requests.
>
> -Philip

[Edward Dekkers]

Philip Prindeville wrote:
> Rex Dieter wrote:
> > Franck Y wrote:
> > > I don t know waht does this thing mean.....like the "get_peer_addr "
> > >
> > > Can you excplain me thk you
> > >
> > > Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
> > > lib/util_sock.c:get_peer_addr(1150)
> > > Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
> > > was Transport endpoint is not connected
> >
> > getpeername failed => failed DNS/hostname lookup.
> >
> > -- Rex
>
> No, that would be gethostbyaddr().  getpeername() is a system call
> that looks for the address of the other side (the remove side) of an
> association or socket pair.
>
> Of course, if you're using connectionless sockets (i.e. datagrams),
> then it won't return an endpoint...  For that, you'd have to use
> recvfrom() and note the endpoint from individual requests.
>
> -Philip

OK guys, I've been following this thread, because I've had this error 
pop up for years. Initially I tried to look it up and fix it. I could 
never find the true answer and it didn't affect my samba so I ignored it.

However, it seems that now you guys actually know what this error means. 
You've given a technical (programmers?) perspective of what has happened.

Is there any chance you can convert that to lamens terms?

i.e. Does anyone know how to get rid of the error message or how my 
configuration for samba is wrong?

P.S. I've had this configuration for samba more or less since RH 5.2 
(meaning I changed things as samba changed). This error did not show up 
till RH9 or Fedora 1 from memory. It's probably caused by something 
deprecated or the like in my smb.conf, but do we have any idea what?

Regards,
Ed.

[Craig White]

On Fri, 2005-10-21 at 10:37 +0800, Edward Dekkers wrote:
> Philip Prindeville wrote:
> > Rex Dieter wrote:
> > > Franck Y wrote:
> > > > I don t know waht does this thing mean.....like the "get_peer_addr "
> > > >
> > > > Can you excplain me thk you
> > > >
> > > > Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
> > > > lib/util_sock.c:get_peer_addr(1150)
> > > > Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
> > > > was Transport endpoint is not connected
> > >
> > > getpeername failed => failed DNS/hostname lookup.
> > >
> > > -- Rex
> >
> > No, that would be gethostbyaddr().  getpeername() is a system call
> > that looks for the address of the other side (the remove side) of an
> > association or socket pair.
> >
> > Of course, if you're using connectionless sockets (i.e. datagrams),
> > then it won't return an endpoint...  For that, you'd have to use
> > recvfrom() and note the endpoint from individual requests.
> >
> > -Philip
>
> OK guys, I've been following this thread, because I've had this error 
> pop up for years. Initially I tried to look it up and fix it. I could 
> never find the true answer and it didn't affect my samba so I ignored it.
>
> However, it seems that now you guys actually know what this error means. 
> You've given a technical (programmers?) perspective of what has happened.
>
> Is there any chance you can convert that to lamens terms?
>
> i.e. Does anyone know how to get rid of the error message or how my 
> configuration for samba is wrong?
>
> P.S. I've had this configuration for samba more or less since RH 5.2 
> (meaning I changed things as samba changed). This error did not show up 
> till RH9 or Fedora 1 from memory. It's probably caused by something 
> deprecated or the like in my smb.conf, but do we have any idea what?

----
No - the explanation was correct. Typically this will come from a Win2K
or WinXP client connection which will simultaneously connect to port 139
and 445 and drop one or the other as unnecessary - hence the log entry.
Samba developers sort of consider this to be rude client behavior. ;-)

If you want that type of activity to not be logged, then in the general
section, declare the smb port...

smb port = 139 #mix of Win95/98/2K/XP as Win95, Win98, WinME only
connect to port 139
or
smb port = 445 #win2K & WinXP clients only

the default is both ports are active for smb

personally, I would recommend that people not concern themselves with
the logged entries and leave it alone since it isn't broken.

If you want an in depth dissection of the ports that Microsoft uses, see
http://support.microsoft.com/kb/832017

Craig

[Franck Y]

Hello,
So we have to forget it ?

Franck
On 10/20/05, Craig White  wrote:
> On Fri, 2005-10-21 at 10:37 +0800, Edward Dekkers wrote:
> > Philip Prindeville wrote:
> > > Rex Dieter wrote:
> > > > Franck Y wrote:
> > > > > I don t know waht does this thing mean.....like the "get_peer_addr "
> > > > >
> > > > > Can you excplain me thk you
> > > > >
> > > > > Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
> > > > > lib/util_sock.c:get_peer_addr(1150)
> > > > > Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
> > > > > was Transport endpoint is not connected
> > > >
> > > > getpeername failed => failed DNS/hostname lookup.
> > > >
> > > > -- Rex
> > >
> > > No, that would be gethostbyaddr().  getpeername() is a system call
> > > that looks for the address of the other side (the remove side) of an
> > > association or socket pair.
> > >
> > > Of course, if you're using connectionless sockets (i.e. datagrams),
> > > then it won't return an endpoint...  For that, you'd have to use
> > > recvfrom() and note the endpoint from individual requests.
> > >
> > > -Philip
> >
> > OK guys, I've been following this thread, because I've had this error
> > pop up for years. Initially I tried to look it up and fix it. I could
> > never find the true answer and it didn't affect my samba so I ignored it.
> >
> > However, it seems that now you guys actually know what this error means.
> > You've given a technical (programmers?) perspective of what has happened.
> >
> > Is there any chance you can convert that to lamens terms?
> >
> > i.e. Does anyone know how to get rid of the error message or how my
> > configuration for samba is wrong?
> >
> > P.S. I've had this configuration for samba more or less since RH 5.2
> > (meaning I changed things as samba changed). This error did not show up
> > till RH9 or Fedora 1 from memory. It's probably caused by something
> > deprecated or the like in my smb.conf, but do we have any idea what?
>
> No - the explanation was correct. Typically this will come from a Win2K
> or WinXP client connection which will simultaneously connect to port 139
> and 445 and drop one or the other as unnecessary - hence the log entry.
> Samba developers sort of consider this to be rude client behavior. ;-)
>
> If you want that type of activity to not be logged, then in the general
> section, declare the smb port...
>
> smb port = 139 #mix of Win95/98/2K/XP as Win95, Win98, WinME only
> connect to port 139
> or
> smb port = 445 #win2K & WinXP clients only
>
> the default is both ports are active for smb
>
> personally, I would recommend that people not concern themselves with
> the logged entries and leave it alone since it isn't broken.
>
> If you want an in depth dissection of the ports that Microsoft uses, see
> http://support.microsoft.com/kb/832017
>
> Craig
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
>
> --
> fedora-list mailing list
> [e-mail]
> To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list

--
Franck

[Craig White]

Either that or make a better effort to comprehend my answer.

Craig

On Thu, 2005-11-17 at 15:42 -0500, Franck Y wrote:
> Hello,
> So we have to forget it ?
>
> Franck
> On 10/20/05, Craig White  wrote:
> > On Fri, 2005-10-21 at 10:37 +0800, Edward Dekkers wrote:
> > > Philip Prindeville wrote:
> > > > Rex Dieter wrote:
> > > > > Franck Y wrote:
> > > > > > I don t know waht does this thing mean.....like the "get_peer_addr "
> > > > > >
> > > > > > Can you excplain me thk you
> > > > > >
> > > > > > Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
> > > > > > lib/util_sock.c:get_peer_addr(1150)
> > > > > > Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
> > > > > > was Transport endpoint is not connected
> > > > >
> > > > > getpeername failed => failed DNS/hostname lookup.
> > > > >
> > > > > -- Rex
> > > >
> > > > No, that would be gethostbyaddr().  getpeername() is a system call
> > > > that looks for the address of the other side (the remove side) of an
> > > > association or socket pair.
> > > >
> > > > Of course, if you're using connectionless sockets (i.e. datagrams),
> > > > then it won't return an endpoint...  For that, you'd have to use
> > > > recvfrom() and note the endpoint from individual requests.
> > > >
> > > > -Philip
> > >
> > > OK guys, I've been following this thread, because I've had this error
> > > pop up for years. Initially I tried to look it up and fix it. I could
> > > never find the true answer and it didn't affect my samba so I ignored it.
> > >
> > > However, it seems that now you guys actually know what this error means.
> > > You've given a technical (programmers?) perspective of what has happened.
> > >
> > > Is there any chance you can convert that to lamens terms?
> > >
> > > i.e. Does anyone know how to get rid of the error message or how my
> > > configuration for samba is wrong?
> > >
> > > P.S. I've had this configuration for samba more or less since RH 5.2
> > > (meaning I changed things as samba changed). This error did not show up
> > > till RH9 or Fedora 1 from memory. It's probably caused by something
> > > deprecated or the like in my smb.conf, but do we have any idea what?
> >
> > No - the explanation was correct. Typically this will come from a Win2K
> > or WinXP client connection which will simultaneously connect to port 139
> > and 445 and drop one or the other as unnecessary - hence the log entry.
> > Samba developers sort of consider this to be rude client behavior. ;-)
> >
> > If you want that type of activity to not be logged, then in the general
> > section, declare the smb port...
> >
> > smb port = 139 #mix of Win95/98/2K/XP as Win95, Win98, WinME only
> > connect to port 139
> > or
> > smb port = 445 #win2K & WinXP clients only
> >
> > the default is both ports are active for smb
> >
> > personally, I would recommend that people not concern themselves with
> > the logged entries and leave it alone since it isn't broken.
> >
> > If you want an in depth dissection of the ports that Microsoft uses, see
> > http://support.microsoft.com/kb/832017
> >
> > Craig
> >
> > --
> > This message has been scanned for viruses and
> > dangerous content by MailScanner, and is
> > believed to be clean.
> >
> > --
> > fedora-list mailing list
> > [e-mail]
> > To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
>
> --
> Franck
>
> -- 
> fedora-list mailing list
> [e-mail]
> To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list

[Franck Y]

Hello
Thank you craig !
I read it 2 times more and now, I understand it !:-)
I only have win2000 and and XP client. But i have a network printer (Xerox),
and she is "writing" the same error.

So i will try tomorow to forces Xp/Win2k  to use the port = 445.

Thank you
Franck
On 11/17/05, Craig White  wrote:
> Either that or make a better effort to comprehend my answer.
>
> Craig
>
> On Thu, 2005-11-17 at 15:42 -0500, Franck Y wrote:
> > Hello,
> > So we have to forget it ?
> >
> > Franck
> > On 10/20/05, Craig White  wrote:
> > > On Fri, 2005-10-21 at 10:37 +0800, Edward Dekkers wrote:
> > > > Philip Prindeville wrote:
> > > > > Rex Dieter wrote:
> > > > > > Franck Y wrote:
> > > > > >
> > > > > > > I don t know waht does this thing mean.....like the "get_peer_addr "
> > > > > > >
> > > > > > > Can you excplain me thk you
> > > > > > >
> > > > > > > Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
> > > > > > > lib/util_sock.c:get_peer_addr(1150)
> > > > > > > Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
> > > > > > > was Transport endpoint is not connected
> > > > > >
> > > > > > getpeername failed => failed DNS/hostname lookup.
> > > > > >
> > > > > > -- Rex
> > > > >
> > > > > No, that would be gethostbyaddr().  getpeername() is a system call
> > > > > that looks for the address of the other side (the remove side) of an
> > > > > association or socket pair.
> > > > >
> > > > > Of course, if you're using connectionless sockets (i.e. datagrams),
> > > > > then it won't return an endpoint...  For that, you'd have to use
> > > > > recvfrom() and note the endpoint from individual requests.
> > > > >
> > > > > -Philip
> > > >
> > > > OK guys, I've been following this thread, because I've had this error
> > > > pop up for years. Initially I tried to look it up and fix it. I could
> > > > never find the true answer and it didn't affect my samba so I ignored it.
> > > >
> > > > However, it seems that now you guys actually know what this error means.
> > > > You've given a technical (programmers?) perspective of what has happened.
> > > >
> > > > Is there any chance you can convert that to lamens terms?
> > > >
> > > > i.e. Does anyone know how to get rid of the error message or how my
> > > > configuration for samba is wrong?
> > > >
> > > > P.S. I've had this configuration for samba more or less since RH 5.2
> > > > (meaning I changed things as samba changed). This error did not show up
> > > > till RH9 or Fedora 1 from memory. It's probably caused by something
> > > > deprecated or the like in my smb.conf, but do we have any idea what?
> > >
> > > No - the explanation was correct. Typically this will come from a Win2K
> > > or WinXP client connection which will simultaneously connect to port 139
> > > and 445 and drop one or the other as unnecessary - hence the log entry.
> > > Samba developers sort of consider this to be rude client behavior. ;-)
> > >
> > > If you want that type of activity to not be logged, then in the general
> > > section, declare the smb port...
> > >
> > > smb port = 139 #mix of Win95/98/2K/XP as Win95, Win98, WinME only
> > > connect to port 139
> > > or
> > > smb port = 445 #win2K & WinXP clients only
> > >
> > > the default is both ports are active for smb
> > >
> > > personally, I would recommend that people not concern themselves with
> > > the logged entries and leave it alone since it isn't broken.
> > >
> > > If you want an in depth dissection of the ports that Microsoft uses, see
> > > http://support.microsoft.com/kb/832017
> > >
> > > Craig
> > >
> > > --
> > > This message has been scanned for viruses and
> > > dangerous content by MailScanner, and is
> > > believed to be clean.
> > >
> > > --
> > > fedora-list mailing list
> > > [e-mail]
> > > To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
> >
> > --
> > Franck
> >
> > --
> > fedora-list mailing list
> > [e-mail]
> > To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
>
> --
> fedora-list mailing list
> [e-mail]
> To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list

--
Franck

[Michael H. Warfield]

On Thu, 2005-10-20 at 13:14 -0500, Rex Dieter wrote:
> Franck Y wrote:
> > I don t know waht does this thing mean.....like the "get_peer_addr "
> >
> > Can you excplain me thk you
> >
> > Oct 20 13:41:09 constellation smbd[3927]: [2005/10/20 13:41:09, 0]
> > lib/util_sock.c:get_peer_addr(1150)
> > Oct 20 13:41:09 constellation smbd[3927]:   getpeername failed. Error
> > was Transport endpoint is not connected
> getpeername failed => failed DNS/hostname lookup.

No...  Actually it means that the peer that connected to the Samba smb
socket on port 139 or 445 dropped the connections before the getpeername
function was called.  That's what the "Transport endpoint is not
connected" means.  I means that your peer went away before you could
look him up.  It's not uncommon with all the worm scanning going on for
ports 139/tcp and 445/tcp.  But it also seems to happen a lot with
normal clients.
> -- Rex

Mike