Strange message in syslog.log file !

[Ivan Azuara]

Hi !

I was checking the syslog.log file on my HPUX 11.00 server, and detect the next message:

Jan 13 16:42:05 s_inb_15 inetd[1360]: telnet/tcp: accept: No buffer space availa
ble
Jan 17 12:35:37 s_inb_15 inetd[1360]: telnet/tcp: accept: No buffer space availa
ble
Jan 17 12:35:47 s_inb_15 inetd[1360]: telnet/tcp: accept: No buffer space availa
ble

Look that always get the [1360] socket. Somebody knows if that is a problem, or maybe can be a sniffer program searching on my network ?

Best Regards!

[Sridhar Bhaskarla]

We did have couple of discussions on accept: No buffer space available error and most of the times we found that it has nothing to do with "ENOBUFS 233" error.

In one case the member posted that increasing nstrtel parameter fixed him problem. You may want to try it. Increase it by 50% and see if it helps.

1360 is the pid of your inetd process. It's not the socket.

-Sri

[Uday_S_Ankolekar]

Check in var/adm/nettl.LOG** for more information, It may be a duplicate ip address on the network conflicting.

Or If you get this error on a service guad node then check for package ip address..


-USA..

[rick jones · Accepted solution]

99 times out of 10, a "no buffer space available" error on an accept on HP-UX 11 and later means simply that by the time the application (in this case inetd) got around to calling accept(), the remote client had already closed the connection.

That stems from either the client being entirely too impatient, or the (in this case) inetd being too slow :) However, they are not "fatal" errors - inetd is still working, its listen endpoint is still good etc etc.

[rick jones]

btw that number in square brackets is not a socket ID. it is a process id.