On Thu, Sep 28, 2000 at 11:23:44AM +0100, Colin Watson wrote: > will trillich <[EMAIL PROTECTED]> wrote: > >i've not deciphered some of these XINETD messages just yet-- > >if you can translate for me, or point me to The Path Of Enlightenment, > >lemme know: > > > >Sep 27 20:37:20 server xinetd[883]: bind failed (Address already in use > >(errno = 98)). service = telnet > >Sep 27 20:37:20 server xinetd[883]: bind failed (Address already in use > >(errno = 98)). service = smtp > >Sep 27 20:37:20 server xinetd[883]: bind failed (Address already in use > >(errno = 98)). service = ident > > > >how do i find out what's already "using an address?" > >and how do i become convinced that ANYTHING is using it now? > > Try 'netstat -a', or 'netstat -an' if it bogs down trying to do DNS > lookups. You might have to look for the port numbers from /etc/services > rather than the service names. (It will throw several pages of > information at you on a busy server.)
aha. very helpful gadget there, that 'netstat' command. Active Internet connections (servers and established) Proto Recv-Q Send-Q Local Address Foreign Address State tcp 0 0 server:4658 mail.pinncomp.net:pop3 TIME_WAIT tcp 0 0 localhost:4656 localhost:smtp TIME_WAIT tcp 0 0 server:4655 mail.speedex.net:imap2 TIME_WAIT tcp 0 1 server:www ftp.colormax.net:59242 FIN_WAIT1 tcp 0 20 linus:ssh wdt:2048 ESTABLISHED tcp 0 0 *:auth *:* LISTEN tcp 0 0 *:pop3 *:* LISTEN tcp 0 0 *:smtp *:* LISTEN tcp 0 0 *:ftp *:* LISTEN tcp 0 0 *:time *:* LISTEN tcp 0 0 *:daytime *:* LISTEN tcp 0 0 *:discard *:* LISTEN tcp 0 0 server:domain *:* LISTEN tcp 0 0 linus:domain *:* LISTEN tcp 0 0 1.2.3.4:domain *:* LISTEN tcp 0 0 localhost:domain *:* LISTEN tcp 0 0 *:ssh *:* LISTEN tcp 0 0 linus:www *:* LISTEN tcp 0 0 server:www *:* LISTEN tcp 0 0 localhost:www *:* LISTEN tcp 0 0 *:mysql *:* LISTEN udp 0 0 *:discard *:* udp 0 0 *:1141 *:* udp 0 0 server:domain *:* udp 0 0 linus:domain *:* udp 0 0 1.2.3.4:domain *:* udp 0 0 localhost:domain *:* raw 0 0 *:icmp *:* 7 raw 0 0 *:tcp *:* 7 raw 0 0 *:ipv6-icmp *:* 7 raw 0 0 *:ipv6-icmp *:* 7 raw 0 0 *:ipv6-icmp *:* 7 in tinkering with tunl0 i got it up and running on 1.2.3.4 which is bogus; now i can't kill it. (don't think that' related to my problem, tho...) > The messages above indicate that some other service on your machine is > already listening on the given ports. > > >the only service i use 'bind' in is telnet, to restrict it to my > >secure internal 192.168.*.*, so it won't show up on any nmap scan > >of my internet-visible ip address... > > bind() is a system call; *everything* that listens on a socket calls it. > 'man 2 bind' if you need to do socket programming. confusing -- in /etc/xinetd.conf there's a "bind" directive that allows you to tell it "listen on this interface but not that one"... > >Sep 27 20:52:00 server xinetd[1005]: warning: can't get client address: > >Invalid argument > >Sep 27 20:54:51 server xinetd[1005]: warning: can't get client address: > >Invalid argument > > > >argument... from where? what client address are we looking for? > > That's how the C library describes the EINVAL error code, which is a > fairly vague error saying "you passed the wrong argument to this > function". I can't see anywhere in the source code where xinetd prints > this message, though; maybe it's something xinetd calls, or maybe I'm > missing something. thanks for looking! i notice that it only happens when i'm logged in, not when there's server stuff going on by itself (http, ftp, etc). only when i've got a shell login is there any trouble. and now i notice in daemon.log that the server gets re-started after every 'can't get client address...' Sep 28 04:24:55 server xinetd[1005]: warning: can't get client address: Invalid argument Sep 28 04:24:55 server identd[3726]: started Sep 28 04:28:35 server xinetd[1005]: warning: can't get client address: Invalid argument Sep 28 04:28:35 server identd[3771]: started Sep 28 04:34:44 server xinetd[1005]: warning: can't get client address: Invalid argument Sep 28 04:34:44 server identd[4179]: started (logcheck filters out the non-alarming lines, so i didn't see this earlier. may not be significant...) anybody else have some clues?
