Re: tcpsvd and ipsvd-instruct
"M.S. Lucas" <[email protected]>
| Newsgroups | gmane.comp.misc.pape.general |
|---|---|
| Organization | TAOS-IT |
| Message-ID | <00a001c43985$5b10fc90$0301a8c0@lepus> |
Gerrit Pape wrote: > On Thu, May 13, 2004 at 01:10:06PM -0400, Charlie Brady wrote: >> On Thu, 13 May 2004, M.S. Lucas wrote: >>> @4000000040a3a8411cb00074 tcpsvd: info: status 1/20 >>> @4000000040a3a8411cb265bc tcpsvd: info: pid 4470 from 192.168.1.3 >>> @4000000040a3a8481512eb9c tcpsvd: info: start 4470 :192.168.1.250 >>>>> 192.168.1.3:3497 /var/qmail/control/peers/0: >>> @4000000040a3a8481517a68c tcpsvd: info: end 4470 exit 111 >>> @4000000040a3a84815185e24 tcpsvd: info: status 0/20 >> ... >>> And another failure. >> > I don't think it's tcpsvd that exits 111 here, it would print an error > message, 111 normally indicates a temporary error. pid xxx in the > tcpsvd log means it forked for a connection and has pid xxx; start xxx > means it now runs the default program, specified at the command line. > But rblsmtpd, recordio, and smtpfront-qmail also shouldn't exit 111 > without printing an error message. M.S., could you please run a > system call trace to find out which program exits 111?, e.g.: > > /usr/local/bin/tcpsvd -vv -u "$QMAILDUID":"$NOFILESGID" -c > "$MAXSMTPD" \ -x /var/qmail/control/peers.cdb 0 smtp \ > strace -f -o/tmp/trace /usr/local/bin/rblsmtpd -b -C \ > -r bl.spamcop.net recordio /usr/local/bin/smtpfront-qmail It took some time (and sleep) to figure this out but the first results are looking toward tcpsvd isn't starting the program cause he thinks the program to start is empty. Somewhere in the start line there was something wrong where tcpserver didn't had any problems with. using the following part gives the error about "unable to run" > "$MAXSMTPD" \ -x /var/qmail/control/peers.cdb 0 smtp \ > strace -f -o/tmp/trace /usr/local/bin/rblsmtpd -b -C \ The (temp) sollution was to put the strace on the same line as the "0 smtp" part and everything worked. Breaking up the line and reinserting the "\" didn't reinsert the problem. Conclusion: There was something wrong in the two lines shown above but what isn't known. Everything is working now with the -x or the -i command with kind regards, Maurice Lucas TAOS-IT