bak@csd_v.UUCP (Bruce) (11/22/87)
Perhaps I am reinventing the wheel here, but I am trying to write a utility that polls a system and continues to retry every so often ( default 1 hour intervals ) until a successful connection is established. I am doing this because if I just issue a $ uucico -r1 sysname & or some equivalent like uusub, etc., the connection is only tryed three times if the destination is busy and then the poll goes away. I would like a utility that continues to exec uucico and then reschedule itself until uucico successfully connects to the destination system. My problems are (1) is this really necessary or is there some uucp facility ( not HONEYDANBER ) which incorporates this function and (2) if not, how do I tell whether or not uucico has successfully connected to the target system. Uucico seems to exit 0 regardless of the circumstances. Any help would be appreciated. -- Bruce Kern | uunet!swlabs!csd_v!bak Computer Systems Design | 1-203-270-0399 29 High Rock Rd., Sandy Hook, Ct. 06482 | This space for rent.
dhesi@bsu-cs.UUCP (Rahul Dhesi) (11/23/87)
In article <134@csd_v.UUCP> bak@csd_v.UUCP (Bruce) writes: >Perhaps I am reinventing the wheel here, but I am trying to write >a utility that polls a system and continues to retry every so often >( default 1 hour intervals ) until a successful connection is established. If you're using 4.3BSD, use the "uupoll" command to queue a null request, and invoke uucico hourly from cron. If you're using System V Release 2, create a uupoll script that includes the following lines: pollfile=/usr/spool/uucp/C.${system}ACPOLL touch $pollfile && chmod a+w $pollfile Here, the variable "system" contains the name of the system for which you wish to queue a null request. (My uupoll script actually validates the system name against those returned by the uuname command, so a mistyped system name doesn't cause trouble.) Invoke this uupoll script when you want to queue a null request. Then whenever uucico is invoked it will attempt to call the system, and will keep on doing so until the connection is successful. Also, since you will be invoking uucico hourly, you may wish to add the -r flag to all uucp requests, so they are all queued instead of invoking uucico immediately. (This flag may be undocumented for some commands.) If you are using smail, add the -r flag to where it calls uux. I don't know how System V Release 3 handles this. -- Rahul Dhesi UUCP: <backbones>!{iuvax,pur-ee,uunet}!bsu-cs!dhesi
zeeff@b-tech.UUCP (Jon Zeeff) (11/27/87)
In article <134@csd_v.UUCP> bak@csd_v.UUCP (Bruce) writes: >Perhaps I am reinventing the wheel here, but I am trying to write >a utility that polls a system and continues to retry every so often >( default 1 hour intervals ) until a successful connection is established. > I found that the most reliable way to tell if a connection has taken place is uusub (SysV.2 w/o HDB). So you get something like: # Execute this script just after midnight with "poll time sites1 site2 .." # -- Jon Zeeff (umix!b-tech!zeeff) . /etc/TIMEZONE quit=$1 shift today=`date +%d` # loop until the quit time while test `date +%H` -le $quit do for j in $* do # trim to 6 chars i=`expr "$j" : '\(.\{1,6\}\).*'` laston=`uustat -M$j` laston=`expr "$laston" : '.*/\([0-9].\).*' 2> /dev/null` if test $laston -ne $today 2> /dev/null then /usr/lib/uucp/uucico -r1 -s$j status=$? # Code to handle the case where the uukill daemon noticed that # disk space was running out and killed incoming uucp connections # if it was killed, then disk is full and we shouldn't call for awhile if test status -eq 15 then sleep 5000 fi # wait in case uuxqt is going to trigger something sleep 40 fi done # Wait a while before trying again sleep 1000 done -- Jon Zeeff Branch Technology, uunet!umix!b-tech!zeeff zeeff%b-tech.uucp@umix.cc.umich.edu
giebelhaus@hi-csc.UUCP (Timothy R. Giebelhaus) (11/29/87)
I use this script. It works very well for me. It also calls again if
the call was not finished (failed a normal termination). One day I
plan to change it just a bit so that if it does turn out the call is
already in progress when the script starts, it checks to see if the
call terminated normally. Now, if the call was already made by the
remote host or uu.hourly, it just drops through.
#! /bin/sh
#
# Call uunet for news
newssystem=uunet
uucico -r1 -s${newssystem}
#
# Try to call again a few more times if I have to
# sleep first to be sure the STST and LCK files are stable.
sleep 60
ststfile=/usr/spool/uucp/STST.${newssystem}
lckfile=/usr/spool/uucp/LCK..${newssystem}
maxtries=aaaaaaaaaaaaaaaa
tries=a
while test -f $ststfile -a $tries != $maxtries -a ! -f $lckfile
do
echo "Failed, STST = `cat $ststfile`; tries = $tries" | mail root
sleep 900
tries=${tries}a
rm $ststfile
uucico -r1 -s${newssystem}
sleep 60
done
# Notify me if the news did not finish
if test $tries = $maxtries
then
echo "Failed to call $newssystem for news" | mail root
fi
# Notify me if the news was already running.
if test -f $lckfile
then
echo "LCK file exists. STST = `cat $ststfile`" | mail root
fi
--
---------------------------------
UUCP: {uunet, ihnp4!umn-cs}!hi-csc!giebelhaus
ARPA: hi-csc!giebelhaus@umn-cs.arpa
Nobody I know admits to sharing my opinions. I don't even have a pet.