Thanks to GertJan Hagenaars for helping me troubleshoot so far: OK: original symptoms: 1 machine, cron stops logging, /var/adm/messages dissapears. doing a /etc/init.d stop (then start) (doesnt always restart, may start for 10-15 mins, then dies again) More info: doing : script % strace -f -p proc id % ^C % exit Shows me a great message: namely: "ERROR: unable to open /dev/log" and if I look at it, I get: mbruntel@hummer: /etc $> ls -la /dev/log lrwxrwxrwx 1 root other 27 Mar 17 2000 /dev/log -> ../devices/pseudo/log@0:log mbruntel@hummer: /etc $> ls -la /devices/pseudo/log@0:log crw-r----- 1 root staff 21, 5 Mar 17 2000 /devices/pseudo/log@0:log ========================^^^^^^^^^========= staff is my "default" programmer group. SO I changed it to the way my other systems look; (chgrp sys) to get crw-r----- 1 root sys 21, 5 Mar 17 2000 /devices/pseudo/log@0:log and do an ls -la again : ls -la /var/adm/messages and I get: /var/adm/messages: No such file or directory Another suggestion was to see if lsof shows anything it shows nothing that I can However, I do get this message: lastlog_get_entry: Error reading from /var/adm/lastlog: Bad file number lastlog_get_entry: Error reading from /var/adm/lastlog: Bad file number Thanks for any additional insights that you can share with me! Mitch Bruntel --original message: -- % -----Original Message----- % From: GertJan Hagenaars [mailto:gj@hagenaars.com] % Sent: Thursday, September 18, 2003 11:18 PM % To: Bruntel, Mitchell L, ALABS % Subject: Re: /var/adm/messages stops, dissapears. % % % Hi Mitch, % % um... I wouldn't copy anything on top of messages/syslog/maillog if % you're having a problem with it. Use this instead (shorter, because you % don't have to check if the file is there or not, and quicker, because it % will create it with the right permissions (if memory serves)). % % umask 133 % touch /var/adm/message % touch /var/adm/syslog % touch /var/adm/maillog % % Meanwhile (and this will be a hassle), get lsof from www.sunfreeware.com % and install it. See if it can tell you if something else is using those % files. % % Also, find syslog's processid, and do this: % % script % strace -f -p 12345 % ^C % exit % % to find out if syslog is doing this by itself (virtually unheard of). % script will create a file called "typescript" with all the system calls % that make up syslog. If syslog is doing it, you should see an "unlink" % in there somewhere. % % Also, see if there's a specific time when this happens (12 minutes past % the hour? 33 minutes past the hour?, and have a look at your crontabs. % % Last thing, if none of the above really help, there's a trick you can % use. Syslog opens the log files and keeps them open while it's still % alive. If someone removes the file, you will not free up the space, % because syslog keeps the file descriptor open (lsof will show you this). % Soooooooooooooooooooooo... if you do the following (once) : % % cd /var/adm % ln messages safe.messages % ln syslog safe.syslog % ln maillog safe.maillog % % you have created a new name for the files that (likely) will not be % removed. When the real names are gone, the only thing you need to do to % get them back (before you run logcheck, and before you restart syslog) % is this: % % for i in messages syslog maillog % do % if [ ! -f /var/adm/$i ]; then % ln /var/adm/safe.$i /var/adm/$i % fi % done % % actually, you can take the above snippet and run that from cron as well. % % It doesn't solve the problem, but at least it shouldn't cause you too % many headaches. % % Apparently, Bruntel, Mitchell L, ALABS wrote: % % interesting problem with my syslog. (frustrating I mean) % % On one system I have the /var/adm/messages file keeps disapearing % % My log shows: % % Unusual System Events % % =-=-=-=-=-=-=-=-=-=-= % % File /var/adm/syslog cannot be read. % % File /var/adm/messages cannot be read. % % % % and I can do a restart (stop/restart) of syslog by doing a /etc/init.d/syslog % % stop, then a start % % % % and it is seen (since this problem started I added the following lines to the % % stock file: % % (based on original creation above) % % if [ ! -f /var/adm/messages ]; then % % /usr/bin/cp /dev/null /var/adm/messages % % /usr/bin/chmod 0644 /var/adm/messages % % fi % % if [ ! -f /var/adm/syslog ]; then % % /usr/bin/cp /dev/null /var/adm/syslog % % /usr/bin/chmod 0644 /var/adm/syslog % % fi % % if [ ! -f /var/adm/maillog ]; then % % /usr/bin/cp /dev/null /var/adm/maillog % % /usr/bin/chmod 0644 /var/adm/maillog % % % % so it seems to restart ok, if only for another 10 mins or so. % % % % I've tried running a log of when it's stopping, lasts about 10-12 mins, then % % needs restart. % % process list when it stops doesnt show any immediate candidates for cause of % % problem. % % % % I'm kind of stumped? % % ideas, suggestions, welcome. % % % % I've included my /etc/syslog.conf to see if someone ELSE see's something funky % % that might be doing this, but I've % % been over it 5-6 times, and dont see much. % % % % Thanks again, and I did try a few searches on the list, but didnt come up with % % anything % % % % Oh yes, the syslogd daemon doesnt stop running, it's just that /var/messages % % disapears! % % % % Mitch Bruntel % % AT&T Labs % % % % ++++++++++++++++++++++++++++++++ % % /etc/syslog.conf follows: % % #ident "@(#)syslog.conf 1.5 99/02/03 SMI" /* SunOS 5.0 */ % % # % % # Copyright (c) 1991-1999 by Sun Microsystems, Inc. % % # All rights reserved. % % # % % # syslog configuration file. % % # % % # This file is processed by m4 so be careful to quote (`') names % % # that match m4 reserved words. Also, within ifdef's, arguments % % # containing commas must be quoted. % % # % % # next is original % % #*.info /var/adm/messages % % *.info;mail.info /var/adm/messages % % #mail.info /var/adm/maillog % % *.err;kern.notice /dev/sysmsg % % *.err;kern.debug;daemon.notice;mail.crit /var/adm/messages % % #kern, mark.debug /var/adm/messages % % # prev line for kernel mark % % % % *.alert;kern.err;daemon.err operator % % *.alert % % *.emerg * % % % % # if a non-loghost machine chooses to have authentication messages % % # sent to the loghost machine, un-comment out the following line: % % #auth.notice ifdef(`LOGHOST', /var/log/authlog, @loghost) % %% % #mail.debug ifdef(`LOGHOST', /var/log/syslog, @loghost) % % % % # % % # non-loghost machines will use the following lines to cause "user" % % # log messages to be logged locally. % % # % % ifdef(`LOGHOST', , % % user.err /dev/sysmsg % % user.err /var/adm/messages % % user.alert `root, operator' % % user.emerg * % % ) % % auth.debug /var/log/authlog % % auth.notice /var/log/authlog % % auth.info /var/log/authlog % % _______________________________________________ % % sunmanagers mailing list % % sunmanagers@sunmanagers.org % % http://www.sunmanagers.org/mailman/listinfo/sunmanagers % % CHeers, % GertJan. % _______________________________________________ sunmanagers mailing list sunmanagers@sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagersReceived on Thu Sep 25 13:09:04 2003
This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:43:20 EST