Date: Fri, 21 Jan 2005 15:51:21 +1100 (EST) From: Bsd B <freebsd04@yahoo.com.au> To: freebsd-questions@freebsd.org Subject: ntp is acting strangely Message-ID: <20050121045121.92762.qmail@web54007.mail.yahoo.com>
next in thread | raw e-mail | index | archive | help
Hi, I am baffled about an ntp problem i am experiencing. Please excuse me if this is a simple question, as I am still a relatively inexperienced user compared to the community. My setup is 3 computers (2 act as game clients and the other acts as a game server) all running FreeBSD4.9 RELEASE. They are each connected to the internet (through the network) and through wireless (private network). Each connection is fully functioning (ie internet connection is good and they can ping each other) The experiment i wanted to carry out was to just have a look at the ntp offsets over time for the three computers, and comparing them to each other. I have written a very simple shell script that simply outputs the date and time along with the result of the ntp offset of the server being synched to (ie the server with a asterik(*) from the result of "ntpq -p"), every 30 seconds. The servers I have selected in /etc/ntp.conf for all 3 computers are reliable stratum 2 university ntp servers. (2 are in the x.x.1 network and the other is on the x.x.20 network) I ran the script on all 3 computers and over a week (from Jan 5 to Jan 13), they seemed to run fine (fluctuating offsets within the range of -50 to 50, however, towards the end of the week, all 3 tests ran showed missing data for about 1/2 day, as if there was some sort of outage. Upon investigation, I found only one of the client machines had the following in /var/log/messages (it is an extract, starting from the first ntpd related entry - it is the exact data except i was advised to edit the ip addresses out - please excuse me) Jan 11 22:34:46 client1 ntpd[95]: sendto(ntp server 1 x.x.20.x): No route to host Jan 11 22:39:04 client1 ntpd[95]: sendto(ntp server 1 x.x.20.x): No route to host Jan 11 23:26:13 client1 ntpd[95]: sendto(ntp server 1 x.x.20.x): No route to host Jan 12 03:06:05 client1 ntpd[95]: sendto(ntp server 2 x.x.1.x): No route to host Jan 12 07:53:46 client1 ntpd[95]: sendto(ntp server 3 x.x.1.x): No route to host Jan 12 11:08:14 client1 ntpd[95]: time reset -0.189678 s Looking at the ntp offset result files, the missing data began at approx 11 Jan 2005 22:29:58 (showing an ntp offset of -5.332) and resumed approx 12 Jan 2005 10:32:26 with an offset of 46.081. The offset gradually decreased to -172.70 at 11:07:57 before dropping out again and resuming minutes later with a offset of -5.712. After that it seemed fine. I don't believe it was just simply the case of 3 of the ntp servers going down. Also, as stated previously, 3 computers are pinging fine, and the script is so simple and basic that it can't be the problem. A previous test i ran was even stranger...one of the computers produced the desired results, one of the computers received frequent drop outs and the other only had a small period of results. Question - is it common to have ntp offsets not providing data? I specified a drift file (/var/db/ntp.dirft) in /etc/ntp.conf, which i thought was used to continue adjusting the offset even when contact was lost with all servers. So how do I fix this problem or more importantly, what is the cause of it? Is it freebsd? I have been searching the internet for the last couple of days and I am yet to find some solid literature or solutions. I am very keen to get on top of this. Any suggestions/ideas/pointers are much appreciated. Deep thanks in advance Bob __________________________________ Do you Yahoo!? The all-new My Yahoo! - What will yours do? http://my.yahoo.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050121045121.92762.qmail>