Re: ntpd confusion
On Nov 27, 2014, at 11:00 AM, mad m...@sharktooth.de wrote: I found it. It was DHCP. The NTP init scripts checks if there is a file /var/lib/ntp/ntp.conf.dhcp in which the local router is configured as only ntp source. The DHCP client package in Debian tries to get as much information as possible out of the DHCP transaction. It then makes it available to as many other packages as it can. I think this is a conscious decision on the part of the Debian Developers, and overall, I support it. But it sometimes leads to surprises, such as we have seen here. Specifically, for NTP, when a new NIC comes online and it’s configured for DHCP, the DHCP client checks to see if there is an NTP server provided in the DHCP response packet. If there is, it copies /etc/ntp.conf to /var/lib/ntp/ntp.conf then strips out *all* lines beginning with “server”. It then puts at the top of the file a single new “server” line directing ntpd to use the DHCP provided server. When the ntp init.d script runs, it checks to see if there is a /var/lib/ntp/ntp.conf.dhcp file, and uses it in place of the conventional /etc/ntp.conf. The result is what you saw. If you don’t like that behavior, there is a workaround. With the ntp version distributed in Wheezy (and, possibly, also Lenny) the ntp.conf file can have a “pool” directive, which is *not* messed with by the DHCP scripts. So, in /etc/ntp.conf replace the four lines server 0.debian.pool.ntp.org iburst server 1.debian.pool.ntp.org iburst server 2.debian.pool.ntp.org iburst server 3.debian.pool.ntp.org iburst with the one line pool us.pool.ntp.org iburst preempt If you aren’t in the US, try using your own country code in place of “us” as long as host CC.pool.ntp.org indicates it exists. If not, try some nearby countries. Or, as a complete last resort, you can use just plain “pool.ntp.org” without a country code. So, assuming you are in Germany, you would use pool de.pool.ntp.org iburst preempt The mechanism behind the “pool” directive is described in more detail at http://www.eecis.udel.edu/~mills/ntp/html/discover.html#pool Enjoy! Rick -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/93bda098-cd18-4cbe-9456-2828663a9...@pobox.com
Re: ntpd confusion
I found it. It was DHCP. The NTP init scripts checks if there is a file /var/lib/ntp/ntp.conf.dhcp in which the local router is configured as only ntp source. Am 11/24/14 16:11, schrieb Chen Wei: On Mon, Nov 24, 2014 at 10:58:29AM +0100, mad wrote: mentioned, other Debian installations not on my home network, with the same configuration show as expected four clock sources. Even starting ntpd on the command line doesn't show any more data and ntpd is compiled without debugging. Probably that is what I will do, recompile ntpd with debug enabled and then see what ntpd is actually doing. Besides recompile with debug level option, I would also suggest try the NTP mailing list. Could it have something to do with upnp, zeroconf or something like that? No idea. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/547774ca.8070...@sharktooth.de
Re: ntpd confusion
mad wrote: I found it. It was DHCP. The NTP init scripts checks if there is a file /var/lib/ntp/ntp.conf.dhcp in which the local router is configured as only ntp source. Thank you for posting the resolution of the problem! Bob signature.asc Description: Digital signature
Re: ntpd confusion
Am 24.11.2014 um 06:13 schrieb Chen Wei: On Sat, Nov 22, 2014 at 09:57:50PM +0100, mad wrote: # ntpq -p remote refidst t when poll reach delay offset jitter fritz.box X.Y.Z.A 3 u- 6411.8740.153 0.052 I use the default ntp configuration and other Debian installations directly on the internet use all four clock sources (0.debian.pool, 1.debian.pool...). Why not 1) double check /etc/ntp.conf, make sure lines such as server 0.debian.pool.ntp.org exist. Already done that multiple times. 2) verify remote ntp server is reachable, # nmap -sU -p123 0.debian.pool.ntp.org # nmap -sU -p123 0.debian.pool.ntp.org Host is up (0.027s latency). Other addresses for 0.debian.pool.ntp.org (not scanned): 85.10.246.226 141.30.228.4 192.53.103.108 rDNS record for 37.120.166.3: olymp.auf-feindgebiet.de PORTSTATE SERVICE 123/udp open ntp Nmap done: 1 IP address (1 host up) scanned in 1.70 seconds On my system, even the unreachable server shows in ntpq -p output. That is my problem! Not on my internal home network systems. As mentioned, other Debian installations not on my home network, with the same configuration show as expected four clock sources. Even starting ntpd on the command line doesn't show any more data and ntpd is compiled without debugging. Probably that is what I will do, recompile ntpd with debug enabled and then see what ntpd is actually doing. Could it have something to do with upnp, zeroconf or something like that? mad -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/54730145.8020...@sharktooth.de
Re: ntpd confusion
On Mon, Nov 24, 2014 at 10:58:29AM +0100, mad wrote: mentioned, other Debian installations not on my home network, with the same configuration show as expected four clock sources. Even starting ntpd on the command line doesn't show any more data and ntpd is compiled without debugging. Probably that is what I will do, recompile ntpd with debug enabled and then see what ntpd is actually doing. Besides recompile with debug level option, I would also suggest try the NTP mailing list. Could it have something to do with upnp, zeroconf or something like that? No idea. -- Chen Wei -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20141124151104.GA16480@localhost
Re: ntpd confusion
Am 22.11.2014 um 22:01 schrieb Erwan David: Le 22/11/2014 21:57, mad a écrit : Hi! I'm stumped. On my home network all my Debian installations _only_ use the router as clock source. # ntpq -p remote refidst t when poll reach delay offset jitter fritz.box X.Y.Z.A 3 u- 6411.8740.153 0.052 I use the default ntp configuration and other Debian installations directly on the internet use all four clock sources (0.debian.pool, 1.debian.pool...). Why is that? It seems to have something to do with IPv4 and IPv6. If I start ntpd with '-4 -n' ntpq exists with 'Connection refused'. If I start it with '-6 -n' ntpq says 'No association ID's returned'. I'm thoroughly confused. Does anyone knows what's happening? TIA mad Does your router announce itself as ntp server in the dhcp answer ? No, at least that is what I read from the answer (removed several lines): # dhcpdump -i eth0 IP: A.B.C.25 A.B.C.1 OP: 1 (BOOTPREQUEST) HTYPE: 1 (Ethernet) HLEN: 6 HOPS: 0 XID: 4cf07154 SECS: 0 FLAGS: 0 CIADDR: 0.0.0.0 YIADDR: 0.0.0.0 SIADDR: 0.0.0.0 GIADDR: 0.0.0.0 SNAME: . FNAME: . OPTION: 53 ( 1) DHCP message type 3 (DHCPREQUEST) OPTION: 50 ( 4) Request IP address0.0.0.0 --- TIME: 2014-11-23 15:33:49.014 IP: A.B.C.1 255.255.255.255 OP: 2 (BOOTPREPLY) HTYPE: 1 (Ethernet) HLEN: 6 HOPS: 0 XID: 4cf07154 SECS: 0 FLAGS: 7f80 CIADDR: 0.0.0.0 YIADDR: 0.0.0.0 SIADDR: A.B.C.1 GIADDR: 0.0.0.0 SNAME: . FNAME: . OPTION: 53 ( 1) DHCP message type 6 (DHCPNAK) --- TIME: 2014-11-23 15:33:49.014 IP: A.B.C.25 A.B.C.1 OP: 1 (BOOTPREQUEST) HTYPE: 1 (Ethernet) HLEN: 6 HOPS: 0 XID: 4cf07154 SECS: 0 FLAGS: 0 CIADDR: 0.0.0.0 YIADDR: 0.0.0.0 SIADDR: 0.0.0.0 GIADDR: 0.0.0.0 SNAME: . FNAME: . OPTION: 53 ( 1) DHCP message type 7 (DHCPRELEASE) --- -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5471f19d.4010...@sharktooth.de
Re: ntpd confusion
mad wrote: I use the default ntp configuration and other Debian installations directly on the internet use all four clock sources (0.debian.pool, 1.debian.pool...). Should see some remapped names from the pool in the list then. Why is that? It seems to have something to do with IPv4 and IPv6. If I start ntpd with '-4 -n' ntpq exists with 'Connection refused'. If I start it with '-6 -n' ntpq says 'No association ID's returned'. I'm thoroughly confused. Does anyone knows what's happening? No idea. Works for me. Is your router blocking ntp port 123 and preventing them from connecting? Bob Example from a machine here using the pool. # ntpq -p remote refid st t when poll reach delay offset jitter == +time.tritn.com 66.220.9.122 2 u 651 1024 377 68.8241.508 22.071 *pool-test.ntp.o 204.123.2.72 2 u 439 1024 377 69.1641.546 1.088 -ntp2.pcloud.com 43.77.130.2542 u 568 1024 377 74.165 -12.416 17.309 +tssnet1.tss.net 204.123.2.5 2 u 940 1024 377 74.379 -0.202 0.965 signature.asc Description: Digital signature
Re: ntpd confusion
On Sat, Nov 22, 2014 at 09:57:50PM +0100, mad wrote: # ntpq -p remote refidst t when poll reach delay offset jitter fritz.box X.Y.Z.A 3 u- 6411.8740.153 0.052 I use the default ntp configuration and other Debian installations directly on the internet use all four clock sources (0.debian.pool, 1.debian.pool...). Why not 1) double check /etc/ntp.conf, make sure lines such as server 0.debian.pool.ntp.org exist. 2) verify remote ntp server is reachable, # nmap -sU -p123 0.debian.pool.ntp.org On my system, even the unreachable server shows in ntpq -p output. -- Chen Wei -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20141124051325.GA6657@localhost
ntpd confusion
Hi! I'm stumped. On my home network all my Debian installations _only_ use the router as clock source. # ntpq -p remote refidst t when poll reach delay offset jitter fritz.box X.Y.Z.A 3 u- 6411.8740.153 0.052 I use the default ntp configuration and other Debian installations directly on the internet use all four clock sources (0.debian.pool, 1.debian.pool...). Why is that? It seems to have something to do with IPv4 and IPv6. If I start ntpd with '-4 -n' ntpq exists with 'Connection refused'. If I start it with '-6 -n' ntpq says 'No association ID's returned'. I'm thoroughly confused. Does anyone knows what's happening? TIA mad -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5470f8ce.5000...@sharktooth.de
Re: ntpd confusion
Le 22/11/2014 21:57, mad a écrit : Hi! I'm stumped. On my home network all my Debian installations _only_ use the router as clock source. # ntpq -p remote refidst t when poll reach delay offset jitter fritz.box X.Y.Z.A 3 u- 6411.8740.153 0.052 I use the default ntp configuration and other Debian installations directly on the internet use all four clock sources (0.debian.pool, 1.debian.pool...). Why is that? It seems to have something to do with IPv4 and IPv6. If I start ntpd with '-4 -n' ntpq exists with 'Connection refused'. If I start it with '-6 -n' ntpq says 'No association ID's returned'. I'm thoroughly confused. Does anyone knows what's happening? TIA mad DOes your router announce itself as ntp server in the dhcp answer ? -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5470f9c3@rail.eu.org
