| View previous topic :: View next topic |
| Author |
Message |
ben_dash
Apprentice
Joined: 12 Aug 2003
Posts: 239
|
 Posted: Fri Aug 04, 2006 3:07 am Post subject: ntpd loses connection to pool.ntp.org |
 |
|
I have noticed that my Gentoo router's ntpd tends to lose its connectivity to the *.pool.ntp.org stratum 3 servers every few weeks or months.
This isn't a big deal, as I usually notice fairly quickly as my MthTV server is one of its ntp clients and time drift is obvious when watching recordings that start to begin either a couple of seconds too early or a couple of seconds to late.
Here's my "ntpq -c pe" output after stopping ntpd on the router and re-running "ntpdate pool.ntp.org":
| Code: | alpha root # ntpq -c pe
remote refid st t when poll reach delay offset jitter
==============================================================================
a.pool.ntp.uq.e 130.102.152.7 2 u 213 256 17 217.831 6.573 1.890
ntp.cis.rv.ua 129.69.1.153 2 u 210 256 17 175.226 -3.592 3.062
evenstar.silver 192.36.134.25 2 u 212 256 17 147.978 -4.070 1.074
alpha root # ntpq -c rv |grep stratum
processor="i686", system="Linux/2.6.11-gentoo-r9", leap=11, stratum=16,
alpha root # ntpq -c rv |grep stratum
processor="i686", system="Linux/2.6.11-gentoo-r9", leap=11, stratum=16,
alpha root # ntpq -c rv |grep stratum
processor="i686", system="Linux/2.6.11-gentoo-r9", leap=00, stratum=3,
alpha root # ntpq -c pe
remote refid st t when poll reach delay offset jitter
==============================================================================
+a.pool.ntp.uq.e 130.102.152.7 2 u 187 512 37 217.831 6.573 1.845
*ntp.cis.rv.ua 129.69.1.153 2 u 187 512 37 174.946 -4.267 3.245
+evenstar.silver 192.36.134.25 2 u 186 512 37 147.978 -4.070 0.937
|
I understand that these servers have a certain amount of unreliability, and that's fine, but it seems that after a couple of months I'm 100% guaranteed to loose connectivity to all three servers that *.pool.ntp.org provides me, because their stratum goes up to 16.
Does anyone have any advice on how often I should restart ntpd in order to re-map my *.pool.ntp.org domains to new ntp servers?
Here's my /etc/ntp.conf for reference:
| Code: | alpha root # cat /etc/ntp.conf
#restrict default noquery notrust nomodify
restrict default kod nomodify notrap nopeer noquery
restrict 127.0.0.1 nomodify
restrict 192.168.0.1 mask 255.255.255.0 nomodify nopeer notrap
fudge 127.127.1.0 stratum 3
server 0.pool.ntp.org
server 1.pool.ntp.org
server 2.pool.ntp.org
#server 127.127.1.1
driftfile /var/lib/ntp/ntp.drift
logfile /var/log/ntp.log |
Also, I'm curious why the polling success seems so varied between my various clients and servers:
| Code: | alpha root # ntpq -c pe
remote refid st t when poll reach delay offset jitter
==============================================================================
+a.pool.ntp.uq.e 130.102.152.7 2 u 43 64 377 217.321 31.321 5.463
*ntp.cis.rv.ua 129.69.1.153 2 u 38 64 377 177.224 18.425 9.038
+evenstar.silver 192.36.134.25 2 u 37 64 377 149.076 16.320 8.834
alpha root # ntpq -c rv |grep stratum
processor="i686", system="Linux/2.6.11-gentoo-r9", leap=00, stratum=3,
beta ~ # ntpq -c pe
remote refid st t when poll reach delay offset jitter
==============================================================================
alpha 194.44.92.10 3 u 807 1024 1 1.376 -2462.9 0.001
omega ~ # ntpq -c pe
remote refid st t when poll reach delay offset jitter
==============================================================================
alpha 194.44.92.10 3 u 1002 1024 1 0.349 -19997. 0.001
gamma ~ # ntpq -c pe
remote refid st t when poll reach delay offset jitter
==============================================================================
*alpha 194.44.92.10 3 u 1 64 377 0.277 -12.140 7.409
|
Why is gamma reaching my ntp server, alpha, so much more than all the other machines? They are all configured exactly the same.
Finally all my clients reached stratum 4 while my router is stratum 3, it seems strange that there was so much difference between the number of ntp requests that "reach"ed and how quickly that number increased. weird.
Thanks for any advice!
_________________
Free Question Answerer |
|
| Back to top |
|
 |
wynn
Advocate
Joined: 01 Apr 2005
Posts: 2421
Location: UK
|
| Posted: Fri Aug 04, 2006 9:39 am Post subject: |
|
|
| Quote: | | Does anyone have any advice on how often I should restart ntpd in order to re-map my *.pool.ntp.org domains to new ntp servers? |
Daily should be a good first step, to be halved if you still have problems.
Although pool.ntp.org is the best idea for spreading the load and getting working servers, you could also consider putting in a fixed list of, say, 10 or 15 servers from the public server list at http://ntp.isc.org/bin/view/Servers/WebHome
| Quote: | | I'm curious why the polling success seems so varied between my various clients and servers |
I'm not sure if I've understood you but beta and omega have set the polling period up to the maximum, 1024, while gamma is still at the first stage, 64. As far as I know, this is to do with the ease with which beta and omega have in getting to alpha over the network and the difficulty that gamma has had.
| Quote: | | number of ntp requests that "reach"ed and how quickly that number increased |
Again, I'm not sure whether I've understood you correctly but the reach number is eight bits with the low order bit being set to 1 when a poll succeeds in reaching the server or set to 0 if it fails.
_________________
The avatar is jorma, a "duck" from "Elephants Dream": the film and all the production materials have been made available under a Creative Commons Attribution 2.5 License, see orange.blender.org for details. |
|
| Back to top |
|
|
ben_dash
Apprentice
Joined: 12 Aug 2003
Posts: 239
|
| Posted: Fri Aug 04, 2006 12:24 pm Post subject: |
|
|
Hmmm, all the machines are on the same LAN with a gigabit switch. I wonder why any of them would have any difference in their abilities to talk to alpha.
It looks like I totally misunderstood what "poll" and "reach" meant... time for a little NTP research for me 
It seems like the issue of the pool.ntp.org servers gradually losing stratum status would be a common one. If the solution is to cycle the ntpd daemon twice a day do you already have a script to do this, I don't want to reinvent the wheel...
Thanks!!!
_________________
Free Question Answerer |
|
| Back to top |
|
|
wynn
Advocate
Joined: 01 Apr 2005
Posts: 2421
Location: UK
|
| Posted: Fri Aug 04, 2006 1:02 pm Post subject: |
|
|
| Quote: | | If the solution is to cycle the ntpd daemon twice a day do you already have a script to do this |
Funny, I was going to answer this question in my first post until I reread the question you had actually asked
Couldn't you put
| Code: | | /etc/init.d/ntpd restart |
in crontab. I'd look up how to do it but I'm sure you know better than me.
_________________
The avatar is jorma, a "duck" from "Elephants Dream": the film and all the production materials have been made available under a Creative Commons Attribution 2.5 License, see orange.blender.org for details. |
|
| Back to top |
|
|
tgh
Apprentice
Joined: 05 Oct 2005
Posts: 222
|
| Posted: Mon Aug 14, 2006 6:22 am Post subject: |
|
|
Looks like you're only talking to (3) NTP servers in your config. You might want to up that to (5) NTP servers from the pool. You can have more, but NTP will only talk to a maximum of 10. You may need to add entries from another country code (if you're in the US, use the Canada pool entries such as 0.ca.pool.ntp.org to gain additionals). For instance:
server 0.us.pool.ntp.org
server 1.us.pool.ntp.org
server 2.us.pool.ntp.org
server 0.ca.pool.ntp.org
#server 1.ca.pool.ntp.org
#server 2.ca.pool.ntp.org
I would suggest not restarting ntp more often then once per week. As long as the NTP daemon has been running for a while and has been tracking a few stable time sources, it should be fine for a day or two until the weekly restart. Of course, YMMV.
If you're servicing multiple clients, you may want to look into adding a single Statum Two entry at the top (to make it easy to see) of your server list. That would be in addition to the 4 or 5 servers from the pool. Whether those stratum two servers listed at the ntp.org website will stay around for a while is a bit of guesswork. |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Networking & Security
