NTP server remain in Soliciting pool server
-
And are you getting answers? Your reach is 0 (zero) you can not sync time if you do not get answers...
You seem to be sending out the queries - but if you do not get an answer, then no you can not sync... That has ZERO to do with pfsense.
What is this
Thu, Feb 7 2036
ntpq> readvar associd=0 status=0618 leap_none, sync_ntp, 1 event, no_sys_peer, version="ntpd 4.2.8p11@1.3728-o Fri Mar 16 18:58:06 UTC 2018 (1)", processor="amd64", system="FreeBSD/11.1-RELEASE-p10", leap=00, stratum=2, precision=-21, rootdelay=0.230, rootdisp=14.828, refid=192.168.3.32, reftime=deeb62ad.1d5fcc1b Sat, Jul 7 2018 11:02:53.114, clock=deeb6310.a0d7f0a0 Sat, Jul 7 2018 11:04:32.628, peer=23915, tc=9, mintc=3, offset=0.120466, frequency=3.615, sys_jitter=0.127323, clk_jitter=0.157, clk_wander=0.017If your sending ref of 2036 - maybe the ntp server your asking just ignores you and does not answer. See my reftime
Looks like your normal queries are using a date of feb 7, 2036... Adjust the date and time on your pfsense box to be atleast close! ;)
-
reftime=00000000.00000000 Thu, Feb 7 2036 7:28:16.000 is not the problem at initial sync...
i've configured another internal ntp server that works and pfsense box ntpd using that server works...
-
So your saying ntp works talking to your ntp server locally, but does not get answers remotely? But it not getting answers is related to what? How is an ntp server not answering something you clearly sent out pfsense issue? If it was something wrong with the query - then your local ntp server wouldn't answer it either.
Your local ntp server taking a different path to sync to outside, is it not syncing to outside?
-
my local ntp server reach internet through pfsense box and works fine, sync and serve.
pfsense box using my local ntp server works fine:
ntpq> readvar associd=0 status=0618 leap_none, sync_ntp, 1 event, no_sys_peer, version="ntpd 4.2.8p11@1.3728-o Wed Feb 28 13:29:50 UTC 2018 (1)", processor="amd64", system="FreeBSD/11.1-RELEASE-p10", leap=00, stratum=3, precision=-22, rootdelay=36.095, rootdisp=47.551, refid=10.10.11.2, reftime=deec69b4.f16fb644 Sun, Jul 8 2018 12:45:08.943, clock=deec69bf.29ca9dae Sun, Jul 8 2018 12:45:19.163, peer=44435, tc=6, mintc=3, offset=2.311372, frequency=3.350, sys_jitter=0.000000, clk_jitter=0.134, clk_wander=0.052ntpq> peers remote refid st t when poll reach delay offset jitter ============================================================================== 10.10.11.2 .POOL. 16 p - 64 0 0.000 0.000 0.000 *wpad.*****.** 193.204.114.105 2 u 35 64 377 0.145 2.311 0.049ntpq> as ind assid status conf reach auth condition last_event cnt =========================================================== 1 44434 8811 yes none none reject mobilize 1 2 44435 161a no yes none sys.peer sys_peer 1 -
Why would you have it marked pool when clearly its not?
So my question is does this ntp server take the same "path" to the internet that pfsense would take on its own? Or are you routing this traffic different pfsense local traffic would take - ie the default route/gateway?
You stated
i have only an internet interface that isn't WAN (now is disabled for migration reason
-
for pool configuration is just an oversight.
ntp server take the same path of pfsense (not WAN)
-
Well that makes zero sense - is the ntp server using the same servers you were trying to sync pfsense with?
You sure the ref time thing is not an issue. I don't recall every seeing such a ref time.. it is one of the rollover dates.. Now that your ntp is syncing and your time is good. What if you change to from your local server to external on pfsense?
So your saying that when you ntp server that is internal talks to the outside to sync its time its using the same bath and "the same source ip" as when pfsense itself goes out to the internet to talk to ntp. That 80.19.a.b address, your sure its not say 80.19.a.c ? or something like that?
Makes not sense that ntp server would answer your client, but not the client on pfsense?
What version of pfsense are you using? 2.4.3p1?
-
yes the same servers.
mtr from my local server:
mtr from my pfsense box
it's the same...
-
Not what I asked exactly... I asked if the source IP is the same..
While you might hit the same gateway that .233, maybe your coming from .A when natting your clients behind pfsense but pfsense is using .B as its IP. For example if you have a VIP setup when you have multiple IPs from the ISP..
You only have the 1 public IP, and no vips setup on pfsense where it could be using a different source IP.
Just makes no sense that your not getting any replies to your queries when pfsense does them, but do get replies when client behind pfsense does the query.
-
80.19.236.233 is my public router
80.19.236.234 is my pfsense public interface directly connected to 80.19.236.233
ntp internal server is 10.10.11.2 connected to pfsense 10.10.11.1
-
Well then why does server X not respond to a ntp query when it comes from pfsense, but does respond when it comes from client behind pfsense? That is the million dollar question.
Since the source IP is the same. The clients are the same ntp version it looks like.
ntpd 4.2.8p11@1.3728-o
So I would suggest you set pfsense to use both your local ntp, and public ntp. It will now query both of them. Are you not seeing answers from the public ones? See my output below, you see pfsense syncs to my local stratum 1 ntp server I run. But it also asks pool as backup to if my local goes down (pi with gps as my ntp server). you will notice that they all show reach of 377.. Means talking to them and getting answers..
-
@johnpoz said in NTP server remain in Soliciting pool server:
Well then why does server X not respond to a ntp query when it comes from pfsense, but does respond when it comes from client behind pfsense? That is the million dollar question.
yes, pfsense box not respond to a ntp query but a client of pfsense box works.
I tried to configured local and public ntp server but public doesn't work as usual...
-
Makes ZERO sense... Really going to need a sniff on your outbound interface so we can see the ntp queries from your client behind pfsense and the queries from pfsense itself so we can try and figure out why the ones from pfsense do not get any responses.
I would suggest you sniff on your outbound interface for 123 udp with host of one of the public servers your client and pfsense are doing queries for.
Then download this and open in wireshark.. Maybe we can spot the difference in the queries that would explain why the server doesn't respond, or its possible maybe it is and its not getting back to you. I run a server in the pool.. So its open to the internet IPv4 and IPv6.. Let me send you the IP in a private chat - and set pfsense to use that server. And I can sniff on my wan to see if getting your query and if I send a response.
I will setup a sniff right now for your 80.x.x.x IP... Give me 2 seconds.
-
Ok I have sniff running for your IP on my wan udp 123... Set pfsense to use this IP and see if you get any responses or not.. Let me know after its been running a few minutes and I will stop the sniff and take a look
BTW: please do not post my full public IP its 64.53.x.x
-
Ok, some ntp packet sent to your ntp but without any response.
-
No nothing captured.. I never got those packets.
-
Not from that IP you say is your public IP. You sure pfsense is sending them out the right hole??
-
Since it seems now via chat you do have VIPs... How about you post your outbound nats so we can figure out what IP pfsense would use vs your clients when talking outbound.
-
-
There are no hybrid rules above that? Or below that? From those your clients and pfsense itself would be using whatever the IP of ADSL20PUB is... Do you have other connections? What interfaces do you have your ntp set to use in the ntp setup page?
Do you have any port forwards setup on your client interfaces for ntp? (udp 123)
