NTP Sync Strange Issue

[scracha]

Morning,

We started getting alerts outside of maintenance window for a port we schedule to switch off at night on a solar site so I investigated.

Date/Time on the netonix was 6 minutes ahead of NTP server. I popped in an invalid NTP server IP address then put correct one in again and it corrected itself. Uptime on netonix 270 days so I'm wondering if the NTP client daemon on the netonix isn't resynching periodically?

I'm seeing this on all our netonixes (some are ahead or beind real time by up to 10 minutes). Haven't corrected the rest yet in the event you need more troubleshooting info. Firmware 1.5.3 (obviously due to the uptime it isn't the latest but I woudln't imagine you've made NTP changes)

thanks

[Stephen]

It should be, the ntpclient in our switch is a package commonly used in ubuntu server's and a few other distro's. Do you have any log's from that switch?

[jpaine619]

It should be, the ntpclient in our switch is a package commonly used in ubuntu server's and a few other distro's. Do you have any log's from that switch?

ahh.. 3 x in a single line.. I have to say something.....

apostrophes are used to denote contractions or possessives. Not pluralities. servers, distros, logs.

[scracha]

It should be, the ntpclient in our switch is a package commonly used in ubuntu server's and a few other distro's. Do you have any log's from that switch?

Hi, there's nothing in the log files pertaining to ntpclient. I've just went to another netonix switch (we have about 60 of them), disabled NTP and then re-enabled it and it corrected the time.


Jan 1 00:00:11 netonix: 1.5.3 on WS-12-250-DC
Jan 1 00:00:15 system: Setting MAC address from flash configuration: EC:13:B2:82:53:F0
Jan 1 00:00:18 admin: adding lan (eth0) to firewall zone lan
Jan 1 13:00:35 admin: removing lan (eth0) from firewall zone lan
Jan 1 13:00:39 admin: adding lan (eth0) to firewall zone lan
Jan 1 13:00:46 admin: adding lan (eth0) to firewall zone lan
Jan 1 13:00:48 Port: link state changed to 'up' (1G) on port 13
Jan 1 13:00:48 system: starting ntpclient
Jan 1 13:00:52 dropbear[1372]: Running in background
Jan 1 13:00:54 switch[1399]: temp sensor version 3
Jan 1 13:00:54 switch[1400]: Detected warm boot
Jan 1 13:00:56 system: starting ntpclient
Apr 17 11:50:44 system: time set by NTP server
May 15 03:07:39 Port: link state changed to 'down' on port 2
May 15 03:08:25 Port: link state changed to 'up' (1G) on port 2
Jul 6 10:27:29 UI: Configuration changed by admin (::ffff:202.1XX.X.X)
Jul 6 10:27:29 UI: NTP_Enable: changed from 'Enabled' to 'Disabled'
Jul 6 10:27:41 UI: Configuration changed by admin (::ffff:202.1XX.X.X)
Jul 6 10:27:41 UI: NTP_Enable: changed from 'Disabled' to 'Enabled'
Jul 6 10:27:45 system: starting ntpclient
Jul 6 10:29:20 UI: Configuration auto backup successful




And linux log file


Jan 1 00:00:10 sysinit: killall: udhcpc: no process killed
Jan 1 00:00:13 kernel: vtss_core: module license '(c) Vitesse Semiconductor Inc.' taints kernel.
Jan 1 00:00:13 kernel: switch: 'Luton26' board detected
Jan 1 00:00:14 kernel: vtss_port: Loaded port module on board Luton26, type 5
Jan 1 00:00:15 kernel: nf_conntrack version 0.5.0 (2048 buckets, 8192 max)
Jan 1 00:00:15 kernel: i2c /dev entries driver
Jan 1 00:00:15 kernel: i2c_vcoreiii i2c_vcoreiii: i2c bus driver on IRQ 19
Jan 1 00:00:17 sysinit: Loading defaults
Jan 1 00:00:17 sysinit: Loading synflood protection
Jan 1 00:00:17 sysinit: Adding custom chains
Jan 1 00:00:17 sysinit: Loading zones
Jan 1 00:00:18 sysinit: Loading forwarding
Jan 1 00:00:18 sysinit: Loading redirects
Jan 1 00:00:18 sysinit: Loading rules
Jan 1 00:00:18 sysinit: sh: bad number
Jan 1 00:00:18 sysinit: Loading includes
Jan 1 00:00:26 kernel: eth0: no IPv6 routers present
Jan 1 13:00:32 syslogd exiting
Jan 1 13:00:33 syslogd started: BusyBox v1.19.4
Jan 1 13:00:33 syslogd exiting
Jan 1 13:00:34 syslogd started: BusyBox v1.19.4
Jan 1 13:00:47 sysinit: killall: telnetd: no process killed
Jan 1 13:00:47 sysinit: killall: ntpclient: no process killed
Jan 1 13:00:48 sysinit: killall: ntpclient: no process killed
Jan 1 13:00:49 sysinit: 1970-01-01 13:00:49: (network.c.272) warning: please use server.use-ipv6 only for hostnames, not without server.bind / empty address; your config will break if the kernel default for IPV6_V6ONLY changes
Jan 1 13:00:49 sysinit: 1970-01-01 13:00:49: (network.c.272) warning: please use server.use-ipv6 only for hostnames, not without server.bind / empty address; your config will break if the kernel default for IPV6_V6ONLY changes
Jan 1 13:00:49 sysinit: 1970-01-01 13:00:49: (log.c.216) server started
Jul 6 10:27:31 syslogd exiting
Jul 6 10:27:32 syslogd started: BusyBox v1.19.4
Jul 6 10:27:43 syslogd exiting
Jul 6 10:27:44 syslogd started: BusyBox v1.19.4

[Stephen]

I'm looking into it now. It's likely just an update to a configuration file for that software to make it poll the server's to stay in sync. I'll find out soon.

[sakita]

Another brand of switch we use has a setting, in minutes, for how often to poll the NTP server. The value ranges from 1 to 1440. For example, we usually set it to 1440 for a daily time update. So, on bootup the switch gets the time and then at the same time every day it gets it again.

I just checked a few of our Netonix switches (running from 50 to 208 days) and their clocks are off by 20 seconds to 1 minute 30 seconds.

[Stephen]

Fix for this has been released, now available in 1.5.6rc1