[opensuse] TW ntpd delays shutdown
I've seen this on multiple installations and just ignored the delay before, but it's seriously irritating that reboot can take more than 10 or so seconds to accomplish. Typically ntpd shows a 1.5m timeout, and often takes more than a minute to timeout and reboot to proceed. The only red I see in 'systemctl status ntpd' is restrict: ignoring line 52, mask '::' unusable. Lines 52-54 of /etc/ntp.conf: # Local users may interrogate the ntp server more closely. restrict 127.0.0.1 restrict ::1 Tail of /var/log/ntp: http://fm.no-ip.com/Tmp/SUSE/Factory/ntptail-TWdelay.txt What can be done to end this delay? -- "Wisdom is supreme; therefore get wisdom. Whatever else you get, get wisdom." Proverbs 4:7 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Felix Miata wrote:
I've seen this on multiple installations and just ignored the delay before, but it's seriously irritating that reboot can take more than 10 or so seconds to accomplish. Typically ntpd shows a 1.5m timeout, and often takes more than a minute to timeout and reboot to proceed.
The only red I see in 'systemctl status ntpd' is
restrict: ignoring line 52, mask '::' unusable.
This is because you have (presumably) disabled ipv6. I'm sure it can be ignored.
Tail of /var/log/ntp:
http://fm.no-ip.com/Tmp/SUSE/Factory/ntptail-TWdelay.txt
What can be done to end this delay?
I can't reproduce such a delay at shutdown, but with Leap15, not TW. Also, I don't use any external time sources. If it were me, I would try enabling ipv6, just to see if it might have any effect. -- Per Jessen, Zürich (18.8°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-05-28 08:22, Per Jessen wrote:
Felix Miata wrote:
I've seen this on multiple installations and just ignored the delay before, but it's seriously irritating that reboot can take more than 10 or so seconds to accomplish. Typically ntpd shows a 1.5m timeout, and often takes more than a minute to timeout and reboot to proceed.
The only red I see in 'systemctl status ntpd' is
restrict: ignoring line 52, mask '::' unusable.
This is because you have (presumably) disabled ipv6. I'm sure it can be ignored.
Tail of /var/log/ntp:
http://fm.no-ip.com/Tmp/SUSE/Factory/ntptail-TWdelay.txt
What can be done to end this delay?
I can't reproduce such a delay at shutdown, but with Leap15, not TW. Also, I don't use any external time sources.
If it were me, I would try enabling ipv6, just to see if it might have any effect.
Only if the router and ISP provide IPv6. On the other hand, why has ntpd got to connect anywhere when told to die? Just die and shut up, fast. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)
On 2018-05-28 12:30, Peter Suetterlin wrote:
Carlos E. R. wrote:
On the other hand, why has ntpd got to connect anywhere when told to die? Just die and shut up, fast.
Issues syncing to the HW clock?
Writing the HW clock takes a second at most (because that clock has one second granularity). Maybe ntp thinks the system clock is not in sync and tries to get a sync before writing the time to the hwclock. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)
Carlos E. R. wrote:
On 2018-05-28 12:30, Peter Suetterlin wrote:
Issues syncing to the HW clock?
Writing the HW clock takes a second at most (because that clock has one second granularity).
If it works. If everything works as expected, ntpd would not hang. But it does. So I was thinking of some timeout accessing the HW clock for whatever reason.... -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-05-28 14:43, Peter Suetterlin wrote:
Carlos E. R. wrote:
On 2018-05-28 12:30, Peter Suetterlin wrote:
Issues syncing to the HW clock?
Writing the HW clock takes a second at most (because that clock has one second granularity).
If it works. If everything works as expected, ntpd would not hang. But it does. So I was thinking of some timeout accessing the HW clock for whatever reason....
Can't be. Nothing I know of can delay that operation for more than a second. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)
Carlos E. R. wrote:
On 2018-05-28 08:22, Per Jessen wrote:
Felix Miata wrote:
I've seen this on multiple installations and just ignored the delay before, but it's seriously irritating that reboot can take more than 10 or so seconds to accomplish. Typically ntpd shows a 1.5m timeout, and often takes more than a minute to timeout and reboot to proceed.
The only red I see in 'systemctl status ntpd' is
restrict: ignoring line 52, mask '::' unusable.
This is because you have (presumably) disabled ipv6. I'm sure it can be ignored.
Tail of /var/log/ntp:
http://fm.no-ip.com/Tmp/SUSE/Factory/ntptail-TWdelay.txt
What can be done to end this delay?
I can't reproduce such a delay at shutdown, but with Leap15, not TW. Also, I don't use any external time sources.
If it were me, I would try enabling ipv6, just to see if it might have any effect.
Only if the router and ISP provide IPv6.
No, that's irrelevant. It's only an issue if the configuration is faulty.
On the other hand, why has ntpd got to connect anywhere when told to die? Just die and shut up, fast.
It's impossible to tell - insufficient info at the moment. The log contains nothing useful. -- Per Jessen, Zürich (23.1°C) http://www.hostsuisse.com/ - virtual servers, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-05-28 12:35, Per Jessen wrote:
Carlos E. R. wrote:
On 2018-05-28 08:22, Per Jessen wrote:
Felix Miata wrote:
I've seen this on multiple installations and just ignored the delay before, but it's seriously irritating that reboot can take more than 10 or so seconds to accomplish. Typically ntpd shows a 1.5m timeout, and often takes more than a minute to timeout and reboot to proceed.
The only red I see in 'systemctl status ntpd' is
restrict: ignoring line 52, mask '::' unusable.
This is because you have (presumably) disabled ipv6. I'm sure it can be ignored.
Tail of /var/log/ntp:
http://fm.no-ip.com/Tmp/SUSE/Factory/ntptail-TWdelay.txt
What can be done to end this delay?
I can't reproduce such a delay at shutdown, but with Leap15, not TW. Also, I don't use any external time sources.
If it were me, I would try enabling ipv6, just to see if it might have any effect.
Only if the router and ISP provide IPv6.
No, that's irrelevant. It's only an issue if the configuration is faulty.
Define faulty :-) Some of my machines have a local fixed IPv6 address, created for testing IPv6 inside the local network. But no IPv6 addresses given by the router, and no IPv6 to outside. But I do not know if my "halt" sequence is slow due to ntp or not, I have not investigated. I never reboot, I always hibernate. I do see halt as slow, but I blame that my machine has been up for many days, and that I reboot because I updated the kernel or something, so that the system is not stable at that point. Sometimes halt fails. Ok, poweroff, not halt.
On the other hand, why has ntpd got to connect anywhere when told to die? Just die and shut up, fast.
It's impossible to tell - insufficient info at the moment. The log contains nothing useful.
My ntp log is too short, too. Look: 25 May 23:24:40 ntpd[3029]: 0.0.0.0 c618 08 no_sys_peer 26 May 23:33:15 ntpd[3029]: 0.0.0.0 0613 03 spike_detect +0.510839 s 26 May 23:37:57 ntpd[3029]: 0.0.0.0 061c 0c clock_step +0.512390 s 26 May 23:37:57 ntpd[3029]: 0.0.0.0 0615 05 clock_sync 26 May 23:37:58 ntpd[3029]: 0.0.0.0 c618 08 no_sys_peer 27 May 10:14:17 ntpd[3029]: 0.0.0.0 0613 03 spike_detect -0.331092 s 27 May 10:14:28 ntpd[3029]: 0.0.0.0 061c 0c clock_step -0.331364 s 27 May 10:14:27 ntpd[3029]: 0.0.0.0 0615 05 clock_sync 27 May 10:14:28 ntpd[3029]: 0.0.0.0 c618 08 no_sys_peer Now, consider that between yesterday and today I hibernated the machine. Surely when I restored the machine ntp had to go nuts! Yet, there is nothing in the log. A spike of 0.3 seconds, that's all. Surely my settings need a change so that it logs more details. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)
Carlos E. R. wrote:
On 2018-05-28 12:35, Per Jessen wrote:
Carlos E. R. wrote:
On 2018-05-28 08:22, Per Jessen wrote:
Felix Miata wrote:
I've seen this on multiple installations and just ignored the delay before, but it's seriously irritating that reboot can take more than 10 or so seconds to accomplish. Typically ntpd shows a 1.5m timeout, and often takes more than a minute to timeout and reboot to proceed.
The only red I see in 'systemctl status ntpd' is
restrict: ignoring line 52, mask '::' unusable.
This is because you have (presumably) disabled ipv6. I'm sure it can be ignored.
Tail of /var/log/ntp:
http://fm.no-ip.com/Tmp/SUSE/Factory/ntptail-TWdelay.txt
What can be done to end this delay?
I can't reproduce such a delay at shutdown, but with Leap15, not TW. Also, I don't use any external time sources.
If it were me, I would try enabling ipv6, just to see if it might have any effect.
Only if the router and ISP provide IPv6.
No, that's irrelevant. It's only an issue if the configuration is faulty.
Define faulty :-)
Your setup. :-) A non-working IPv6 setup. If the router pretends it has IPv6, but it doesn't, that'll screw things up.
On the other hand, why has ntpd got to connect anywhere when told to die? Just die and shut up, fast.
It's impossible to tell - insufficient info at the moment. The log contains nothing useful.
My ntp log is too short, too. Look:
25 May 23:24:40 ntpd[3029]: 0.0.0.0 c618 08 no_sys_peer 26 May 23:33:15 ntpd[3029]: 0.0.0.0 0613 03 spike_detect +0.510839 s 26 May 23:37:57 ntpd[3029]: 0.0.0.0 061c 0c clock_step +0.512390 s 26 May 23:37:57 ntpd[3029]: 0.0.0.0 0615 05 clock_sync 26 May 23:37:58 ntpd[3029]: 0.0.0.0 c618 08 no_sys_peer 27 May 10:14:17 ntpd[3029]: 0.0.0.0 0613 03 spike_detect -0.331092 s 27 May 10:14:28 ntpd[3029]: 0.0.0.0 061c 0c clock_step -0.331364 s 27 May 10:14:27 ntpd[3029]: 0.0.0.0 0615 05 clock_sync 27 May 10:14:28 ntpd[3029]: 0.0.0.0 c618 08 no_sys_peer
You have different log options set too, it's much more verbose. -- Per Jessen, Zürich (23.9°C) http://www.cloudsuisse.com/ - your owncloud, hosted in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2018-05-28 13:09, Per Jessen wrote:
Carlos E. R. wrote:
On 2018-05-28 12:35, Per Jessen wrote:
Carlos E. R. wrote:
Only if the router and ISP provide IPv6.
No, that's irrelevant. It's only an issue if the configuration is faulty.
Define faulty :-)
Your setup. :-)
A non-working IPv6 setup. If the router pretends it has IPv6, but it doesn't, that'll screw things up.
AFAIK my router does not pretend to provide IPv6.
On the other hand, why has ntpd got to connect anywhere when told to die? Just die and shut up, fast.
It's impossible to tell - insufficient info at the moment. The log contains nothing useful.
My ntp log is too short, too. Look:
25 May 23:24:40 ntpd[3029]: 0.0.0.0 c618 08 no_sys_peer 26 May 23:33:15 ntpd[3029]: 0.0.0.0 0613 03 spike_detect +0.510839 s 26 May 23:37:57 ntpd[3029]: 0.0.0.0 061c 0c clock_step +0.512390 s 26 May 23:37:57 ntpd[3029]: 0.0.0.0 0615 05 clock_sync 26 May 23:37:58 ntpd[3029]: 0.0.0.0 c618 08 no_sys_peer 27 May 10:14:17 ntpd[3029]: 0.0.0.0 0613 03 spike_detect -0.331092 s 27 May 10:14:28 ntpd[3029]: 0.0.0.0 061c 0c clock_step -0.331364 s 27 May 10:14:27 ntpd[3029]: 0.0.0.0 0615 05 clock_sync 27 May 10:14:28 ntpd[3029]: 0.0.0.0 c618 08 no_sys_peer
You have different log options set too, it's much more verbose.
logfile /var/log/ntp # alternate log file # logconfig =syncstatus + sysevents logconfig =all It should log "all". It does not. statsdir /var/log/ntpstat/ # directory for statistics files filegen peerstats file peerstats type day enable filegen loopstats file loopstats type day enable filegen clockstats file clockstats type day enable The directory is empty. -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" at Telcontar)
Carlos E. R. wrote:
On 2018-05-28 13:09, Per Jessen wrote:
Carlos E. R. wrote:
On 2018-05-28 12:35, Per Jessen wrote:
Carlos E. R. wrote:
Only if the router and ISP provide IPv6.
No, that's irrelevant. It's only an issue if the configuration is faulty.
Define faulty :-)
Your setup. :-)
A non-working IPv6 setup. If the router pretends it has IPv6, but it doesn't, that'll screw things up.
AFAIK my router does not pretend to provide IPv6.
ISTR that was the exact problem last time we investigated it. Your router thought you had IPv6, but your provider doesnt. Something like that. We're straying from the topic - but unless your router sends out routing announcements, or you have an active dhcpv6, your network interfaces will never allocate global addresses, only link-local. No problem.
logfile /var/log/ntp # alternate log file # logconfig =syncstatus + sysevents logconfig =all
It should log "all". It does not.
Anyway, it doesn't help Felix with his shutdown problem. I've upgraded ntp on my TW installation, still can't reproduce. Felix, maybe post your complete config? -- Per Jessen, Zürich (25.4°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (4)
-
Carlos E. R. -
Felix Miata -
Per Jessen -
Peter Suetterlin