home.social

#chrony — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #chrony, aggregated by home.social.

  1. From Milliseconds to 26 Nanoseconds: How a $20 eBay SFP Module Beat My Entire NTP Setup

    austinsnerdythings.com/2026/04

    I think we can get Subspace 3.0 up and running cheaper now @joew

  2. From Milliseconds to 26 Nanoseconds: How a $20 eBay SFP Module Beat My Entire NTP Setup

    austinsnerdythings.com/2026/04

    I think we can get Subspace 3.0 up and running cheaper now @joew

    #NTP #GPS #Chrony

  3. From Milliseconds to 26 Nanoseconds: How a $20 eBay SFP Module Beat My Entire NTP Setup

    austinsnerdythings.com/2026/04

    I think we can get Subspace 3.0 up and running cheaper now @joew

    #NTP #GPS #Chrony

  4. From Milliseconds to 26 Nanoseconds: How a $20 eBay SFP Module Beat My Entire NTP Setup

    austinsnerdythings.com/2026/04

    I think we can get Subspace 3.0 up and running cheaper now @joew

    #NTP #GPS #Chrony

  5. From Milliseconds to 26 Nanoseconds: How a $20 eBay SFP Module Beat My Entire NTP Setup

    austinsnerdythings.com/2026/04

    I think we can get Subspace 3.0 up and running cheaper now @joew

    #NTP #GPS #Chrony

  6. Stratum 1 NTP-сервер на Raspberry Pi в 2026 году: DCF77 + GPS/PPS, chrony, libgpiod v2

    Stratum 1 NTP-сервер на Raspberry Pi в 2025 году: DCF77 + GPS/PPS, chrony, libgpiod v2, когда все туториалы по «точному времени на Raspberry Pi» сломаны на современных системах. Разбираю почему — и показываю как сделать правильно.

    habr.com/ru/articles/1018554/

    #raspberrypi #C++ #linux #ntp #chrony #gps #dcf77

  7. Stratum 1 NTP-сервер на Raspberry Pi в 2026 году: DCF77 + GPS/PPS, chrony, libgpiod v2

    Stratum 1 NTP-сервер на Raspberry Pi в 2025 году: DCF77 + GPS/PPS, chrony, libgpiod v2, когда все туториалы по «точному времени на Raspberry Pi» сломаны на современных системах. Разбираю почему — и показываю как сделать правильно.

    habr.com/ru/articles/1018554/

    #raspberrypi #C++ #linux #ntp #chrony #gps #dcf77

  8. Stratum 1 NTP-сервер на Raspberry Pi в 2026 году: DCF77 + GPS/PPS, chrony, libgpiod v2

    Stratum 1 NTP-сервер на Raspberry Pi в 2025 году: DCF77 + GPS/PPS, chrony, libgpiod v2, когда все туториалы по «точному времени на Raspberry Pi» сломаны на современных системах. Разбираю почему — и показываю как сделать правильно.

    habr.com/ru/articles/1018554/

    #raspberrypi #C++ #linux #ntp #chrony #gps #dcf77

  9. Stratum 1 NTP-сервер на Raspberry Pi в 2026 году: DCF77 + GPS/PPS, chrony, libgpiod v2

    Stratum 1 NTP-сервер на Raspberry Pi в 2025 году: DCF77 + GPS/PPS, chrony, libgpiod v2, когда все туториалы по «точному времени на Raspberry Pi» сломаны на современных системах. Разбираю почему — и показываю как сделать правильно.

    habr.com/ru/articles/1018554/

    #raspberrypi #C++ #linux #ntp #chrony #gps #dcf77

  10. Secure time synchronization with NTS

    To set the correct time on computers, usually NTP (Network Time Protocol) is used. However NTP is an insecure time synchronization protocol and allows a man-in-the-middle to shift time on an NTP client. To tackle these issues, NTS (Network Time Security) was created which aims to solve these problems. In this article I show how you can configure ntpd-rs and Chrony to use NTS servers. systemd's timesyncd does not support yet NTS, but it's currently in development so hopefully this will be […]

    blog.frehi.be/2026/03/29/secur

  11. Secure time synchronization with NTS

    To set the correct time on computers, usually NTP (Network Time Protocol) is used. However NTP is an insecure time synchronization protocol and allows a man-in-the-middle to shift time on an NTP client. To tackle these issues, NTS (Network Time Security) was created which aims to solve these problems. In this article I show how you can configure ntpd-rs and Chrony to use NTS servers. systemd's timesyncd does not support yet NTS, but it's currently in development so hopefully this will be […]

    blog.frehi.be/2026/03/29/secur

  12. Secure time synchronization with NTS

    To set the correct time on computers, usually NTP (Network Time Protocol) is used. However NTP is an insecure time synchronization protocol and allows a man-in-the-middle to shift time on an NTP client. To tackle these issues, NTS (Network Time Security) was created which aims to solve these problems. In this article I show how you can configure ntpd-rs and Chrony to use NTS servers. systemd's timesyncd does not support yet NTS, but it's currently in development so hopefully this will be […]

    blog.frehi.be/2026/03/29/secur

  13. Secure time synchronization with NTS

    To set the correct time on computers, usually NTP (Network Time Protocol) is used. However NTP is an insecure time synchronization protocol and allows a man-in-the-middle to shift time on an NTP client. To tackle these issues, NTS (Network Time Security) was created which aims to solve these problems. In this article I show how you can configure ntpd-rs and Chrony to use NTS servers. systemd's timesyncd does not support yet NTS, but it's currently in development so hopefully this will be […]

    blog.frehi.be/2026/03/29/secur

  14. Secure time synchronization with NTS

    To set the correct time on computers, usually NTP (Network Time Protocol) is used. However NTP is an insecure time synchronization protocol and allows a man-in-the-middle to shift time on an NTP client. To tackle these issues, NTS (Network Time Security) was created which aims to solve these problems. In this article I show how you can configure ntpd-rs and Chrony to use NTS servers. systemd's timesyncd does not support yet NTS, but it's currently in development so hopefully this will be […]

    blog.frehi.be/2026/03/29/secur

  15. Needed an #NTP server on a Raspberry Pi. To keep it light, I chose #OpenNTPD.

    Not a great idea.

    If your clock is off by more than 180 seconds, it won't adjust it; logs say

    adjtime failed: Invalid argument

    Now, according to the man page, it will jump to adjust larger offsets on startup, but apparently #Debian has patched this to require adding `-s` to /etc/default/openntpd.

    They also don't support `constraint from` config stanzas used to get a trusted date via HTTPS.

    Trying #chrony now.

  16. Needed an #NTP server on a Raspberry Pi. To keep it light, I chose #OpenNTPD.

    Not a great idea.

    If your clock is off by more than 180 seconds, it won't adjust it; logs say

    adjtime failed: Invalid argument

    Now, according to the man page, it will jump to adjust larger offsets on startup, but apparently #Debian has patched this to require adding `-s` to /etc/default/openntpd.

    They also don't support `constraint from` config stanzas used to get a trusted date via HTTPS.

    Trying #chrony now.

  17. Needed an #NTP server on a Raspberry Pi. To keep it light, I chose #OpenNTPD.

    Not a great idea.

    If your clock is off by more than 180 seconds, it won't adjust it; logs say

    adjtime failed: Invalid argument

    Now, according to the man page, it will jump to adjust larger offsets on startup, but apparently #Debian has patched this to require adding `-s` to /etc/default/openntpd.

    They also don't support `constraint from` config stanzas used to get a trusted date via HTTPS.

    Trying #chrony now.

  18. Needed an #NTP server on a Raspberry Pi. To keep it light, I chose #OpenNTPD.

    Not a great idea.

    If your clock is off by more than 180 seconds, it won't adjust it; logs say

    adjtime failed: Invalid argument

    Now, according to the man page, it will jump to adjust larger offsets on startup, but apparently #Debian has patched this to require adding `-s` to /etc/default/openntpd.

    They also don't support `constraint from` config stanzas used to get a trusted date via HTTPS.

    Trying #chrony now.

  19. Needed an #NTP server on a Raspberry Pi. To keep it light, I chose #OpenNTPD.

    Not a great idea.

    If your clock is off by more than 180 seconds, it won't adjust it; logs say

    adjtime failed: Invalid argument

    Now, according to the man page, it will jump to adjust larger offsets on startup, but apparently #Debian has patched this to require adding `-s` to /etc/default/openntpd.

    They also don't support `constraint from` config stanzas used to get a trusted date via HTTPS.

    Trying #chrony now.

  20. I configured #chrony shipped with #ProxmoxVE to operate as a server and set up #DNS records on my router to point ntp[1-4].aliyun.com to the local IP of the #Proxmox node with chrony.

    According to the internet, ntp[1-4].aliyun.com are 4 of 8 NTP servers used by #PhilipsHue bridge.

    Then I blocked internet access for the bridge.

    Now we wait and see if it works.

    #chrony #proxmoxve #dns #proxmox #philipshue #homelad #selfhost #selfhosted #selfhosting #offline

  21. I configured #chrony shipped with #ProxmoxVE to operate as a server and set up #DNS records on my router to point ntp[1-4].aliyun.com to the local IP of the #Proxmox node with chrony.

    According to the internet, ntp[1-4].aliyun.com are 4 of 8 NTP servers used by #PhilipsHue bridge.

    Then I blocked internet access for the bridge.

    Now we wait and see if it works.

    #chrony #proxmoxve #dns #proxmox #philipshue #homelad #selfhost #selfhosted #selfhosting #offline

  22. I configured #chrony shipped with #ProxmoxVE to operate as a server and set up #DNS records on my router to point ntp[1-4].aliyun.com to the local IP of the #Proxmox node with chrony.

    According to the internet, ntp[1-4].aliyun.com are 4 of 8 NTP servers used by #PhilipsHue bridge.

    Then I blocked internet access for the bridge.

    Now we wait and see if it works.

    #chrony #proxmoxve #dns #proxmox #philipshue #homelad #selfhost #selfhosted #selfhosting #offline

  23. I configured #chrony shipped with #ProxmoxVE to operate as a server and set up #DNS records on my router to point ntp[1-4].aliyun.com to the local IP of the #Proxmox node with chrony.

    According to the internet, ntp[1-4].aliyun.com are 4 of 8 NTP servers used by #PhilipsHue bridge.

    Then I blocked internet access for the bridge.

    Now we wait and see if it works.

    #chrony #proxmoxve #dns #proxmox #philipshue #homelad #selfhost #selfhosted #selfhosting #offline

  24. I configured #chrony shipped with #ProxmoxVE to operate as a server and set up #DNS records on my router to point ntp[1-4].aliyun.com to the local IP of the #Proxmox node with chrony.

    According to the internet, ntp[1-4].aliyun.com are 4 of 8 NTP servers used by #PhilipsHue bridge.

    Then I blocked internet access for the bridge.

    Now we wait and see if it works.

    #chrony #proxmoxve #dns #proxmox #philipshue #homelad #selfhost #selfhosted #selfhosting #offline

  25. NTP in practice. Mostly it works. Time to time it's like 8 hours off. Internet connection is working.

    I already debugged unreliable NTP once and replaced the unreliable "ntp" NTP daemon with much more reliable "chrony" NTP daemon. But still unreliable.

    "Old people don't like computers" - because computers are unreliable! Why would I use something that's unreliable? A tuning fork ground out of quartz, invented by Seiko, connected to a battery, semiconductor amplifier, 15-bit counter, few logic gates, stepper motor, cogwheels, hands and a clock face is much more reliable!

    #ntp #unreliable #linuxfail #linux #chrony #ntpd #reliability

  26. NTP in practice. Mostly it works. Time to time it's like 8 hours off. Internet connection is working.

    I already debugged unreliable NTP once and replaced the unreliable "ntp" NTP daemon with much more reliable "chrony" NTP daemon. But still unreliable.

    "Old people don't like computers" - because computers are unreliable! Why would I use something that's unreliable? A tuning fork ground out of quartz, invented by Seiko, connected to a battery, semiconductor amplifier, 15-bit counter, few logic gates, stepper motor, cogwheels, hands and a clock face is much more reliable!

    #ntp #unreliable #linuxfail #linux #chrony #ntpd #reliability

  27. NTP in practice. Mostly it works. Time to time it's like 8 hours off. Internet connection is working.

    I already debugged unreliable NTP once and replaced the unreliable "ntp" NTP daemon with much more reliable "chrony" NTP daemon. But still unreliable.

    "Old people don't like computers" - because computers are unreliable! Why would I use something that's unreliable? A tuning fork ground out of quartz, invented by Seiko, connected to a battery, semiconductor amplifier, 15-bit counter, few logic gates, stepper motor, cogwheels, hands and a clock face is much more reliable!

    #ntp #unreliable #linuxfail #linux #chrony #ntpd #reliability

  28. NTP in practice. Mostly it works. Time to time it's like 8 hours off. Internet connection is working.

    I already debugged unreliable NTP once and replaced the unreliable "ntp" NTP daemon with much more reliable "chrony" NTP daemon. But still unreliable.

    "Old people don't like computers" - because computers are unreliable! Why would I use something that's unreliable? A tuning fork ground out of quartz, invented by Seiko, connected to a battery, semiconductor amplifier, 15-bit counter, few logic gates, stepper motor, cogwheels, hands and a clock face is much more reliable!

    #ntp #unreliable #linuxfail #linux #chrony #ntpd #reliability

  29. NTP in practice. Mostly it works. Time to time it's like 8 hours off. Internet connection is working.

    I already debugged unreliable NTP once and replaced the unreliable "ntp" NTP daemon with much more reliable "chrony" NTP daemon. But still unreliable.

    "Old people don't like computers" - because computers are unreliable! Why would I use something that's unreliable? A tuning fork ground out of quartz, invented by Seiko, connected to a battery, semiconductor amplifier, 15-bit counter, few logic gates, stepper motor, cogwheels, hands and a clock face is much more reliable!

    #ntp #unreliable #linuxfail #linux #chrony #ntpd #reliability

  30. Made progress on packaging by @trifectatech for @gentoo :
    ✅ proper hermetic Rust build
    ✅ cooked up OpenRC init scripts
    ✅ runs unprivileged as ntp:ntp
    ✅ operational metrics exporter
    ✅ man pages

    Still a few rough edges but getting there!
    is currently still the better choice but alternatives are always good.

  31. Made progress on packaging #ntpd-rs by @trifectatech for @gentoo :
    ✅ proper hermetic Rust build
    ✅ cooked up OpenRC init scripts
    ✅ runs unpriviligeded as ntp:ntp
    ✅ operational #prometheus metrics exporter
    ✅ man pages

    Still a few rough edges but getting there!
    #chrony is currently still the better choice but alternatives are always good.

    #ntp #rust

  32. Made progress on packaging #ntpd_rs by @trifectatech for @gentoo :
    ✅ proper hermetic Rust build
    ✅ cooked up OpenRC init scripts
    ✅ runs unprivileged as ntp:ntp
    ✅ operational #prometheus metrics exporter
    ✅ man pages

    Still a few rough edges but getting there!
    #chrony is currently still the better choice but alternatives are always good.

    #ntp #rust #gentoo

  33. Made progress on packaging #ntpd-rs by @trifectatech for @gentoo :
    ✅ proper hermetic Rust build
    ✅ cooked up OpenRC init scripts
    ✅ runs unpriviligeded as ntp:ntp
    ✅ operational #prometheus metrics exporter
    ✅ man pages

    Still a few rough edges but getting there!
    #chrony is currently still the better choice but alternatives are always good.

    #ntp #rust

  34. Made progress on packaging #ntpd_rs by @trifectatech for @gentoo :
    ✅ proper hermetic Rust build
    ✅ cooked up OpenRC init scripts
    ✅ runs unprivileged as ntp:ntp
    ✅ operational #prometheus metrics exporter
    ✅ man pages

    Still a few rough edges but getting there!
    #chrony is currently still the better choice but alternatives are always good.

    #ntp #rust #gentoo

  35. Von kompliziert zu simpel: PTP-Synchronisation mit Timemaster statt ptp4l

    Timemaster macht die PTP-Synchronisation unter Linux endlich einfach. Statt mühsam ptp4l und phc2sys zu konfigurieren, reicht eine zentrale Datei – und schon werden PTP- und NTP-Quellen automatisch eingebunden.

    #chrony #HighPrecisionTime #linux #Netzwerkzeit #ntp #phc2sys #PrecisionTimeProtocol #ptp #ptp4l #systemd #Timemaster #zeitserver #Zeitsynchronisation

    dirkwouters.de/von-kompliziert

  36. Von kompliziert zu simpel: PTP-Synchronisation mit Timemaster statt ptp4l

    Timemaster macht die PTP-Synchronisation unter Linux endlich einfach. Statt mühsam ptp4l und phc2sys zu konfigurieren, reicht eine zentrale Datei – und schon werden PTP- und NTP-Quellen automatisch eingebunden.

    #chrony #HighPrecisionTime #linux #Netzwerkzeit #ntp #phc2sys #PrecisionTimeProtocol #ptp #ptp4l #systemd #Timemaster #zeitserver #Zeitsynchronisation

    dirkwouters.de/von-kompliziert

  37. GPS-gestützter Zeitserver mit Raspberry Pi: So baust Du Deine eigene Grandmaster Clock

    Mit einem Raspberry Pi und einem GPS-Empfänger lässt sich ein präziser Stratum-1 Zeitserver aufbauen – inklusive PPS-Signal und PTP-Unterstützung. So entsteht eine kostengünstige Alternative zu teuren Grandmaster Clocks, die exakte Zeit im ganzen Netzwerk be

    #chrony #Clock #gps #grandmaster #nmea #ntp #ntpd #pi #pps #rasbian #raspberry #zeitserver

    dirkwouters.de/gps-gestuetzter

  38. GPS-gestützter Zeitserver mit Raspberry Pi: So baust Du Deine eigene Grandmaster Clock

    Mit einem Raspberry Pi und einem GPS-Empfänger lässt sich ein präziser Stratum-1 Zeitserver aufbauen – inklusive PPS-Signal und PTP-Unterstützung. So entsteht eine kostengünstige Alternative zu teuren Grandmaster Clocks, die exakte Zeit im ganzen Netzwerk be

    #chrony #Clock #gps #grandmaster #nmea #ntp #ntpd #pi #pps #rasbian #raspberry #zeitserver

    dirkwouters.de/gps-gestuetzter

  39. GPS-gestützter Zeitserver mit Raspberry Pi: So baust Du Deine eigene Grandmaster Clock

    Mit einem Raspberry Pi und einem GPS-Empfänger lässt sich ein präziser Stratum-1 Zeitserver aufbauen – inklusive PPS-Signal und PTP-Unterstützung. So entsteht eine kostengünstige Alternative zu teuren Grandmaster Clocks, die exakte Zeit im ganzen Netzwerk be

    #chrony #Clock #gps #grandmaster #nmea #ntp #ntpd #pi #pps #rasbian #raspberry #zeitserver

    dirkwouters.de/gps-gestuetzter

  40. GPS-gestützter Zeitserver mit Raspberry Pi: So baust Du Deine eigene Grandmaster Clock

    Mit einem Raspberry Pi und einem GPS-Empfänger lässt sich ein präziser Stratum-1 Zeitserver aufbauen – inklusive PPS-Signal und PTP-Unterstützung. So entsteht eine kostengünstige Alternative zu teuren Grandmaster Clocks, die exakte Zeit im ganzen Netzwerk be

    #chrony #Clock #gps #grandmaster #nmea #ntp #ntpd #pi #pps #rasbian #raspberry #zeitserver

    dirkwouters.de/gps-gestuetzter

  41. Cài đặt Chrony trên Server Linux (2023)

    Cài đặt Chrony trên Server Linux (2023) #Chrony #Linux #NTP #Server #CấuHình #ThờiGian #HệThống Chrony là một daemon đồng bộ hóa thời gian mạng mạnh mẽ và chính xác, được thiết kế để hoạt động hiệu quả trên nhiều loại máy chủ Linux. Nó là một giải pháp thay thế đáng tin cậy cho NTP (Network Time Protocol) truyền thống, với nhiều ưu điểm vượt trội.

    maychu.top/2025/06/09/cai-dat-

  42. My overnight tests finished!

    In my environment, I get the best #NTP accuracy with #Chrony when using `minpoll -2 maxpoll -2` and not applying any filtering. That is, have the client poll the NTP server 4 times per second. Anything between `minpoll -4` (16x/second) and `minpoll 0` (1x/second) should have similar offsets, but the jitter increases with fewer than 4 polls per second.

    scottstuff.net/posts/2025/06/0

    Chrony has a `filter` option that applies a median filter to measurements; the manual claims that it's useful for high-update rate local servers. I don't see any consistent advantage to `filter` in my testing and larger filter values (8 or 16) consistently make everything worse.

    When polling 4x/second on a carefully constructed test network, NTP on the client machine is less than 2 ns away from #PTP with 20 ns of jitter. I know that PTP on the client is 4 ns away from PTP on the server (w/ 2 ns of jitter), as measured via oscilloscope.

    So, you could argue that this counts as single-digit nanosecond NTP error, although with 20 ns of jitter that's probably a bit optimistic. In any case, that's *well* into the range where cable lengths are a major factor in accuracy. It's a somewhat rigged test environment, but it's still much better than I'd have expected from NTP.

  43. My overnight tests finished!

    In my environment, I get the best #NTP accuracy with #Chrony when using `minpoll -2 maxpoll -2` and not applying any filtering. That is, have the client poll the NTP server 4 times per second. Anything between `minpoll -4` (16x/second) and `minpoll 0` (1x/second) should have similar offsets, but the jitter increases with fewer than 4 polls per second.

    scottstuff.net/posts/2025/06/0

    Chrony has a `filter` option that applies a median filter to measurements; the manual claims that it's useful for high-update rate local servers. I don't see any consistent advantage to `filter` in my testing and larger filter values (8 or 16) consistently make everything worse.

    When polling 4x/second on a carefully constructed test network, NTP on the client machine is less than 2 ns away from #PTP with 20 ns of jitter. I know that PTP on the client is 4 ns away from PTP on the server (w/ 2 ns of jitter), as measured via oscilloscope.

    So, you could argue that this counts as single-digit nanosecond NTP error, although with 20 ns of jitter that's probably a bit optimistic. In any case, that's *well* into the range where cable lengths are a major factor in accuracy. It's a somewhat rigged test environment, but it's still much better than I'd have expected from NTP.

  44. My overnight tests finished!

    In my environment, I get the best accuracy with when using `minpoll -2 maxpoll -2` and not applying any filtering. That is, have the client poll the NTP server 4 times per second. Anything between `minpoll -4` (16x/second) and `minpoll 0` (1x/second) should have similar offsets, but the jitter increases with fewer than 4 polls per second.

    scottstuff.net/posts/2025/06/0

    Chrony has a `filter` option that applies a median filter to measurements; the manual claims that it's useful for high-update rate local servers. I don't see any consistent advantage to `filter` in my testing and larger filter values (8 or 16) consistently make everything worse.

    When polling 4x/second on a carefully constructed test network, NTP on the client machine is less than 2 ns away from with 20 ns of jitter. I know that PTP on the client is 4 ns away from PTP on the server (w/ 2 ns of jitter), as measured via oscilloscope.

    So, you could argue that this counts as single-digit nanosecond NTP error, although with 20 ns of jitter that's probably a bit optimistic. In any case, that's *well* into the range where cable lengths are a major factor in accuracy. It's a somewhat rigged test environment, but it's still much better than I'd have expected from NTP.

  45. My overnight tests finished!

    In my environment, I get the best #NTP accuracy with #Chrony when using `minpoll -2 maxpoll -2` and not applying any filtering. That is, have the client poll the NTP server 4 times per second. Anything between `minpoll -4` (16x/second) and `minpoll 0` (1x/second) should have similar offsets, but the jitter increases with fewer than 4 polls per second.

    scottstuff.net/posts/2025/06/0

    Chrony has a `filter` option that applies a median filter to measurements; the manual claims that it's useful for high-update rate local servers. I don't see any consistent advantage to `filter` in my testing and larger filter values (8 or 16) consistently make everything worse.

    When polling 4x/second on a carefully constructed test network, NTP on the client machine is less than 2 ns away from #PTP with 20 ns of jitter. I know that PTP on the client is 4 ns away from PTP on the server (w/ 2 ns of jitter), as measured via oscilloscope.

    So, you could argue that this counts as single-digit nanosecond NTP error, although with 20 ns of jitter that's probably a bit optimistic. In any case, that's *well* into the range where cable lengths are a major factor in accuracy. It's a somewhat rigged test environment, but it's still much better than I'd have expected from NTP.