Based on what awakerix said “…more like a race on health check when you’re suspend/resume.”, I updated my router’s QoS setting to give NTP service a higher prioerity. It might have solved the issue!
Thanks, All
Appearently, TrueNAS prefer the system stays on 24/7. I have been experiencing not only NTP no peers issues. My pool drive was randomly put out of service. The flaged drive always tested clean with “smartctl long” and “badblock” in a separate Linux environment.
I disabled the “rtcwake” corn job early November and the system has been running without issues. I put the drives (two of them) that were put out of service back in the pool and they are running error free.
I believe how TrueNAS handle all system resources and events upon system wake up need some scrutiny.
Best.
My system is 24/7 up.
I see. I do have us.pool.ntp.org added as the first NTP server beside the debian server. Maybe you can try it. If you’re not in US, going to https://www.ntppool.org and look up one that will work for you.
Best
IIRC, NTP gets really unhappy if the clocks run too far apart between updates and TrueNAS has a allergic reaction if system times bounce around.
ISPs have been known to turn off NTP access due to potential hacking. That may or may not allow your NAS to drift more than it ought to.
I am a big fan of CenterClicks NTP2xx series’ to give you an excellent, inexpensive, stable, easy-to-use NTP at home. I you like, read a bit more about it on the old forum.
Most security-related libraries do not like time to bounce around. Kerberos, totp, etc, etc. This isn’t really a TrueNAS thing. We just alert on it because such things matter to non-home use case.
3 Likes
I got this error this night for the first time, and that was after I upgraded to 25.10.2 last evening:
NTP health check failed - No Active NTP peers: [{‘SERVER: NOT_SELECTABLE [94.198.159.11]’}, {‘SERVER: SELECTABLE [77.174.62.158]’}, {‘SERVER: NOT_SELECTABLE [178.238.99.189]’}]
I am on the default NTP settings.
This is my backup server which is virtualized in proxmox.
The main server has not been upgraded yet.
Did you compare to the NTP settings in your main server or try to ping those addresses from the upgraded server?
From my 25.10.1 - Goldeye
I have those NTP servers, but did not try to ping them. Didn’t get around to that yet.
Strangely, this afternoon the error cleared, exactly 12hours to the minute after the warning this night.
I added 0.us.pool.ntp.org as the first NTP server besides the default debian servers, and stop putting the system to sleep seemed did the trick. I havn’t encunter this NTP peers issue since December post.