CN pool collapse a few hours every day

TLDR, adding ipv6 to the main domain and let everybody to use it is the way to go.

2 Likes

Sure, and flipping the switch on the wall is what makes the light bulb glow. And when it doesn’t, just keep yanking the switch until it does. :wink:

But now is the case that they have cn.pool.ntp.org configurated and
not something else.

Its just to much traffic during peak but I have identified a few networks
that just sends too many requests. No idea if its just broken
or intensional. Or maybe many behind NAT.

I have had them blocked for many years. I notice now that a few of
them actually have stopped sending traffic and some are still sending
traffic.

Many wise and interesting comments in this thread.
No need to repeat things.

However, no word from ask and no action has happend and I cant wait
so I have lowerd the bandwidth even more - so hopefully I can survive
the CN pool collapse better next time.

Sadly, a consequence of this, alot of bandwidth and resources wont be
used at all.

It would be better if they could be used in the global pool but not in
the CN pool during peak time. If that would be possible @ask ?

Can the CN zone server operators take a look to their traffic in the next days ?

For me it has been greatly improved.
Almost no red dots and not as bad score anymore.

Im not sure exactly how many servers that was in the pool during the
peak but I noticed that it was much better and that is the same for
servers that the change I did 2 days ago didnt affect.

What did you do? :slightly_smiling_face:

Tomorrow I will take a closer look during peak.

Sounds good. Pressing both thumbs that it will work in the future.

Long storry - short: I’ve asked a big company which is already present in the pool for some help and they saied: yes, sure :slight_smile:

4 Likes

It seems like it did.
I was a bit suprised and I wonder where did the traffic go…

I did as I said and checked during peak 4 days ago.
Yes, it was a bit higher during peak but it was a quite “flat” bump.
Approx 26 active servers and -8 during peak, that has been repeated
over the next days.

I guess the problem is solved now.
Say thanks to the company, it really helped :slightly_smiling_face: