[Bloat] Router congestion, slow ping/ack times with kernel 5.4.60

Thomas Rosenstein thomas.rosenstein at creamfinance.com
Sat Nov 7 08:33:32 EST 2020


On 7 Nov 2020, at 13:37, Thomas Rosenstein wrote:
>
> I have also tried to reproduce the issue with the kernel on a virtual 
> hyper-v machine, there I don't have any adverse effects.
> But it's not 100% the same, since MASQ happens on it .. will 
> restructure a bit to get a similar representation
>

I reproduced something similar now, for the VM it only happens at 120 
threads but I have every 5 or so seconds one packet that's delayed 
heavily:

PING 10.10.0.69 (10.10.0.69) 56(84) bytes of data.
64 bytes from 10.10.0.69: icmp_seq=1 ttl=63 time=0.273 ms
64 bytes from 10.10.0.69: icmp_seq=2 ttl=63 time=0.260 ms
64 bytes from 10.10.0.69: icmp_seq=3 ttl=63 time=0.621 ms
64 bytes from 10.10.0.69: icmp_seq=4 ttl=63 time=0.218 ms
64 bytes from 10.10.0.69: icmp_seq=5 ttl=63 time=52.7 ms
64 bytes from 10.10.0.69: icmp_seq=6 ttl=63 time=0.261 ms
64 bytes from 10.10.0.69: icmp_seq=7 ttl=63 time=0.381 ms
64 bytes from 10.10.0.69: icmp_seq=8 ttl=63 time=0.230 ms
64 bytes from 10.10.0.69: icmp_seq=9 ttl=63 time=0.378 ms
64 bytes from 10.10.0.69: icmp_seq=10 ttl=63 time=0.268 ms
64 bytes from 10.10.0.69: icmp_seq=11 ttl=63 time=0.313 ms
64 bytes from 10.10.0.69: icmp_seq=12 ttl=63 time=0.395 ms
64 bytes from 10.10.0.69: icmp_seq=13 ttl=63 time=0.269 ms
64 bytes from 10.10.0.69: icmp_seq=14 ttl=63 time=44.1 ms
64 bytes from 10.10.0.69: icmp_seq=15 ttl=63 time=0.279 ms
64 bytes from 10.10.0.69: icmp_seq=16 ttl=63 time=0.406 ms
64 bytes from 10.10.0.69: icmp_seq=17 ttl=63 time=0.245 ms
64 bytes from 10.10.0.69: icmp_seq=18 ttl=63 time=0.202 ms
64 bytes from 10.10.0.69: icmp_seq=19 ttl=63 time=0.229 ms
64 bytes from 10.10.0.69: icmp_seq=20 ttl=63 time=0.357 ms
64 bytes from 10.10.0.69: icmp_seq=21 ttl=63 time=0.462 ms
64 bytes from 10.10.0.69: icmp_seq=22 ttl=63 time=0.300 ms
64 bytes from 10.10.0.69: icmp_seq=23 ttl=63 time=60.3 ms
64 bytes from 10.10.0.69: icmp_seq=24 ttl=63 time=0.200 ms
64 bytes from 10.10.0.69: icmp_seq=25 ttl=63 time=15.8 ms
64 bytes from 10.10.0.69: icmp_seq=26 ttl=63 time=0.296 ms
64 bytes from 10.10.0.69: icmp_seq=27 ttl=63 time=0.273 ms
64 bytes from 10.10.0.69: icmp_seq=28 ttl=63 time=0.591 ms
64 bytes from 10.10.0.69: icmp_seq=29 ttl=63 time=0.583 ms

during that time the interface is loaded with 40 Mbits.

this does not appear with 3.10 kernel:

PING 10.10.0.69 (10.10.0.69) 56(84) bytes of data.
64 bytes from 10.10.0.69: icmp_seq=1 ttl=63 time=0.322 ms
64 bytes from 10.10.0.69: icmp_seq=2 ttl=63 time=0.242 ms
64 bytes from 10.10.0.69: icmp_seq=3 ttl=63 time=0.461 ms
64 bytes from 10.10.0.69: icmp_seq=4 ttl=63 time=0.244 ms
64 bytes from 10.10.0.69: icmp_seq=5 ttl=63 time=0.429 ms
64 bytes from 10.10.0.69: icmp_seq=6 ttl=63 time=0.268 ms
64 bytes from 10.10.0.69: icmp_seq=7 ttl=63 time=0.540 ms
64 bytes from 10.10.0.69: icmp_seq=8 ttl=63 time=0.465 ms
64 bytes from 10.10.0.69: icmp_seq=9 ttl=63 time=0.237 ms
64 bytes from 10.10.0.69: icmp_seq=10 ttl=63 time=0.216 ms
64 bytes from 10.10.0.69: icmp_seq=11 ttl=63 time=0.248 ms
64 bytes from 10.10.0.69: icmp_seq=12 ttl=63 time=0.247 ms
64 bytes from 10.10.0.69: icmp_seq=13 ttl=63 time=0.282 ms
64 bytes from 10.10.0.69: icmp_seq=14 ttl=63 time=0.256 ms
64 bytes from 10.10.0.69: icmp_seq=15 ttl=63 time=0.268 ms
64 bytes from 10.10.0.69: icmp_seq=16 ttl=63 time=0.272 ms
64 bytes from 10.10.0.69: icmp_seq=17 ttl=63 time=0.236 ms
64 bytes from 10.10.0.69: icmp_seq=18 ttl=63 time=0.287 ms
64 bytes from 10.10.0.69: icmp_seq=19 ttl=63 time=0.288 ms
64 bytes from 10.10.0.69: icmp_seq=20 ttl=63 time=0.297 ms
64 bytes from 10.10.0.69: icmp_seq=21 ttl=63 time=0.319 ms
64 bytes from 10.10.0.69: icmp_seq=22 ttl=63 time=0.294 ms
64 bytes from 10.10.0.69: icmp_seq=23 ttl=63 time=0.313 ms
64 bytes from 10.10.0.69: icmp_seq=24 ttl=63 time=0.208 ms
64 bytes from 10.10.0.69: icmp_seq=25 ttl=63 time=0.222 ms
64 bytes from 10.10.0.69: icmp_seq=26 ttl=63 time=0.279 ms
64 bytes from 10.10.0.69: icmp_seq=27 ttl=63 time=0.257 ms
64 bytes from 10.10.0.69: icmp_seq=28 ttl=63 time=0.266 ms
64 bytes from 10.10.0.69: icmp_seq=29 ttl=63 time=0.398 ms
64 bytes from 10.10.0.69: icmp_seq=30 ttl=63 time=0.281 ms
64 bytes from 10.10.0.69: icmp_seq=31 ttl=63 time=0.263 ms
64 bytes from 10.10.0.69: icmp_seq=32 ttl=63 time=0.276 ms

might be a hint that it's actually reproduceable.


More information about the Bloat mailing list