Open 1o1o1 opened 2 years ago
you can use --min-interval
for that. it will introduce artificial rate limiting to all the jobs. Another option would be to reduce the amount of jobs running by using --scale
with some fractional value
it's not a full scale speed limiting (you would need to experiment to find the value that works for you) but doing that would require synchronizing work threads resulting in overall performance degradation
Ok) looks like working well 👍 "-scale 0.6" is good for Archlinux on a VPS ... "-min-interval 2s -scale 0.6" is well for archlinuxarm on my routerboard espressobin :)
First of all you may just increase the nf_conntrack
limit with smth like
sudo sysctl net.netfilter.nf_conntrack_max=128000
You may also limit the overall outgoing throughput of your network interface with smth as simple as
sudo tc qdisc add dev eth0 root tbf rate 20mbps latency 50ms burst 1mb
this would limit the outgoing traffic on eth0
interface to 20 megabytes (160 megabit) per second, but for the whole interface, not just db1000n.
Can you add limit download/upload speed? :)
trickle don't work whit db1000n)
I have this msg in logs: "kernel: nf_conntrack: nf_conntrack: table full, dropping packet" :)