Self sign-up has been disabled due to increased spam activity. If you want to get access, please send an email to a project owner (preferred) or at gitlab(at)nic(dot)cz. We apologize for the inconvenience.
It would be nice if we could set more than 4 IP within forward policy (http://knot-resolver.readthedocs.io/en/stable/modules.html#query-policies)
Because forward policy is set as default in omnia, we have to cut these additional IP in omnia init script and few users were complaining about it.
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information
Child items
...
Show closed items
Linked items
0
Link issues together to show that they're related.
Learn more.
Yes, reputation tracking currently doesn't work with stub nor with forwarding https://gitlab.labs.nic.cz/knot/resolver/issues/125 but I thought it might be desirable in some use cases. Otherwise I can't see why to have so many IPs in the list...
There's a very fast (250ms by default, 100ms in Omnia) re-transmit that iterates through the list very quickly. We might improve that logic a bit (see #208 (closed)), but generally with caching it should be already good enough. Having more that four (2 IPv4 + 2 IPv6) addresses on that list doesn't make any sense for static network device like router.
If you want to do the selection of fastest FORWARD target, you can do something like this in the init script:
#!/bin/sh for NS in "$@"; do RTT=$(dig +noall +stats +time=1 +retry=0 . IN DNSKEY @$NS | sed -ne 's/;; Query time: \(.*\) msec/\1/p'); if [ -z "$RTT" ]; then RTT=1000 # anything bigger than timeout equals timeout elif [ "$1" != "${1#*:[0-9a-fA-F]}" ]; then RTT=$(($RTT - 20)) # give IPv6 a 20 ms headstart fi printf "%s;%s\n" "$RTT" "$NS";done | sort -n | cut -f 2- -d \; | head -4
I suppose a significant part of the point is that the init/config scripts on Omnia have to decide somehow which four NSs to keep if it gets more of them.
Well, I think it wouldn't take too much work to extend the NS list, perhaps exclusively for FORWARD+STUB modes and not recursive one, but let's close until we plan to actually implement it.
Is it really common on Omnia to get more than four addresses?
Personally I don't think it's too common, but I can imagine that this can happen in case that someone is using 3G or some other type of fallback connection.