As a stopgap measure that allows old config to load, I've made the script cap it at 100:

Cherry-picked into crux.

@rgrant I've cherry-picked it into Crux, will be in the 1.2.3 release soon.

Cherry-picked into crux.

Cherry-picked into crux.

Cherry-picked into crux.

CHerry-picked into crux.

Cherry-picked into crux.

Cherry-picked into crux.

Imported 2.0.17 into crux.

Actually, there are already run monitor traffic interface $intf commands and they work fine in 1.2.3.

If it reappears, feel free to reopen.

@alkersan Looks good, thanks!

I think we should use it as a chance to improve it. The nesting does not do anyone any favors, tcpdump applies to all interface type, so a single level "monitor interface $intf" command is going to be easier to use.

Fixed. Should in the next nightly build.

I've fixed the setup and re-populated the equuleus repo.

I could reproduce it. The DHCP client script overwrites those entries.

We cannot backport anything that introduces a syntax or behaviour change, much less if it may possibly need a migration script. ;)

@zsdc I've included keepalived 2.0.17 in the rolling release repo and it will be in tomorrow's nightly build. Please test.

I would be wary of including it in Crux, since it's a behaviour change.

This is really a broken abstraction. There is no separate namespaces for IPv4 and IPv6 groups in IPset.
We'll have to autoprefix the groups or similar, if we want it to work that way.

@ekim First of all, a quick fix for your situation with next-hop: add { } after the next-hop that is missing it. I.e.:

@c-po We have plurals everywhere we have more than one option under a subtree. dhcpv6-options, offload-options, "system options"
Since there's also the hostname option, and possibly other options we may want to send, that makes sense I guess.