Protobuf schema has been written.

@jpbostic My idea for interacting with vyconf from outside the interactive shell is a bit different. The issue with 'vyshell -c "set interfaces ethernet eth0 disable' is that it needs to setup a session first, and store the session ID between commands, so either it will be limited to 'vyshell -c "configure; set interfaces ethernet eth0 address 192.0.2.1/24; set interfaces ethernet eth0 mtu 1400"' (i.e. long command strings in single call), or it will be dependent on specific environment setup, and from VyOS 1.x we already know how problematic it will be.

Worse, such config can be saved, but cannot be loaded afterwards because the formatter doesn't bother to quote such strings and you end up with a syntactically invalid config.

@jpbostic I think @mickvav has a point here, making it configurable is relatively cheap (it's just a field in /etc/passwd after all, or a call to chsh). People who don't want it to be the deault like myself and service accounts for ansible etc. can just change the default.

Once @tmartinson setups a physical server for us (I'd like to say thanks to him, by the way!), it will become a permanent place for the jenkins VM and build hosts where we can give access all maintainers without worrying about mixing Sentrium corporate stuff with it.

@tmartinson Well, you should change your vote then (votes are not final here, for the better I guess).

@systo Just to make sure you are looking at it the right way, in the large it's actually less verbose than old syntax. The vif may not be the best example but firewall would make it apparent:

@rps An serious issue with "interfaces { eth0" is that when there is no parent subtree of all ethernet interfaces specifically, we don't know which script to call when something in "eth0" changes. We'd have to have one big script that handles the whole "interfaces" subtree, which is very problematic when it comes to adding new interface types. If eth* interfaces are children of the "ethernet" node and tun* interfaces are children of the "tunnel" node, it's easy to attach ethernet script to the "ethernet" node and "tunnel" script to the "tunnel" node, if we want to add "openvpn" later, we won't have to modify that large script to accomodate it

@rps No, that's not the biggest challenge. Semicolon at the end of leaf nodes makes them unambiguous enough and easy to tell from tag nodes (this is especially bad with valueless nodes by the way, think "disable", colon wouldn't help there, but semicolon at the end does the job). The biggest challenge is that with "ethernet eth0" the parser must be fully stateful and capable of tracking which parent nodes it's already seen. "eth0", "eth1" etc. are really children of the same node called "ethernet", but in the config they appear separately. Consider this unusual but logically valid config:

@Merijn I'm still not sure why JunOS has that "unit" thing. To me it looks redundant, redundant ©. Though what we are discussing is "unit 0" vs "unit { 0" grammatic distinction, rather than specific syntax of ethernet interfaces.

@tmartinson No, "vlan-id 99" is the old style. And, at that stage we don't know if it's ethernet or not.

@Merijn Now that you remind me of it, I think "edit interfaces tunnel; copy tun10 to tun11" or similar should be possible regardless of the config syntax. No matter how it looks in the config, internally "tunnel" is a node with children "tun0", "tun1" and so on, and there's no reason why it shouldn't be possible to use it as edit level.

@dsteinkopf Not sure, we'll have to devise some rules regarding line breaks, and past some number of leaf nodes inside we are back to the original aesthetic issue (and then there can be non-leaf nodes inside too!
On a fresh look today, I'm convinced that the old tag node formatting is aesthetically superior, so myself as a user of my own project I'm probably voting no, though as a developer I want to see how many people also think it's worth it.

Yes, related. I was just talking to myself really, we get the CI back first, and then we can look into adding vyconf to it.

Our gateway is bad and we should feel bad. When jenkins migration to the new site is complete (we are migrating build hosts too), this should work again.

Thanks! Unit tests pass.

Unit tests pass for me too.

I made it return Unknown if the file in question doesn't exist, hope this fixed the issue.

Seems to work. @hexes, please test and tell me if for you it doesn't.

Duplicate indeed.

Does openfastpath really work? Have you tried it? It all looks great, and if it works reliably, we indeed should integrate it.

I moved the "task" part to T215.

Hi Carl,

Someone please remove the vyconf tag from this task, it has nothing to do with it. It can be added to 1.2.0 in fact!

Looks like it works, and the tests pass.

@thomas.courbon I think we fixed a similar bug before, but it seems the fix wasn't quite complete...

@hexes Could you update the task and specify which image you use and which error you get in it?

Looks like a bug indeed.

Hi Alex,

@Alexis I wish, with this shortage of contributors, we were really in position to make specific plans regarding the timeframe. ;)
This also applies to the security updates issue. We really need a dedicated security watcher, but, sadly, no one wants to take up this role, so it's always done in a haphazard manner, which is a bad experience for both end users and developers, but that's what we've got.
In a few years of project life, the number of people committed to using VyOS in production grew, but the number of people committed to developing it almost did not, it's still just a few people who have to do everything, and, frankly, it's taxing. At this point, none of us can turn it into a full time job (the commercial support thing @syncer and I started may change it in the future and give some of the maintainers guaranteed N hours a week to spend on it, but it's still a very early stage).

@Alexis By the way, if you are into the quagga source code, maybe you want to join the work on switching to the upstream or cumulus quagga and "forth-porting" vyatta changes to it?

@Alexis Please don't panic. This bug is only exploitable if RA handling is enabled in quagga, and by default it is not. Setting interface's IPv6 to autoconf doesn't enable it in zebra either.
I agree it should be included in 1.1.8, but it's not urgent. I suppose we'll build 1.1.8 some time next week anyway, there are other issues to be addressed.

Maybe really generic one should not reference specific companies, not sure.

I've added the mirror to the load balancer. Please add it to the wiki page: http://wiki.vyos.net/wiki/Mirrors
Also, what's your update interval?

Anything local should hardly be considered high priority, but the first one is remote. We definitely should include the fix in the 1.1.8

Hey, could you rename the task to make it more obvious what it's about. The default meaning of GCC is GNU Compiler Collection. ;)

Looks like a worthy endeavor.

I have some ideas why can it be... We'll have a look.

As @UnicronNL says, lines about nonexistent interfaces have no effect on dnsmasq functionality.

Could you tell us why these hostnames are wrong and what are the right hostnames you expected?

Well, I'll reiterate. Interactice SSH works fine to the device in question. The library is giving timeouts. Interactice SSH is not giving me timeouts. From this I deduce that the problem must be in the library, or in exscript.

Well, you realize that "execute one more time" is absolutely not an option for a library that is supposed to allow making configuration changes unattended from SDN controllers etc. among other things, right? ;)

Oops. Fixed, try now.

What we need for certain, other than git and build-essential, assuming the base system is minimal install:

Also, I can't edit this task sadly, but no way the difficulty is less than an hour... It's pretty involved, open-vm-tools source repo doesn't make it particulary easy to debianize...

@brona Do you know exact reproducing procedure?
We noticed this problem with Veeam Backup and Replication tool that uses quiescence, but we could never reproduce it simply by making a snapshot of a test VM.

For the jessie branch, we are likely staying with systemd, so systemd-related code will have to be added anyway.
Whether to remove the inittab-related code or not, not sure. I don't think we should implement both right away, but if we leave some room for extending it to support other init systems, it should be fine I think.