Rare project developers

[Alexander Gall <>]

Hi

On Mon, 28 Mar 2022 19:06:05 +0200, mc36 <> said:

> and yet another, /etc/freertr/p4-profile need to be set to GEANT-TESTBED...

You mean as the default? It is set here:
https://bitbucket.software.geant.org/projects/RARE/repos/rare-nix/browse/default.nix#38

But GEANT_TESTBED is probably not a good default since there is only
one ONIE installer for everybody and this profile is not really useful
to anyone else. It seems clear that some amount of customization needs
to be done after a fresh ONIE installation. How this is done is kind
of out of the scope of the installer. In the lab, this could be done
via Ansible, I guess.

> i failed to find this and the other location... where these files come from?

The ONIE installer is not part of the rare-nix repo.  The core of it
is a separate repo
https://github.com/alexandergall/onie-debian-nix-installer (and that's
where I'll add the capability to pin packages), but the way it is used
by rare-nix is a bit more complicated.  A SDE-specific installer is
provided by the bf-sde-nixpkgs package
(https://github.com/alexandergall/bf-sde-nixpkgs/tree/master/bf-sde/support#mkonieinstaller),
which is then called by rare-nix here:
https://bitbucket.software.geant.org/projects/RARE/repos/rare-nix/browse/default.nix#86

I'll take care of this as it involves quite a bit of very Nixy stuff.

-- 
Alex

> thanks,
> cs


> On 3/28/22 10:27, mc36 wrote:
>> hh,
>> any one more thing, when you're installing in the chroot environment in 
>> the onie loader,
>> please consider executing "apt-mark hold linux-image-amd64" otherwise a 
>> apt update / apt upgrade openssh-server would ruin the box!
>> thanks,
>> cs
>> 
>> 
>> On 3/28/22 10:20, mc36 wrote:
>>> hi,
>>> thanks... i'll execute this as part of my regular weekend routine...
>>> yeahh, i already depolyed the new one at par0001 running the latest nix 
>>> builds already...
>>> now i'll play a bit with the nix way of upgrade if you don't mind....
>>> i see in release-manager/activation.sh that you stop then restart the 
>>> service...
>>> imho you're having a good reason to do so, but, since we'll call it from 
>>> the service itself, it never reaches the reactivation part...
>>> i'll add a small daemonRun.bin too to overcome this... or any better idea?
>>> thanks,
>>> cs
>>> 
>>> 
>>> 
>>> On 3/28/22 10:16, Alexander Gall wrote:
>>>> Hi
>>>> 
>>>> On Mon, 28 Mar 2022 09:24:33 +0200, mc36 <> said:
>>>> 
>>>>> hi,
>>>>> what about this commit?
>>>>> https://bitbucket.software.geant.org/projects/RARE/repos/rare-nix/commits/0a015696e95de22cf94a2eabeddc011299b7c303
>>>> 
>>>> You can check the status of the builds at
>>>> http://hydra.nix.net.switch.ch/jobset/RARE-OS/development
>>>> 
>>>> A build is started when a commit is pushed to the master branch (to be
>>>> precise, Hydra polls the Git repo every 5 minutes to check for new
>>>> commits).
>>>> 
>>>> Any errors during the build would show up there.     You got lucky with
>>>> ptyRun sinc libutil is already part of the standard build environment
>>>> :)
>>>> 
-- 
SWITCH
--------------------------
Alexander Gall, Network
Werdstrasse 2, P.O. Box, 8021 Zurich, Switzerland
phone +41 44 268 15 15, direct +41 44 268 15 22
, http://www.switch.ch