Mailinglist Archive: yast-devel (54 mails)

< Previous Next >
Re: [yast-devel] RFC: Yast Team Infrastructure: Yast Jenkins Node
On Tue, 24 Jun 2014 09:51:41 +0200
Lukas Ocilka <lukas.ocilka@xxxxxxxx> wrote:


Today we have an outage of the Yast Jenkins node because of "No space
left on device". This is a virtual machine maintained by a team member
(including scripts around). It's not the first time this node has died
for some reason.
Thread has died No space left on device

As this is a core infrastructure of the Yast team and everyone depends
on it (as we got used to it and OBS/IBS also expects SRs to be done
this way), I would like to change the current rules this way:

1. This is a core infrastructure --> it should run on a
Yast-team-owned server (we have a HW ready in server room)

Nice idea, but I think that this won't work, as external jenkins node
need to be in separated location to allow access from external network.
But we can discuss it with Daniel.

2. Any (core) infrastructure, including scripts, should be owned by
Yast team

I agree.

3. Everything needs to be at GitHub following the same (but stricter)
rules for merging changes (continuous deployment from GitHub?)
For instance, I'd recommend to require two different LGTMs instead
of one as it is now

Well, now it is on susestudio including script and excluding
credentials as I don't want to have it public.

4. Everything needs to be well documented in a way that "everyone" in
the team should be able to recover from such error or even start a
new node if something goes really bad

My plan for internal node is to run two nodes, so if one goes wrong,
then it should not be problem. For external it is a bit tricky.
Currently access is granted for me, mvidner and Daniel and Bernhard.

5. Everyone in the team needs to have a root (e.g., via ssh keys)
access but not to change anything they do not understand

not a problem if I have ssh public keys of everyone.

6. We need monitoring, maybe internal IT guys can do that for us and
they might have and access to that system for urgent cases as well

jenkins itself do monitoring, maybe just some reporting missing?

All this doesn't need to be changed "right now". One little step at
the time. Then another.


Thanks in advance

To unsubscribe, e-mail: yast-devel+unsubscribe@xxxxxxxxxxxx
To contact the owner, e-mail: yast-devel+owner@xxxxxxxxxxxx

< Previous Next >