(Transcribed by TurboScribe.ai. Go Unlimited to remove this message.)
[Speaker 6]
Alexander, I would have also one request from the 5G project to the OCS. There is one question if we turn on mTLS. If OCS is flexible to use mTLS or not mTLS and if it can handle a notify URI or callback URI with HTTP or HTTPS or is Amdox OCS aware about mTLS active or not active?
This question I got from the security colleagues. I don't know if I can write an email and put this together.
[Speaker 4]
Or maybe somebody from Amdox can already answer that. But in case not just send me a one line. I'll write an email.
Yes, just a one line.
[Speaker 6]
Please check this. It's regarding if we activate 5G mTLS, how the surrounding elements could deal with it. That mean do we have to do a big bang or is it possible that the platform can live with both options?
[Speaker 4]
So I guess we're pretty much complete. I guess we have the elephant in the room and that's the repository structure. We already had a call earlier between Amdox and Eduard.
What's the current stance of both parties involved?
[Speaker 3]
I'm just trying to pull in Harsh. He's not able to join. I think we should wait at least.
So MJ is on leave today. I want at least Harsh to be on a call for this discussion. Give me a couple of minutes to check if he's able to join.
Otherwise, I think it would be better to discuss this topic once. Either MJ will be back on Monday. But at least Harsh should be there.
[Speaker 8]
Did Martin also say he will join here?
[Speaker 2]
Martin is here.
[Speaker 8]
Martin is here. I'll just jump in. Hi Martin.
Hi.
[Speaker 5]
Stan the robot today.
[Speaker 8]
My audio is not good. For me.
[Speaker 6]
No, like last time. Also for us. But we don't.
We ignored it.
[Speaker 9]
Maybe it's something with my headphones. Yes, this is Stan. Robotic voice.
[Speaker 3]
Is there any other topic maybe which we want to cover in the meantime?
[Speaker 4]
Well, let me start in line eight. Secondary side VPN. Peering IP.
I have a request and a remark. Could somebody who is accessing from your team, who's accessing us via VPN, give it a shot via your VPN endpoints in either Dublin or India? Because that should be possible to my knowledge.
[Speaker 3]
It is. You're asking about the Citrix access, is it? Yes.
Yeah, we are able to do. I think we tried with India Pune and it worked. We are able to access Magenta Citrix when we are on Pune Amlogs VPN.
[Speaker 4]
It is working fine. Okay. That's good news.
I don't see Tamas in the call yet. So that point is still open. Then I have line 15, the host names or the names for Kubernetes.
That's something I'm still discussing with our security with Karsten Stepanek. Harsh returned the first questionnaire I sent and I will send him the second round later this week. That's pretty much all I have except for the repository issues.
Any other points from somebody else?
[Speaker 3]
For the Citrix part, Neha correct me if I'm wrong. I think we're still not able to access the Citrix from our laptop, but we are able to do it via the SVN. So I think there are two ways.
Last time I remember, when we were connected to Ottawa VPN, we could launch the Citrix directly from our laptop. Now that we have a dedicated SVN and if we want to use Citrix from there, we could be on any VPNs. Owen, Neha, can you double confirm just this bit?
[Speaker 1]
Yeah, correct, what you said.
[Speaker 7]
The way it's set up so that Ottawa is the only place that can access Magento currently, the SVN is in Ottawa. So that's why it works for everyone. It's not great because it's a lot of layers of remote desktops.
You're going to Ottawa and then back to Austria as well. But if you're saying that it should be set up now for the endpoints in Ireland and India, yeah. Okay, we'll test those then.
[Speaker 4]
Yeah, give it a shot and let me know.
[Speaker 7]
Just see if you can hit the Citrix URL.
[Speaker 3]
Yeah, Neha, you can try as well. For me, it doesn't load. It simply hangs on the Pune VPN.
I just switched to Pune VPN for testing. It hangs for me. Let me check the Citrix URL.
[Speaker 1]
Did you try using the Ottawa VPN? I think that was the most easy one.
[Speaker 7]
I think Alex is saying harsh that it's set up now for the Irish ones, maybe the Pune ones.
[Speaker 4]
Yeah, it should work. That's at least what I heard.
[Speaker 3]
No, then Alex, we have to update the comment, which we wrote on column 7.8. It's still not working. I was saying it works from within the SVN, which is again another layer, which is different from Ottawa. I'm talking about if you talk about direct access from laptop, it's only working from Ottawa.
From Pune, at least it's not working for me. Yeah, it's not working for me as well. Yeah, not working.
So, you can say not working from anywhere except Ottawa still.
[Speaker 4]
Yeah, I will need to clarify this. Okay, so as you hear, I guess these are the most pressing issues right now. So, we had a troubleshooting call earlier with Neha, Eduard.
[Speaker 1]
So, what was the agreement on this? You guys had a call earlier, right, as you said?
[Speaker 4]
Yeah, but that was just a technical call. So, we need to discuss this issue right now.
[Speaker 8]
So, if someone can just repeat what was proposed there that we have clear understanding all of us in this round here. I think maybe you can sum up again, please, what was the requirement and what is the collision there? Just to have an understanding and good explanation again.
[Speaker 2]
Okay, so I don't need an email for that. Yeah, so proposition from Andocs is basically that they will not follow the GitLab structure that we have and how Flags is set up in behind because we are able to create some additional work for them. And, yeah, our proposition is exactly opposite of that, right?
So, we have already created a structure like the unified one for all console vendors which we try to follow. So, for us, it would mean if we are not following that, then basically we just point Flags to their GitLab, basically to their group and we let them do whatever they want. But then that means that our team TDI will not then support anything else outside of the Flags functionality.
So, then the operations team of Magenta Austria or whatever may come, they will be trained by Andocs or I don't know how, but we will just not have the power over how they are doing things. This is my understanding, if I'm correct.
[Speaker 8]
Yeah, that's how I understood it too when you explained it earlier before the meeting.
[Speaker 2]
So, of course, both sides have some leverages, like what's good for each, but then at the end, it's not me who can push Magenta Austria, right? At the end, it's they who need to decide which way to go. So, it's either making TDI to follow what Andocs proposes or it's, yeah, you're going with Andocs way and then we take our hands off it.
[Speaker 6]
What are the gaps? We had the discussion with the hostnames in the past. I think this is more or less accepted to have here the different approach, but it's still deviating, Petrash.
[Speaker 1]
Okay, so there are two items. Sorry, Edward, if I can speak for a minute. You can correct me, right?
Like I had a meeting internally yesterday, right? So, what Pratik or I think MG was proposing. So, in general, I think first and the most important thing, right?
We are saying we will only have one master branch and we need to always work on this master branch, right? For all the deployments, whether it's the lab or it's the staging or it's the production environments, right? So, every time the master branch is there, we don't have a strategy for branching, correct?
[Speaker 6]
That means you don't have a reference. That means you have one branch and this is reference production. No, no, it's not that.
[Speaker 2]
No, no. What he means is just GitLab branching. So, what we do with every vendor in every NATCO is basically you have a main branch there and that is the production branch.
So, you can do whatever you want outside of that. You can create how many branches you want, but only what is the main branch, like the code and main branch, it's getting into the cluster by flags. What they propose is to have more main branches.
So, flags would not only read the main branch, but they want to switch flags to read different branches at different points of time.
[Speaker 1]
Yeah, and why we are saying this is we have the experience in or in TDG where this is the way we are doing the things and it's a proven way for production handling also, right? So, both the places, the system is live taking the 5G SA traffic, right? And this is the work is actually flying, right?
So, even they are very happy with this. So, is there a difference between TDG and Megia? I think they're also on Chef or on the same TDI platform, right?
So, what's the difference here?
[Speaker 2]
So, Megia is doing it differently as well. So, okay, let's not talk about Megia. So, but this is the other issue.
This is not the one that we were supposed to talk about. Okay.
[Speaker 1]
I tell you my thing, right? We are only responsible. We are doing much more than what we are responsible here in this project to handle the software deployment, right?
As per the written black and white agreements, it very clearly says it will be Megia Telecom who will be using their own pipelines and deploying the stuff, okay? In order to progress quickly, because we have lost so much of time for this environment, branching, ways of working, methodologies, blah, blah, blah, right? We went ahead and we did our own, like, we checked with both the networks where we are doing with something with Flux, right?
And based on this, how our product is best suited, based on the best practices which are proven in the production in touch telecom group, right? So, that's what we have adopted, right? We see one basic issue here in the strategy which you guys are proposing is like we only have one master branch.
So, let's say if I make a change, right, without even testing in one environment, the change will go boom, boom, boom in all the environments, right? Which can have a very negative impact on your production later on. That's the way we can foresee one of those things.
Pratik, correct me if I'm wrong here, right? Like, in our example yesterday, when we were simulating based on what was the proposal, there were like a lot many such touch points which can create some havoc in the production, right, if we make something by mistake, correct?
[Speaker 2]
Isn't that the reason why you have more environments, like, so you have test environment, you have ref environment, you have prod environment. So, if you want to test something, you don't need to test it on main branch in prod, right? You can test it on main branch in ref or in test.
So, that's what I don't understand.
[Speaker 1]
But if we want to apply change, let's say in one environment.
[Speaker 5]
The problem is, and here we have to be careful, of course, MDocs, as far as I know, is taking care about the more or less complete life cycle. So, Karl, Magenta team, it's more or less up to you. I see the biggest issue is that it is very crucial, very difficult to see on the first glance which branch is now on which cluster, the one which Flux is monitoring and doing something.
And it makes a lot of confusion if this is not completely documented and everybody is aware of it. Otherwise, everybody is doing something on a different branch. If only the main branch is the one everybody knows, only the main branch is the one which is contributed or deployed into the cluster.
But as we are saying, at the end, we are not operating it later. And if you think a branch concept will not confuse you later, and it could be that in the one cluster today it's branch X and tomorrow it's branch Y, and next week it's main back, and everybody is fine with it. We are not the one who is saying no.
We only say that we see a lot of confusion potential and you have to change all the time the Flux config to change on which branch it should look.
(Transcribed by TurboScribe.ai. Go Unlimited to remove this message.)