The Grafana instance at https://accessdev.nci.org.au/grafana will need a new home when Accessdev is retired.
One solution would be to stand up a new instance on another cloud tenant. Perhaps we could have a shared ansible setup so each institution can have their own server.
Any thoughts @Aidan, @dale.roberts?
It’s one of the thing we would like to discuss at the NRI workshop, I would like for this to be part of a broader plan on all shared infrastructure as opposed to tackle them one by one.
Sounds good to me, does this fit into the existing discussions or should we set up a separate meeting?
Definitely keen to collaborate. We had (tentative) plans to host on a Nectar/ARDC VM, but made no progress on that to date.
I agree with @Paola-CMS that this is part of a broader discussion. My apology, I never did resuscitate the infrastructure meetings as their raison d’être was taken away a bit by the NCI spack model building effort of Ben’s team.
We can revive those meetings though, if we think that is worthwhile.
There’s no meeting setup at the workshop I don’t think there’s a session for the infrastructure, all CMS is going to be there, however not every day, definitely Dale and I will be there all week, so maybe Friday if that’s less busy? I’m not sure who exactly should be there from a NRI point of view
That’s correct, but I’m sure we could carve out some time, and clearly visitors are the time constrained ones, so happy to do what works for you all.
I am taking a COSIMA training session at 11.30 on th 8th I believe, but apart from that I am free all Friday.
Hey @Scott
I can’t say that I’m a fan of either of those ideas. I’m trying to move CLEX more towards managed services and away from NCI at the same time. Letting people like us loose on blank VMs with sensitive data and public DNS entries is a recipe for getting said sensitive data stolen. Here is a particularly pertinent example: U.S. ‘No Fly List’ Leaks After Being Left in an Unsecured Airline Server. Unfortunately I haven’t come across a managed service similar to grafana that doesn’t cost a lot. This is why, as @Paola-CMS says, we need a plan. Is there a plan for NRI to hire sysadmins and/or cybersec specialists? If so, blank VMs stood up with ansible start to look more viable. What about reducing the reliance on NCI? Looking beyond NCI offers opportunities for researchers to access more compute and storage resources, and the accessdev shutdown can be a catalyst for that.
I’ll be in Canberra Monday to Thursday
I’m sad that the CLEX Accounting Dashboard is no longer working for the ANU projects (x77/g40). The note there says “This dashboard is deprecated. …some projects may be removed without notice.” I’m wondering if we can come up with any solutions for this moving forward?
Rui at NCI has set up a ‘voila’ thingie on ARE that can do jupyter-based dashboards. Could be a solution for monitoring stuff like this
I’ll be at the workshop next week, if you would like to discuss this. We tried to find someone who would maintain this service and unfortunately there was no uptake, I think the NRI has their own grafana and so they were not interested. The current CLEX dashboard will close completely in December, unless we find someone who is interested in taking it over.
Was this sorted? I’m trying to find who is causing inode limits to blow up in v46 and in the past I’d open grafana… What’s the latest way of doing that?
ACCESS-NRI is tracking our projects, but not external projects.
The CLEX service has been closed with the end of the centre.
There is a COECMS blog about managing storage etc
Hope that is helpful.
Projects ACCESS-NRI is responsible for. This includes the ACCESS Working Group projects and some data projects.
There is nothing stopping others doing this, but yes as far as I know there is no-one doing this for v46, but I have nothing to do with that project, so what I know isn’t that useful.
is not a thing of the past, right?