MLABs server down

[Jack Binks]

Hi There,

We've recently been having some issues with running speedtests to the IP 212.113.31.24 .

We use this IP as our fallback, however it appears to no longer be up. Is this server just down for maintenance? or will it no longer be used?

Best,

Jack

[Nathan Kinkade]

Hi Jack,

The IPv4 address you specified was part of the M-Lab site named LHR05, which was retired a few weeks ago. Is your integration using the M-Lab Locate Service, or is it directly targeting machines?

Best regards,

Nathan

--
You received this message because you are subscribed to the Google Groups "discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to discuss+u...@measurementlab.net.
To view this discussion on the web visit https://groups.google.com/a/measurementlab.net/d/msgid/discuss/CA%2B03w5ivRYZ352mktw06hfpySVk%2Bj0G2mb%3DGhStMTpQu5D4WjQ%40mail.gmail.com.

[Jack Binks]

Hi Nathan,

+Luke Meppem 

We are using the MLABs Locate service in our integration. However, we had the 212.113.31.24 IP hardcoded as a fallback, should the locator service not work for any reason.

Would it be possible to provide another server IP (or multiple) that could be used as a fallback(s) in this scenario?

Best,

Jack

[Nathan Kinkade]

Hi Jack,

I apologize that the retirement of M-Lab site LHR05 is causing problems with your production environment. Unfortunately, the circuit for LHR05 was turned down by Lumen, along with the IPv4 prefix that went with it. We haven't generally attempted to announce when new sites will be turned up or older sites turned down, because the expectation has largely been that all integrations should be using the M-Lab Locate Service to get a list of nearby, healthy machines to test against. You really shouldn't need statically configured fall-back addresses to test against, as the Locate Service should be highly available. Have you found that it isn't?

Detailed information about every M-Lab machine is publicly available:

https://siteinfo.mlab-oti.measurementlab.net/v2/sites/sites.json

There are various other more abbreviated formats of the same data:

https://siteinfo.mlab-oti.measurementlab.net/v2/index.html

This may not be possible in your case, at least in the short term, but I would recommend against configuring a new set of static IP addresses to test against, and instead put the engineering effort into relying 100% on the Locate Service (different from "mlab-ns"):

https://locate.measurementlab.net/v2/nearest/ndt/ndt7

https://locate.measurementlab.net/v2/nearest/ndt/ndt5

There is considerable churn of M-Lab's physical infrastructure at the moment, and you may very well statically configure new addresses that will themselves go away.

Best,

Nathan

On Tue, Sep 26, 2023 at 9:50 AM Jack Binks <jack....@netduma.com> wrote:

Hi Nathan,

Just following up on my previous email. I know this may be a longshot, but is it possible to get this server running again for the next couple of weeks?

Unfortunately, this is causing a lot of issues in our production environment as we are currently unable to hotfix the issue.

Best,

Jack

On Mon, Sep 25, 2023 at 11:37 PM Nathan Kinkade <kin...@measurementlab.net> wrote:

[Nathan Kinkade]

Hi Cleiton,

Yes, gig01 was retired back in July, consistent with the last test date you specified. gig02 was retired last week. Indeed, gig03 and gig04 are scheduled to be retired later this year.

Best,

Nathan

On Mon, Oct 9, 2023 at 9:10 AM Cleiton Moya de Almeida <c...@cos.ufrj.br> wrote:

Hello Nathan,

Do you know if the M-Lab site gig01 (at Rio de Janeiro) was also retired, or is it in maintenance? It is not listed anymore in the Platform Status (https://www.measurementlab.net/status/). 

We are using the Locate Service, the last test to gig01 we got was 2023-07-07.

Also, since last Friday (2023-10-06) we are not getting tests to gig02. Do you know if it is in maintenance, or it was also retired? 

Best regards,

Cleiton

To view this discussion on the web visit https://groups.google.com/a/measurementlab.net/d/msgid/discuss/CANEqZDGNJ%3DquHuYu7W2vaKo1frAEORGNG5qV8SWrc9o1Qi22bA%40mail.gmail.com.