JIRA, Wiki, Maven repo, CI are all down

Community Infrastructure
1

One of our infrastructure providers appears to be having a problem.

  • JIRA / issue tracker
  • wiki
  • all redirects (including errors, om.rs, mavenrepo)
  • CI
  • almost all test and qa environments
  • Atlas
  • Modules
  • Sonar
  • ID
2

Thanks for the update @cintiadr!

Probably not much we can do, but this is a pretty big blocker for development… since the mavenrepo redirect is down, I can’t even start up an OpenMRS instance via the SDK, no mind build anything. (Assumedly I can temporarily change it to point directly to artifactory, will try that now).

@burke @darius

Take care,

Mark

3

It looks like it may be related to this notice (from status.iu):

“A road construction crew on State Road 37 accidentally cut a network cable, causing a network outage that affected a number of services and locations.”

4

Thanks @burke

For what it’s worth, I just posted another thread about running the SDK in offline mode for those that are trying to use it this morning… :slight_smile:

Take care, Mark

1 Like
5

I hear you, I really do. I will see if I can get a failover system or something for maven repo (this one is doable now).

1 Like
6

@cintiadr, I see our DNS provider has the HTTP Redirection Records feature. Could we use that for mavenrepo?

I guess the answer is no, since it allows us to redirect a domain to a URL, whereas we need to redirect a url to a url.

1 Like
7

Thanks guys!

@cintiadr just to be clear, I don’t want you to take my comments as directed at the infrastructure team directly. You guys already go above and beyond on a volunteer basis. I did want to make sure the issue was highlighted to the broader community.

It occurred to me soon after I posted that the workaround was just to run maven in offline mode (duh). (I posted that in a separate post OpenMRS SDK Offline Mode)

It does seem unfortunate that although we moved away from our own hosted repo, we still are at the mercy of our provider because it provides the DNS (though I agree with the decision to keep our own URL to make us platform-independent if we need to switch in the future).

Yeah, was thinking of some sort of redirect like @raff mentioning but figured it might not be possible for the same reasons.

Honestly, with the offline Maven option, we are probably okay as-is. (Though certainly if there was an easy workaround we should implement it).

Take care, Mark

8

@cintiadr could this be the reason for this failure? https://ci.openmrs.org/browse/DEMO-RD-85/log

9

Yes, that :slight_smile: But as jfrog is hosted, I was wondering if we could have a more robust/resilient redirect (I am going to explore Api Management - Amazon API Gateway - AWS for example to see what I can do). Maven repo is probably the service that affects the community the most (after DNS!) when it goes down. I mean, we can survive if JIRA/Confluence/CI is down for 2-3h, but maven repo is quite of a big pain.

Of course, @mogoodrich :slight_smile: I didn’t read it on any other way.

11

It was a broken DNS configuration. It should be fixed now :slight_smile: