@jeapostrophe ping re pkgd

@lexi.lambda Just talked to @jeapostrophe, he plans to fix soon

thanks!

@lexi.lambda it’s up

@mflatt ok, thanks for investigating. Is there a best way to generate that error myself so that I can see what the error behavior is from the client side?

You could modify the server to close the connection (or even just exit) before returning ’ok. I guess you could also have a slow checker and disconnect your client machine from the network before the success message comes back.

Ok, thanks, I’ll try that

Just noticed that the snapshots page: https://pre.racket-lang.org/installers/ lists northwestern but this doesn’t seem to work anymore. i.e. the link http://plt.eecs.northwestern.edu/snapshots/ times out.

@greg ^^^

@pocmatos: I believe there’s some network shuffling happening today at NWU.

@robby: ^

I’ve created a #drracket channel for anyone interested in working on DrRacket.

The site seems to be back now

@robby For Travis CI builds that need Racket snapshot, is Northwestern or Utah preferable (meaning, better uptime, as well as friendly to your resources)? (I ask because I seem to recall switching this in years past, but was advised to switch it back. So I’m not sure whether to accept @pocmatos’ PR to switch it from NW to Utah.)

I don’t ever intend for it to be down and there are no resource constraints from our end

@robby Cool I’ll stick with Northwestern then — thanks!

Sorry it has been down recently.

Does it make sense to send requests to both and take the one that comes back first?

No worries, I wouldn’t expect any one server to have 100% uptime.

> Does it make sense to send requests to both and take the one that comes back first? Maybe. I guess the question is where to put that. e.g. Are you proposing that folks do this try-both logic in various shell scripts etc.? Or are you proposing to put both behind one load balancer that does this, and to which we point? :grin:

did somebody say load balancing? :grin: