Zezo forum:

Post by zezo | 2020-09-09 | 16:41:23

NOAA has announced a maintenance window for the data servers. Weather may be out of sync tonight.

Today, Sept 9, between 17Z and 19Z, there will be network maintenance in our Boulder data center that will most likely take down the para NOMADS website for those hours. It should be restored after that, and data will be backfilled.

15 Comments

Post by zezo | 2020-09-09 | 16:42:24

Another announcement just arrived:

We are aware that there are problems accessing many of our websites and services. We are working on the problem now but have no ETR at this time.

We will send updates as we have them.

Post by GeGaX | 2020-09-09 | 18:37:26

Hi Cvetan 😉,
Thanks for these infos
Where do you get this ? (URL ?)
Do you receive them by email? If yes where can we register to get them also?
Thanks a lot 😉

Post by zezo | 2020-09-09 | 20:20:16

Search for ncep.list.nomads-ftpprd@lstsrv.ncep.noaa.gov

Post by GeGaX | 2020-09-09 | 20:29:24

Great ! Thanks 😉

Post by Luis75 | 2020-09-09 | 22:35:27

Now it looks fine.

Post by M63 | 2020-09-10 | 11:25:52

Noaa HS again for 06z (2020-09-10).

Post by zezo | 2020-09-10 | 12:09:39

Yep.

... gfs.t06z.pgrb2.1p00.f009.idx ...500 read timeout

Post by BooBill | 2020-09-10 | 14:06:41

Cvetan, is the same issue going to be reflected in the game predicted path line? ie) more than 12 hours out are we looking at the previous forecast?

Post by zezo | 2020-09-10 | 15:10:50

Yes, that would affect long-time routing.

BTW the game uses 0z data right now for all forecasts.

And NOAA is still unresponsive.

Post by BooBill | 2020-09-10 | 16:33:49

How can we tell what the game is using?

So, am I correct in concluding we'd do better using the game predicted path than using ECMWF?

If this problem persist till the start of the Figaro 3, it's going to be really interesting.

Post by zezo | 2020-09-10 | 16:58:27

The game server sends the GFS models in use, but the file is encrypted, so not directly available.

The data will get in sync sooner or later, and then there will be an abrupt change of the situation. So you may want to route towards a place when you would really like to be in a day or two.

Post by BooBill | 2020-09-10 | 21:20:00

Is the time stamp of the weather file the game is using anywhere that we can see it? Is that something that is in the raw log on Dashboard. I've turned that on and the NMEA output just for good measure. Don't tell me the solution is in that forum thread please ;)

Post by zezo | 2020-09-10 | 21:42:25

Not easily visible. Won't say more in public. Drop me a line on email for some details.

Post by zezo | 2020-09-11 | 08:50:59

This is (a bit awkward) way to see the grib in use.

The game downloads the forecast in the form of .wnd files. Those can be monitored in the network panel of the browser Developer Tools.

The files are cached, so you have to reload the game before proceeding. Open the dev tols, enter a race and you should see the following:

Click 006.wnd, 009.wnd etc. and check the full file path. It contains the GFS model in use:

https://static.virtualregatta.com/winds/live/20200911/00/006.wnd

in the example above is GFS 2020-09-11 00Z, offset 6.

Post by BooBill | 2020-09-11 | 15:50:49

Looks like we got a full download at 00Z. The 06Z update seemed to stop at +120h. 12Z so far looking like a success.

Expected winds in zezo have matched Dashboard, although I did notice game client was off by a full knot sometimes.

I had been following a line to a turning point I'd picked in Windy.com using ECMWF that kept me above the windline with an eye on the game predicted line for the next 6 hours and it seems to have worked like a charm. Just happy this didn't happen as I was rounding the horn or crossing a tricky transition.

Thanks for the help Cvetan. Definitely have some resources to look at when we experience odd download or sync problems in the future.