McMahon - Last Good ECMWF Mix Run was ( 2023-07-11-12 z ) and Last Should Be ( 2023-07-11-18 z ).
It is possible that the Data is running Late.
Note: this is an automated Post, check data to ensure it is correct!
Kindest Regards,
Tony
I have a suspicion about why this is missing. I saw a suggestion yesterday that the directory naming convention haa reverted to the original name for 2 of the 4 daily runs.
I’ll check when I get my laptop powered up later on. Editing Python code via SSH on a phone isn’t the easiest thing!
Hi Chris.
Sorry I am off grid mostly until I get home late Thursday night .
Not sure this will send or not.
Tony.
18z for 11th July seems to have downloaded and processed without any problems so my worry about changed URLs wasn’t relevant. I’m assuming that this was just a timing issue with the data not being ready when Tony’s scripts ran.
1 Like
Hi Chris,
With ref to
I’m assuming that this was just a timing issue with the data not being ready when Tony’s scripts ran.
Here are my Run Times, I have always used.
It will be easier to check the average data availability times once my recent mods have gathered some run completion stats.
I know historically from investigating GFS availability issues that there are many reasons why run data isn’t always available when expected, e.g.
- NOAA GFS forecast run speed is degraded due to operational issues
- NOAA delivery of model output to the publicly accessible storage is slow due to internal network issues
- NOAA public storage is running slow due to operational issues
- Access to NOAA public storage is slow due to external (Internet) issues
- Access to NOAA public storage is slow due to internal network issues on my server provider’s network
- My server is slow due to other operations taking place at the same time
An example of the last one is when, GFS data is late arriving and ECMWF data arrives a bit early both processing runs happen in parallel. There’s a lot of data being downloaded, run through complex scripts, and a few million data elements calculated and stored into the database. So doing both runs in parallel slows both down significantly.
These are all reasons why it’s not possible to guarantee that run data will always be available for specific times.
I run WxSimate and WxSim hourly for that reason. If some data is a bit late I get it in the next hourly run.
