HYCOM forecasts - DAP server error

[James Turner]

 

Hi,

 

I am writing to report a server error when accessing the latest forecasts on expt_91.2:

I get a DAP server error when accessing the forecast “http://tds.hycom.org/thredds/dodsC/GLBu0.08/expt_91.2/forecasts/hycom_glb_912_2018011900_t168_uv3z.nc”.

The hindcast URL access is fine.  Please see screenshot:

 

 

 

I am requesting the data at 09:47 2018-01-20 UTC.

 

Best regards,

 

Jim

 

 

---

James Ferguson Turner BSc PhD MIMarEST

Lead Consultant

j...@jtgeo.co.uk

+44 (0) 7500 118988

 

 

[Michael McDonald]

Jim,

Sorry. python cmd line queries are not good about returning the actual error response from the http server. If you made a query against this file in a browser it returns "Stale File Handle" for 1/2 the requests. We had two of our four THREDDS servers with stale file handles for this frequently updated forecasts folder, so a simple tomcat restart was all that was needed (we cron this restart task daily between 1:00->1:30PM EST to clear these data update errors).

All servers in the pool are now responding properly. 

>>> import netCDF4

>>> nc = netCDF4.Dataset("http://tds.hycom.org/thredds/dodsC/GLBu0.08/expt_91.2/forecasts/hycom_glb_912_2018011900_t168_uv3z.nc")

>>>

 

I am writing to report a server error when accessing the latest forecasts on expt_91.2:

I get a DAP server error when accessing the forecast “http://tds.hycom.org/thredds/dodsC/GLBu0.08/expt_91.2/forecasts/hycom_glb_912_2018011900_t168_uv3z.nc”.

The hindcast URL access is fine.  Please see screenshot:

   

I am requesting the data at 09:47 2018-01-20 UTC.

[James Turner]

I am happy to confirm the error has gone. Thank you.

 

Jim

[James Turner]

Hi Michael,

 

I’m accessing the server at 0949 UTC and it appears the stale file handle error has returned:

 

 

Best regards,

 

Jim

[Michael McDonald]

Jim,

Thanks for the follow-up. I think we've tracked down this issue now. Our data extraction cron for this dataset was accidentally "duplicated" on two different servers. They were extracting/updating at exactly the same time causing this NFS file locking conflict. It would take several tries for one side to "win" and delete the incoming data file. Please let us know if tomorrow's data extraction does not work as expected.

https://www.dropbox.com/s/faztdien4h919gj/Screenshot%202018-01-23%2018.48.18.png?dl=0

 

I’m accessing the server at 0949 UTC and it appears the stale file handle error has returned:

 

--

Michael McDonald
HYCOM.org Administrator

[James Turner]

All good now – thank you Michael.

 

Best regards,

 

Jim

 

 

From: e.m.mc...@gmail.com [mailto:e.m.mc...@gmail.com] On Behalf Of Michael McDonald
Sent: 23 January 2018 23:54
To: James Turner <j...@jtgeo.co.uk>
Cc: fo...@hycom.org; Warren Gray (warre...@seisintel.com) <warre...@seisintel.com>
Subject: Re: HYCOM forecasts - DAP server error

 

Jim,