[LDM #QHD-410218]: Redundancy Issues with products

Joel,

> Thanks, for the quick response...  I am a little surprised at the idea of the
> data products being purged quickly from the product queue.

Use the pqmon(1) utility to determine how long data-products reside in the 
product-queue.

> Now as far as
> latency goes, I would estimate that we will receive the same products from
> the multiple sources within a 15 minute period.  Maybe even a few minutes.
> The print exec commands would be run multiple times (again on receipt of
> multiple data feeds).  How would one debug pqact executions?  How would I
> find which statement is firing what?  Is there a log or should I create an
> identifiable naming scheme associated with receipt from the various sources?

You can put a pqact(1) process into verbose logging, in which it will log every 
product that it processes, by sending the process a SIGUSR2. Another SIGUSR2 
will put the process into debug logging mode, and another SIGUSR2 will return 
the process to its original mode.

> Our queue is set to 2G, and I just can't see where the product-queue product
> would be purged prior to the other similar products arriving?

You can do the same logging trick with downstream LDM processes. Alternatively, 
you can restart the LDM system (but not the pqact(1) processes) in verbose 
logging mode via the command "ldmadmin restart -v".

You might be interested in the LDM workshop that will be held July 14th and 
15th. See <http://www.unidata.ucar.edu/blogs/news/entry/training_workshop_2011> 
for more information.

> Joel Asunskis, P.E.
> Hydraulic Engineer
> 
> US Army Engineer District, St. Louis
> Engineering & Construction Division
> Hydrologic & Hydraulics Branch
> 
> WATER DATA MANAGEMENT SECTION
> (CEMVS-EC-HM)
> 
> 1222 Spruce St.
> St. Louis, MO 63103
> 
> (314) 331-8338
> (314) 331-8342
> BB: (314) 580-4147

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: QHD-410218
Department: Support LDM
Priority: Normal
Status: Closed