[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[LDM #BOA-542231]: Problem writing 0.5 deg GFS to disk
- Subject: [LDM #BOA-542231]: Problem writing 0.5 deg GFS to disk
- Date: Mon, 07 Dec 2009 11:44:03 -0700
Hi Gerry,
re:
> Quick note: I changed the ldmd.conf request that triggers this to get
> CONDUIT and a product including pgrb2f. We're also pretty specifically
> requesting RUC2. This gets the 2 files in for processing we need. I'm
> still decomposing the individual gribs, but they should also be
> available on the dc-ldm2 ldmd logs...
I am not sure if this means that I should be looking at something...
A quick comment from me:
I notice that you have a couple 'pqact' exec actions in the ~ldm/etc/ldmd.conf
file on dc-ldm2 that specify the use of the same pattern-action file for
different
feeds. I recommend that you separate actions pertinent to a specific feed/set
of feeds into its own pattern action file.
For example, I see the following:
exec "pqact -f IDS|DDPLUS etc/pqact.conf"
exec "pqact -f FSL2|FSL3|FSL4 etc/pqact.conf"
These lines will cause two invocations of 'pqact' and _both_ will be forced to
consider actions not specific to the feeds they are told to act on:
- the first will have to compare each product from the IDS|DDPLUS feed against
every FSL2|FSL3|FSL4 action in ~ldm/etc/pqact.conf
- the second will have to compare every product from the FSL2|FSL3|FSL4 feeds
against every IDS|DDPLUS action in ~ldm/etc/pqact.con
For efficiency sake, I suggest creating a separate pattern-action file
with actions for just the IDS|DDPLUS and a second one with actions just
for the FSL2|FSL3|FSL4 feeds.
The same comment applies to the two currently active 'pqact' exec lines
for CONDUIT:
exec "pqact -f CONDUIT -p ruc2......pgrb20f etc/pqact.conduit.conf"
exec "pqact -vf CONDUIT -p gfs etc/pqact.conduit.conf"
Another question comes to mind from the observation that you are requesting
CONDUIT with pattern 'gfs' from three different
machines:
- are you sure that each of the machines is ingesting the same set of CONDUIT
data?
If they are not, then you could miss products when your rpc.ldmd connection
changes from secondary to primary or visa versa.
Cheers,
Tom
--
****************************************************************************
Unidata User Support UCAR Unidata Program
(303) 497-8642 P.O. Box 3000
address@hidden Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage http://www.unidata.ucar.edu
****************************************************************************
Ticket Details
===================
Ticket ID: BOA-542231
Department: Support LDM
Priority: Normal
Status: Open