[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[TIGGE #BGR-338705]: A problem when perform LDM tests

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: [TIGGE #BGR-338705]: A problem when perform LDM tests
  • Date: Thu, 24 May 2007 18:30:50 -0600
Manuel,
 
> This morning, we have made some changes to the configuration of our LDM.
> Basically:
> - increased product queue size from 4 GB to 12 GB (RAM of the machine is
> 16GB)
> - changed the number of slots in product queue from "default" to the
> following formula:
> $pq_slots = $pq_size;
> $pq_slots =~ s/G//;
> $pq_slots *= 1000*1000*1000;
> $pq_slots /= 430000;
> $pq_slots = int($pq_slots);
> 
> ie, we take into the account the size of each product (we take 1 field
> from ECMWF as example, 430 KB) and we divide the size of the product
> queue by the size of each product.
> 
> This is now the output of pqmon:
> May 21 17:46:22 pqmon NOTE: Starting Up (2728)
> May 21 17:46:22 pqmon NOTE: nprods nfree  nempty      nbytes  maxprods  
> maxfree  minempty  maxext   age
> May 21 17:46:22 pqmon NOTE:  27424   481       1 11952616232     27905  1347  
>        0     22494536 2020
> May 21 17:46:22 pqmon NOTE: Exiting

The above indicates that the product-queue is sized almost perfectly:
the number of bytes used is almost 12 GB and almost all the slots are
used.  Congratulations!

I would create a time-series of the "nbytes" and "nprods" parameters
to ensure that this condition continues to hold.

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: BGR-338705
Department: Support IDD TIGGE
Priority: High
Status: Closed