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

[Support #HIT-987436]: pqmon output is not what I expect



Hi Gregg,

> I'm running LDM 6.6.5 on RHEL with the following feeds into the the LDM:
> 
> -NOAAPORT ingest (local)
> -UNISYS single site and mosaic imager (local satellite ingest)
> -equivalent of a FOS feed from the NWSTG/TOC using a TCP/IP socket
> connection to the TOC.
> 
> I'm running 7 pqacts, different configuration files, and will eventually
> add another one to handle BUFR decoding (profiler).
> 
> EXEC    "pqact"
> EXEC    "pqact
> /nfsbak/ops_users/ldmbakng/etc/pqact.conf_part_1_Satellite"
> EXEC    "pqact
> /nfsbak/ops_users/ldmbakng/etc/pqact.conf_part_2_Decoders_NOT_radar"
> EXEC    "pqact
> /nfsbak/ops_users/ldmbakng/etc/pqact.conf_part_3_Decoders_Radar_NON_level_2"
> EXEC    "pqact
> /nfsbak/ops_users/ldmbakng/etc/pqact.conf_part_4_Decoders_Radar_level_2"
> EXEC    "pqact /nfsbak/ops_users/ldmbakng/etc/pqact.conf_part_6_Other"
> EXEC    "pqact /nfsbak/ops_users/ldmbakng/etc/pqact.conf_part_7_Nwx"
> 
> 
> The system seems to be fine except I receive continuous ldmd.log entries
> similar to:
> 
> 
> Oct 06 14:55:49 rhesrv22 pqact[18068] WARN: Processed oldest product in
> queue: 0.002214 s
> 
> 
> So I'm running pqmon and see results like the following:
> 
> 
> Oct 06 14:55:35 pqmon NOTE:    135     1  244004      870232     11736        
> 9    232403 999131304 6
> Oct 06 14:55:40 pqmon NOTE:     51     1  244088      504344     11736        
> 9    232403 999497192 1
> Oct 06 14:55:45 pqmon NOTE:     60     1  244079     2679216     11736        
> 9    232403 997322320 3
> Oct 06 14:55:50 pqmon NOTE:     67     1  244072      305032     11736        
> 9    232403 999696504 1
> Oct 06 14:55:55 pqmon NOTE:    271     1  243868     4705904     11736        
> 9    232403 995295632 2
> Oct 06 14:56:00 pqmon NOTE:     63     1  244076       17384     11736        
> 9    232403 999984152 0
> Oct 06 14:56:05 pqmon NOTE:    176     2  243962     1073728     11736        
> 9    232403 998927584 3
> Oct 06 14:56:10 pqmon NOTE:    139     1  244000      540856     11736        
> 9    232403 999460680 3
> Oct 06 14:56:15 pqmon NOTE:     54     1  244085     5090296     11736        
> 9    232403 994911240 1
> Oct 06 14:56:20 pqmon NOTE:    137     2  244001     1095744     11736        
> 9    232403 994652824 4
> Oct 06 14:56:25 pqmon NOTE:     68     1  244071      742832     11736        
> 9    232403 999258704 1
> Oct 06 14:56:30 pqmon NOTE:     85     1  244054      351480     11736        
> 9    232403 999650056 2
> Oct 06 14:56:35 pqmon NOTE:     51     1  244088      607512     11736        
> 9    232403 999394024 0
> 
> The LDM queue is set at 1.5 GB.

The "Processed oldest product" messages are consistent with the low values for 
the age of the oldest product in the queue as output by pqmon(1). They're a 
little hard to believe, however, given that your queue is 1.5 GB -- unless 
products are coming-in very fast and furiously.

Can you do some back-of-the-envelope calculations to verify or refute the low 
age values?

> I didn't think I would be obtaining these warnings (ldmd.log) and low
> values of "age of oldest product" (pqmon) as I am with such a large
> queue and having 7 pqacts running.

The number of pact(1) processes running is relatively irrelevant.

> Any suggestions or is there a need to be concerned?  I'm thinking
> something needs to be tweaked.

You could try increasing the size of the queue to see if that makes a 
difference.

You could also upgrade to the latest version of the LDM. Yours is over three 
years old.

> Thanks,
> Gregg

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: HIT-987436
Department: Support LDM
Priority: Normal
Status: Closed