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

[LDM #IGM-878445]: Frequent ldmadmin check emails



Hi Julie,

It would appear that your product-queue is *way* too small relative to your
maximum-latency parameter. The size of the product-queue is determined by the
registry parameters "/queue/size" and "/queue/slots" and the maximum-latency
by the parameter "/server/max-latency". In general, the size of the queue
should be large enough so that the minimum residence time of a data-product
is greater than the maximum-latency parameter. We like to run our queues with
a maximum-latency of about one hour; consequently, we have relatively large
queues (75 MB). The reason for having the queue larger than the maximum-latency
is so that you don't miss any data-products.

If you have gnuplot(1) installed, then the "ldmadmin addmetrics" and
"ldmadmin plotmetrics" commands can help you determine values for the size
parameters. See
<http://www.unidata.ucar.edu/software/ldm/ldm-current/basics/configuring.html#cron>
and search for "addmetrics". Execute the command "ldmadmin printmetrics" to see
if it will work.

Once you in the ballpark, you can set the registry parameter
"/reconciliation-mode" to either "/increase queue" or
"decrease maximum latency" to have "ldmadmin check" automatically adjust
the corresponding parameter(s).

Alternatively, you could stop executing the "ldmadmin check" command (not
recommended) or modify the crontab(1) entry so that it ignores an exit
status of 5, which indicates a bad queue size.

> I have "ldmadmin check" running every 15 minutes in crontab.  I have been
> getting frequent (every 15 minutes) emails with vetQueueSize() messages
> (see partial messages below).  They tell me what the queue size should be
> with the proper number of slots and offer an alternative, setting the
> maximum-latency parameter to a certain number of seconds.
> 
> I am often unable to change the values right away, but as soon as I can, I
> change the queue size and slots in the registry and restart the ldm, but a
> bit later, I receive more messages.
> 
> How can I determine more of an appropriate queue size and number of slots
> so that I'm not told that it should be different?
> 
> Thank you for your help in advance.
> 
> Julie
> 
> 
> 
> The queue should be 31734954370 bytes in size with
> 1353 slots or the maximum-latency parameter should be decreased to 22
> seconds.
> 
> The queue should be 7000531720 bytes in size with
> 323 slots or the maximum-latency parameter should be decreased to 120
> seconds.
> 
> The queue should be 30283603502 bytes in size with
> 1460 slots or the maximum-latency parameter should be decreased to 24
> seconds
> 
> The queue should be 29839864279 bytes in size with
> 949 slots or the maximum-latency parameter should be decreased to 24
> seconds
> 
> The queue should be 30283603502 bytes in size with
> 1460 slots or the maximum-latency parameter should be decreased to 24
> seconds

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: IGM-878445
Department: Support LDM
Priority: Normal
Status: Closed
===================
NOTE: All email exchanges with Unidata User Support are recorded in the Unidata 
inquiry tracking system and then made publicly available through the web.  If 
you do not want to have your interactions made available in this way, you must 
let us know in each email you send to us.