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

20040513: LDM and CPU load (fwd)

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: 20040513: LDM and CPU load (fwd)
  • Date: Sun, 13 Jun 2004 14:28:57 -0600
>From:  "Kevin R. Tyle" <address@hidden>
>Organization:  SUNY Albany
>Keywords:  200405131522.i4DFM3tK005922 LDM system load

Kevin,

Did you get a response to the following inquiry? I don't see one in
the LDM section of our inquiry tracking system, so I am afraid that
it slipped through the cracks somehow.

>Hi Steve,
>
>Any thoughts on what a "typical" CPU load for an LDM server
>such as ours (gusher.atmos.albany.edu) should be?  This server
>requests a subset of CONDUIT, and the full suite of UNIDATA products
>(including FNEXRAD/NNEXRAD), and serves UNIDATA downstream.
>I notice that the CPU load is typically quite high (5.5-7.5) on
>this machine.  Following a restart, the load drops off to around
>1.00 or so, but after a few hours it spikes back up, where it remains
>until another restart.  The high CPU load does not seem to impact
>the quality of data reception, but it seems "strange".  We noticed
>this around the time we began requesting all NNEXRAD products instead
>of just local ones, but I am not sure if this is related or just
>coinicidental.
>
>Any feedback or things to check would be appreciated . . .
>
>Thanks,
>Kevin

Since I am worried that you did not get a response to this inquiry,
I will offer the following:

The load average you will see on a machine running the LDM will be
a strong function of:

- the amount of data processing the machine is doing
- the number of requesting rpc.ldmds that are running (the number
  that this machine is running)
- the speed of the machine
- the size of the queue in relation to the amount of physical RAM
  the machine has
- depending on the type of machine (e.g., Solaris SPARC vs PC OS
  like Linux), the type of the file system being used

and a weaker function of:

- the number of downstream feeds the machine is running

Can you provide more information on the setup on gusher?  It would
help cut down on the number of possibilities.

Thanks, and sorry for the delayed response...

Cheers,

Tom Yoksas
--
+-----------------------------------------------------------------------------+
* Tom Yoksas                                             UCAR Unidata Program *
* (303) 497-8642 (last resort)                                  P.O. Box 3000 *
* address@hidden                                   Boulder, CO 80307 *
* Unidata WWW Service                             http://www.unidata.ucar.edu/*
+-----------------------------------------------------------------------------+