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

20020726: info on ldm and RAM file systems

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: 20020726: info on ldm and RAM file systems
  • Date: Fri, 26 Jul 2002 15:41:29 -0600
>From: Darien Davis <address@hidden>
>Organization: NOAA/FSL
>Keywords: 200207261625.g6QGPQ925025 LDM

Darien,

>Hi support person (maybe Anne Wilson! HI!)

Anne is on vacation right at the moment, but she will be back in time
for the workshop she is putting on starting next Thursday.

>We were wondering if every product was written to disk
>when it is put into the ldm queue.

Yes.

>Since the queue is a file, we were
>wondering if every product routed thru the ldm daemon is
>written to disk and then "actioned" (if there is a product flow diagram
>that you could refer me to, that would be great).

The queue is a memory-mapped file, but yes each product will exist
in the queue after being received.

>If a file is written, has anyone experimented with using a RAM file
>system?

Not that I know of.

>We were going to try this (since we have a product that may
>produce upwards of 30,000 points per minute which may translate to
>60 files per second for data with a burst nature) to see if it
>was feasible. I thought I would ask if anyone else had done this
>and had adverse problems with it. The RAM file systems are so
>clean since you can use the standard I/O calls.

This sounds like an interesting experiment.  Please let us know what
you find out!

>Thanks for any info -- I'll post to ldm users group next!

OK.

Tom Yoksas