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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
Dan, The behavior you describe is the behavior expected when the process that's inserting the data products into the LDM product-queue is not in the process group of the LDM server. When there are no more products to send, an upstream LDM will sleep for 30 seconds or until it receives a CONTINUE signal (SIGCONT), whichever comes first. When it inserts a data-product into the queue, the product-queue library sends a SIGCONT to its process group. Your insertion process is outside the LDM process group, so its SIGCONT isn't being received by any upstream LDM. The solution is to ensure reception of the SIGCONT. This can be done by either executing the insertion process in the process group of the LDM via an EXEC entry in the LDM configuration file, or by having the insertion process explicitly send a SIGCONT to the LDM process group. The former is probably easier. > Full Name: Dan Fredette > Email Address: address@hidden > Organization: WSI > Package Version: ldm-6.11.2 > Operating System: > Hardware: > Description of problem: > > Hello, > > I have a questions about the LDM and data latencies. Is it typical that > when a file is added to a LDM queue on one server, that another server can > expect to get the data 1-30 seconds later? Is that a normal timeframe? > > The way I have it setup, is I have LDM running on two servers on the > same network. I am just posting grib data thats about 3MB in size. > The other server is just listening for this data and no other data. > I see the data posted to the local LDM queue immediately, but the LDM > server listening for this data, sometimes can get it within a second, > sometimes 20-30 seconds later. > > Is there any way to consistently get this to under 5 seconds, or is this > normal behavior and there is not much else we can do with it? > > Thanks. > > Regards, Steve Emmerson Ticket Details =================== Ticket ID: UUZ-520900 Department: Support LDM Priority: Normal Status: Closed