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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
Hi Leo, re: > If you do not mind, I have been given two more questions about clustering. OK, and I see that you asked some other questions yesterday that we have not responded to. I will attempt to address them in this reply... re: > First, I did mean Level II and Level III data and products respectively. :-) Its those darn keyboards :-) re: > What has been proposed here is to run two servers active/active with an LDM > for Level II and one for Level III. If the Level II LDM server goes down, > it would failover to the Level III LDM server putting two LDMs on the one > server. OK. This is not a typical setup that cries out for a clustered solution, but there are many ways of attacking any problem. re: > So when you said you use two ldmd.conf REQUESTs, is that using two LDM > servers or one LDM and is it on the same server. I was referring to using a single LDM running on one server. re: > Would it be preferable to > put two LDM's on a single server and have the second one just be a standby? I don't see how running two LDMs on a single machine provides the redundancy that I think you are trying to achieve... if the machine has a problem, both LDMs will be affected, etc. re: > And I do still think that a call to discuss LDM failover would be very > beneficial here. We would like to have it setup within the next couple of > days due to our tight schedule. How about Friday morning sometime. We have a long meeting this afternoon, and I will be out of the office all day tomorrow. Right now, Friday morning up until about noon is open for me, but another user is trying to setup a conference call to discuss use of our visualization packages in his research, and the target for that call would also be on Friday morning. Below are responses to questions you posed yesterday which we were going to answer today: re: > I do have a couple of additional questions asked to me this afternoon > that deal with clustering. I hope I pose them so that they are > understandable. > > 1) Can LDM listen on a specific port:IP combination? Yes. The LDM can be built to listen on a specific port. The default is to use port 388 which, by the way, is registered with IANA for use by the LDM. Also, if the machine has multiple IPs, the LDM can be configured to listen on the specific IP. We use this feature ourselves on machines where we are running multiple LDMs (in different accounts) simultaneously. re: > 2) Who initiates the initial connection - NWS LDM on a push or AFWA > LDM on a pull? The downstream initiates the connection by issuing one or more REQUESTs to one or more upstreams. The lead LDM process at the upstream checks to see if the machine making the REQUEST(s) is ALLOWed, and, if it is, the connection sense is turned around and the upstream will send the data REQUESTed to the downstream which is then waiting for data. re: > Thanks again for your assistance. No worries. Cheers, Tom -- **************************************************************************** Unidata User Support UCAR Unidata Program (303) 497-8642 P.O. Box 3000 address@hidden Boulder, CO 80307 ---------------------------------------------------------------------------- Unidata HomePage http://www.unidata.ucar.edu **************************************************************************** Ticket Details =================== Ticket ID: MBG-654954 Department: Support LDM Priority: Normal Status: Closed