Alternative schemes could result in chaos, such as when many sites try to feed from one site or when many lower tier sites have to change their LDM configurations. When a data outage occurs, configuration changes are minimized if only the top tier sites failover to alternate sources, so that the lower level tiers of the IDD topology can avoid reconfigurations.
A data outage can occur at any level in the IDD topology, so the first action should be to contact your upstream site to determine their status. If the upstream site is getting data, the problem is local between you and the upstream site; otherwise the upstream site should contact it's upstream site. This chain of events will usually lead to a quicker resolution of the problem than contacting support at Unidata to solve the problem, because support would have to contact the upstream sites as well. However, please CC: support@unidata.ucar.edu in email to the upstream contacts so support will be aware of the data outage and be able to work on a possible secondary solution.
The IDD is a community with a shared goal: receiving meteorological data in a reliable and timely manner. The cooperation of all the IDD community members is needed to achieve this goal.
It's probably a good idea to make a hard copy of this page in case the Unidata web server is unreachable. There are currently four NOAAport source machines for the IDD:
The duplication of the NOAAport streams makes a source machine robust because this single point of failure is eliminated. Also, if the network connection to a source machine fails, then the top tier sites can failover to another source site. The source ring concept has proven practical in covering data outages, sometimes without other nodes on the IDD even noticing that there was a problem. The goal is to have at least four source sites geographically distributed to cover outages due to signal interference, ISP outages, etc.
Sites that are currently feeding from motherlode or thelma should failover to one of these sites in the following order:
The top tier sites that are currently feeding from frost.wunderground.com should failover to one of these sites in the following order:
The top tier sites that are currently feeding from sunshine.ssec.wisc.edu should failover to one of these sites in the following order:
The first action should be to contact your upstream site to determine their
status. You can use the notifyme
program to easily determine whether your upstream site is receiving data of
a particular feedtype. Often this leads to quicker resolution to the data outage
problem and it helps to determine the extent of the problem. If the problem
is with the immediate upstream site, then try your failover site. All IDD sites
should have a failover site. If your site doesn't have a failover, then send
a message to:
support@unidata.ucar.edu
to obtain one. If your failover is unreachable then it's possible to request
a temporary site by sending a message to support. If support is not available
then make a request to the
e-mail list "ldm-users". The IDD community has been very helpful in providing
ad-hoc failover sites when a site has no alternatives. These reconfigurations
have greatly increased the reliability for all sites in the IDD topology.