NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
Don On 1/13/16 9:02 AM, Brendon Hoch wrote:
Hi Don, IDV 5.2 seems to be robust enough under Linux to allow for cronjobs & ISL to produce products with Xvfb on a reliable basis. We're running on CentOS 7.2.1511; the server also has other jobs plus a full LDM suite. We have about 8 IDV cronjobs running at various intervals, ranging from every 10 minutes for a local radar product to hourly for satellite imagery on a 24/7 basis. Each job runs IDV with a ISL script. I've enclosed one example bundle, ISL script and run script of a NE current temperature map. The one thing we've noticed is that once in a while (once a week to once every few weeks), an IDV job gets hung up and we see a high number of kernel context switches. To mitigate this, we use the timeout utility embedded in our runIDV script to kill any IDV job that lasts more than 10 minutes: /bin/timeout 10m ${java} -Xmx${idv_memory}m -XX:+DisableExplicitGC -Didv.enableStereo=false -jar ${dirname}/idv.jar $* We're not sure which script produces the hangup or why, but it's infrequent enough that I'm willing to live with it. It might be a situation where IDV is looking for data on a server and can't find it or there's a network connectivity issue. It would be great if there was a way for IDV to exit cleanly in these unattended ISL situations. Overall, I am very satisfied with the reliability, so much so that I now see this as a viable alternative to replacing many of our legacy WXP & Gempak scripts. Hope this helps, Brendon
-- Don Murray NOAA/ESRL/PSD and CU-CIRES 303-497-3596 http://www.esrl.noaa.gov/psd/people/don.murray/
idvusers
archives: