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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
>From: Mark Tucker <address@hidden> >Organization: Lyndon State >Keywords: 199909271950.NAA15960 McIDAS ADDE performance Mark, re: >> This is starting to sound like a Linux configuration problem, not a >> McIDAS mcserv problem. Is there a process limit on your machine that >> could be causing this? >It seems that inetd may be the problem. The default installation of Linux >(both Debian and RedHat) is to allow 40 connections per minute. I edited >the inetd.conf on our server to allow 255 connections/min. and that >appears to have made a significant difference in the repsonsiveness of >mcserv. I'm sure that this value of 255 is not optimal and will probably >need to be adjusted in the future. 40 connections per minute is very low. In looking at inetd.conf on our RedHat 6.0 system, I saw an entry for 'swat' that specified 400 connections per minute. I think that this might be a good place to start instead of 255. >I haven't yet had a lab full of >students (15-20) hit mcserv at once since I made this change but it is >already noticable at lower loads. Excellent. >Also, ther haven't been any >inetd/mcserv failures in the log files since this change either. Also a very good indication that this was the problem. >You may >want to consider documenting this setting for other sites who may be >serving several clients simultaneously. Here's what I added to >inetd.conf: > >mcserv stream tcp nowait.255 mcadde >/usr/local/mcidas/bin/mcservsh mcservsh -H /usr/local/mcidas >mccompress stream tcp nowait.255 mcadde >/usr/local/mcidas/bin/mcservsh mcservsh -H /usr/local/mcidas I am updating my Notes and Warnings section for Linux users right now. >Thanks for your help on this. Thanks for the update. Tom