[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[GEMPAK #MEU-566855]: nagrib2 segmentation fault error



> <FONT face="Default Sans Serif,Verdana,Arial,Helvetica,sans-serif" 
> size=2><DIV>Hi Steve..</DIV><DIV>&nbsp;</DIV><DIV>Thanks for the response and 
> for the setflag clarification.</DIV><DIV>&nbsp;</DIV><DIV>Nagrib2 or 
> nagrib1&nbsp;literally crashes with a "segmentation&nbsp; fault error" as it 
> is running - there is no other error message produced and at times doenst 
> even leave a gplt job running in the 
> background.</DIV><DIV>&nbsp;</DIV><DIV>The crash does not occur consistently 
> either - in otherwords - nagribing the same grib file sometimes works (no 
> crash) and sometimes does not.&nbsp;&nbsp;&nbsp; I've seen this happen when 
> multiple gempak progs are running from the same directory (and overwriting 
> the .nts files) but I made sure this wasnt happening (by having this be the 
> only program running).</DIV><DIV>&nbsp;</DIV><DIV>Also Ive seen this happen 
> when I have not set my env var DISPLAY to :0.0 in scripts launching gempak 
> programs invoking gplt processes.&nbsp; However - this wa
 s happening to me running nagrib from the command line interface as well as 
running it from a script.</DIV><DIV>&nbsp;</DIV><DIV>PErhaps unrelated or maybe 
not - I noticed the same happens with gpmap - however in this case - when I 
delete the .nts file prior to launching gpmap I get no problems.&nbsp; But once 
that .nts file is written it loses its mind with a "segmentation fault 
error".</DIV><DIV>&nbsp;</DIV><DIV>Pete<BR></DIV><FONT 
color=#990099>-----"Unidata GEMPAK Support" &lt;address@hidden&gt; wrote: 
-----<BR><BR></FONT>&gt;To:&nbsp;address@hidden<BR>&gt;From:&nbsp;"Unidata&nbsp;GEMPAK&nbsp;Support"&nbsp;&lt;address@hidden&gt;<BR>&gt;Date:&nbsp;12/10/2007&nbsp;05:05PM<BR>&gt;cc:&nbsp;address@hidden<BR>&gt;Subject:&nbsp;[GEMPAK&nbsp;#MEU-566855]:&nbsp;nagrib2&nbsp;segmentation&nbsp;fault&nbsp;error<BR>&gt;<BR>&gt;<BR>&gt;Pete,<BR>&gt;<BR>&gt;The
 USE_GFORTRAN=1 flag is only set for compiling the package by
 
<BR>&gt;telling<BR>&gt;which&nbsp;Makinc.linux64_xxxxxx&nbsp;file&nbsp;to&nbsp;use.<BR>&gt;<BR>&gt;This&nbsp;flag&nbsp;has&nbsp;no&nbsp;effect&nbsp;for&nbsp;the&nbsp;runtime&nbsp;environment.<BR>&gt;<BR>&gt;The
 na programs can be linked directly to the gplt, but it 
does<BR>&gt;produce<BR>&gt;a&nbsp;much&nbsp;larger&nbsp;executable,&nbsp;so&nbsp;I&nbsp;don't&nbsp;do&nbsp;it,&nbsp;but&nbsp;it&nbsp;is&nbsp;possible.<BR>&gt;<BR>&gt;It
 would help to know at what point you see a segmentation fault, 
eg<BR>&gt;in&nbsp;launching,&nbsp;or<BR>&gt;somewhere within running the 
program, and what your input source is.<BR>&gt;The&nbsp;problem<BR>&gt;you are 
describing could be a Tab" character within a 
FORMAT<BR>&gt;statement&nbsp;within&nbsp;the<BR>&gt;code that gfortran doesn't 
handle and I didn't catch in the<BR>&gt;distribution.<BR>&gt;Can you provide 
the text output of your session and show me 
the<BR>&gt;error?<BR>&gt;<BR>&gt;Thanks,<BR>&gt;<BR>&gt;Steve&nbsp;Chiswell<BR>&gt;Unidata&nb
 
sp;User&nbsp;Support<BR>&gt;<BR>&gt;<BR>&gt;&gt;&nbsp;Hi..<BR>&gt;&gt;&nbsp;<BR>&gt;&gt;
 I upgraded directly from GEMPAK 5.10.2 to GEMPAK 5.10.4 
(using<BR>&gt;RHEL5)&nbsp;and<BR>&gt;&gt; am noticing many Segmentation Fault 
messages when I use nagrib2 and<BR>&gt;nagrib<BR>&gt;&gt; - didnt see any 
threads or messages regarding this on the 
gembud<BR>&gt;email<BR>&gt;&gt;&nbsp;archive.<BR>&gt;&gt;&nbsp;<BR>&gt;&gt;&nbsp;I&nbsp;downloaded/installed&nbsp;gempak_upc5.10.4_linux64.tar.gz<BR>&gt;&gt;&nbsp;I&nbsp;am&nbsp;sourcing&nbsp;the&nbsp;5.10.4&nbsp;Gemenviron&nbsp;(which&nbsp;was&nbsp;modified&nbsp;for&nbsp;system<BR>&gt;&gt;&nbsp;specific&nbsp;data&nbsp;locations)<BR>&gt;&gt;
 To invoke the new version of GEMPAK.. I killed all cron jobs 
then<BR>&gt;logged<BR>&gt;&gt;&nbsp;out/in<BR>&gt;&gt; I noted nagrib and 
nagrib2 still require gpend to function in 
both<BR>&gt;5.10.2<BR>&gt;&gt;&nbsp;and&nbsp;5.10.4<BR>&gt;&gt; After the 
errors - for fun I rolled back GEMPAK to ver 5.10.2 an
 
d<BR>&gt;noted<BR>&gt;&gt;&nbsp;nagrib2&nbsp;and&nbsp;nagrib&nbsp;work&nbsp;just&nbsp;fine&nbsp;&nbsp;(I&nbsp;am&nbsp;still&nbsp;running&nbsp;5.10.2)<BR>&gt;&gt;&nbsp;<BR>&gt;&gt;&nbsp;Any&nbsp;advice..&nbsp;does&nbsp;it&nbsp;have&nbsp;to&nbsp;do&nbsp;something&nbsp;with&nbsp;how&nbsp;the&nbsp;"set<BR>&gt;&gt;
 USE_GFORTRAN=1" is uncommented in the Gemenviron file (my system 
is<BR>&gt;a<BR>&gt;&gt;&nbsp;xw8400&nbsp;Linux&nbsp;workstation&nbsp;(HP)&nbsp;running&nbsp;RHEL5&nbsp;x86_64)??<BR>&gt;&gt;&nbsp;<BR>&gt;&gt;&nbsp;Pete<BR>&gt;&gt;&nbsp;<BR>&gt;&gt;&nbsp;<BR>&gt;&gt;&nbsp;-----------------------------------------<BR>&gt;&gt;&nbsp;The&nbsp;information&nbsp;contained&nbsp;in&nbsp;this&nbsp;message&nbsp;is&nbsp;intended&nbsp;only&nbsp;for&nbsp;the<BR>&gt;&gt;&nbsp;personal&nbsp;and&nbsp;confidential&nbsp;use&nbsp;of&nbsp;the&nbsp;recipient(s)&nbsp;named&nbsp;above.&nbsp;If<BR>&gt;&gt;&nbsp;the&nbsp;reader&nbsp;of&nbsp;this&nbsp;message&nbsp;is&nbsp;not&nbsp;the&nbsp;intende
 
d&nbsp;recipient&nbsp;or&nbsp;an<BR>&gt;&gt;&nbsp;agent&nbsp;responsible&nbsp;for&nbsp;delivering&nbsp;it&nbsp;to&nbsp;the&nbsp;intended&nbsp;recipient,&nbsp;you<BR>&gt;&gt;&nbsp;are&nbsp;hereby&nbsp;notified&nbsp;that&nbsp;you&nbsp;have&nbsp;received&nbsp;this&nbsp;document&nbsp;in&nbsp;error<BR>&gt;&gt;&nbsp;and&nbsp;that&nbsp;any&nbsp;review,&nbsp;dissemination,&nbsp;distribution,&nbsp;or&nbsp;copying&nbsp;of<BR>&gt;&gt;&nbsp;this&nbsp;message&nbsp;is&nbsp;strictly&nbsp;prohibited.&nbsp;If&nbsp;you&nbsp;have&nbsp;received&nbsp;this<BR>&gt;&gt;&nbsp;communication&nbsp;in&nbsp;error,&nbsp;please&nbsp;notify&nbsp;us&nbsp;immediately,&nbsp;and&nbsp;delete<BR>&gt;&gt;&nbsp;the&nbsp;original&nbsp;message.<BR>&gt;&gt;&nbsp;<BR>&gt;<BR>&gt;<BR>&gt;Ticket&nbsp;Details<BR>&gt;===================<BR>&gt;Ticket&nbsp;ID:&nbsp;MEU-566855<BR>&gt;Department:&nbsp;Support&nbsp;GEMPAK<BR>&gt;Priority:&nbsp;Normal<BR>&gt;Status:&nbsp;Closed</FONT>
> 


Pete,

I have a nagrib2 process here running every 5 minutes for the FNEXRAD nathr 
composites I create and haven't see this problem,
but will pick through and see if there is a nts problem. 

The think you can check on your end is do a "wc -l gemglb.nts" and make sure 
there are less than 300 entries.
Also make sure your disk space didn't fill up and corrupt the file (not likely 
given it occurs represtedly).

Steve Chiswell
Unidata User Support

Ticket Details
===================
Ticket ID: MEU-566855
Department: Support GEMPAK
Priority: Normal
Status: Closed