Use of Crash logs

Questions and other topics related to UEMS 15.
Howard
Posts: 62
Joined: Wed Aug 14, 2013 3:13 am

Re: Use of Crash logs

Post by Howard » Tue Jan 12, 2016 12:16 am

Did some work on hostname
which allowed it to assign an ip address to workstation1

Crashed again different error in real_0002.err log
run_real.log looks OK

Code: Select all

taskid: 2 hostname: workstation1
-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE:  <stdin>  LINE:     854
 input_wrf.F: SIZE MISMATCH:  namelist ide,jde,num_metgrid_levels=          100          125           27 ; input data ide,jde,num_metgrid_levels=          100          130           27
-------------------------------------------
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 2
and also real_0001.err log

Code: Select all

taskid: 1 hostname: workstation1
-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE:  <stdin>  LINE:     854
 input_wrf.F: SIZE MISMATCH:  namelist ide,jde,num_metgrid_levels=          100          125           27 ; input data ide,jde,num_metgrid_levels=          100          130           27
-------------------------------------------
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 1
unsure what is going on ...

Howard

Howard
Posts: 62
Joined: Wed Aug 14, 2013 3:13 am

Re: Use of Crash logs

Post by Howard » Tue Jan 12, 2016 1:01 am

Problem appears to be fixed ... loop back to get an IP address fixed it (I think)

Thanks for all your help ...

smartie
Posts: 97
Joined: Sat May 21, 2011 7:34 am

Re: Use of Crash logs

Post by smartie » Tue Jan 12, 2016 10:25 am

Hi Howard,
glad to hear it's working.

For the record can you say explicitly what the fault was and your fix.?

There are a number of reports of similar issues, but no fixes. Similar MPI errors do identify problems with Hostname and the loopback interface individual to machines.

Best,
David

Ryan_Sutton
Posts: 3
Joined: Thu Nov 10, 2016 8:28 pm

hostname and segmentation fault (signal 11)

Post by Ryan_Sutton » Thu Nov 10, 2016 10:49 pm

I am a new IT linux support staff for Embry-Riddle Prescott. My first task is to build and get UEMS functional for student classroom use. I am not a EMS user, so my knowledge of WRF/EMS is minimal. Regardless I came across 2 issues in this forum relating to a fresh install of UEMS. I decided to post my experience as it might help others.

First issue is DNS lookup and resolution. This environment uses a specific 'prwks' sub domain dns zone for all dynamic client host registrations. The default root parent/search zone is not used for dns client updates. I found that Redhat 7.2 using hostname of "uemshost" or "uemshost.localdomain" did not process correctly with the campus DNS service which impacted UEMS install, it would not even start, complaining about null addr variable. To correct this, Redhat hostname is the FQDN of the client registration zone: uemshost.prwks.erau.edu.

Second issue is stock Redhat 7 Workstation needs perl-Data-Dumper module installed or the run_wrfm process failed with Segmentation fault (signal 11). Note the 'sysinfo' script also fails with Segfault (signal 11) error without this perl module.

System details:

Optiplex o9040 i7 hyper threading off
OS: Redhat 7.2 currently upgraded to 7.3 (Maipo) 3.10.0-514.el7.x86_64
Default partitioning LVM, plus standard partition ext4 mount: /usr1
Gnome desktop
RAM: 16GB
Openjava 1.8 (stock) is being used for Iced Tea to run IDV viewer on Redhat.

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest