Opened 11 years ago

Closed 11 years ago

#411 closed defect (invalid)

rasgeo reports errors anddumps backtrace almost every time run

Reported by: Marcus Sen Owned by: Dimitar Misev
Priority: major Milestone: 8.4.4
Component: rasgeo Version: 8.4
Keywords: Cc:
Complexity: Medium

Description

When I run rasgeo commands like rasimport and raserase I get errors like those shown below even though the commands seem to run correctly and do the requested insert or delete. I give 3 examples, one raserase command, one rasimport and one where I've deliberately made a mistake in the input syntax where there is still a backtrace although all it has to do is report the syntax error.

I am not sure whether this is a major or minor problem as the commands seem to be run correctly when I do them one at a time. I have had problems with some failing when I run several in succession from a shell script but I don't know whether this is related or a different problem. Also the error status makes it difficult to test for succesful running when executing from, for example, a python script.

I am using 8.4.3 RPMs

The example commands are below, the output is attached in 3 files named appropriately.

Example 1 - a collection previously inserted with rasimport using raserase:

$ raserase -conn ~/.rasdaman/rasconnect -coll test_mb8

Example 2 - Inserting a coverage using rasimport. Note that the nmrat_48641 table which the error message says already exists did not exist before executing the rasimport command but does after running it.

$ rasimport -conn ~/.rasdaman/rasconnect -f Landsat_205_20/landsat_205_20_stack.tif -coll test_mb8 -t x8CharImage:x8CharSet

Example 3 - Trying to execute a command with invalid syntax

$ raserase faierqae

Attachments (3)

example1_output.txt (6.3 KB ) - added by Marcus Sen 11 years ago.
Example 1 output
example2_output.txt (6.7 KB ) - added by Marcus Sen 11 years ago.
Example 2 output
example3_output.txt (7.0 KB ) - added by Marcus Sen 11 years ago.
Example 3 output

Download all attachments as: .zip

Change History (6)

by Marcus Sen, 11 years ago

Attachment: example1_output.txt added

Example 1 output

by Marcus Sen, 11 years ago

Attachment: example2_output.txt added

Example 2 output

by Marcus Sen, 11 years ago

Attachment: example3_output.txt added

Example 3 output

comment:1 by abeccati, 11 years ago

I noticed rasimport segmentation faults usually happen when using a GDAL version that is not the one against which rasdaman was compiled. Are you also using the packaged GDAL version from EOX maintained repository on the server?

comment:2 by Marcus Sen, 11 years ago

I'm using gdal-1.9.0-1.eox2.el6.x86_64.

and rasdaman packages:

rasdaman-docs-8.4.3-1.el6.noarch
rasdaman-debuginfo-8.4.3-1.el6.x86_64
rasdaman-petascope-8.4.3-1.el6.noarch
rasdaman-8.4.3-1.el6.x86_64
rasdaman-examples-8.4.3-1.el6.noarch
rasdaman-devel-8.4.3-1.el6.x86_64
rasdaman-rasgeo-8.4.3-1.el6.x86_64
rasdaman-rasdaview-8.4.3-1.el6.x86_64

comment:3 by Marcus Sen, 11 years ago

Resolution: invalid
Status: newclosed

By setting up a new machine from scratch I found that I should have gdal-1.9.0-1.eox3.el6.x86_64 rather than gdal-1.9.0-1.eox2.el6.x86_64. This hadn't been showing up in my yum updates because my /etc/yum.repos.d.eox-testing file had an

includepkgs = rasdaman*

directive so it wasn't picking up gdal from this repository. I think some earlier instructions I had received had suggested to include this directive but I changed it to

includepkgs = rasdaman* gdal*

did yum update and got the eox3 package and now rasimport works fine.

Note: See TracTickets for help on using tickets.