E_FAIL error

[GoogleCodeExporter]

I reran an annual social swap run with the newest version (data.zip on 
FTP, I am calling it revision 104B). During I received a bunch of error 
messages that said:

01:45:46:281:  Error HRESULT E_FAIL has been returned from a call to a COM 
component.
   at ESRI.ArcGIS.Geodatabase.IFeatureClass.Update(IQueryFilter filter, 
Boolean Recycling)
   at SEARCH.Map.resetFields(String fieldName, String oldValue, String 
newValue) in C:\v104B\DataCentric\Map.cs:line 1158

In the animal.log the time stamps seemed to all be linked to when the 
social map is switched and many of the animals die due to insufficient 
area (notice time jump):

01:45:20:609:  inside winter kill for resident number 39
01:45:20:609:  inside winter kill with roll of 0.180736884558917
01:45:20:609:  my chance of winter kill is 0
01:51:20:359:  inside set inital sleeptime curr time is 10/2/2010  6:00 AM
01:51:20:359:  inside calc sleep getting the number of hours before I go 
to sleep

I received 36 of these errors. This corresponds to the number of animals 
that died after the map switch so I think that is probably what is causing 
the error. I have attached the xml and the dispersal_error.log for you to 
look at.

Original issue reported on code.google.com by bpa...@purdue.edu on 21 Oct 2009 at 7:34

Attachments:

• ann_soc_swap.xml
• DispersalError.log

[GoogleCodeExporter]

I am also receiving this error on a test where I had residents subject to a 10% 
timestep mortality. In most instances when an animal died there were no 
problems and 
the HR polygon was reset to unoccupied. For some reason, during one particular 
timestep (8/28/2009 09:00:00 which was right after the last disperser set up a 
HR 
for that year), when residents died of timestep mortality their HR polygon was 
not 
removed from the social map and a E_FAIL error was written out to the 
Dispersal_error.log. The seven error messages seem to perfectly match up with 
the 
seven animals that died during that timestep (#4,5,21,32,33,37,38). The animals 
seemed to die in the text files (line written out) and the logs (no longer 
looped 
through during mortality loops) but just not in the social maps. I have 
attached the 
error log and the xml for you to look at.

Original comment by bpa...@purdue.edu on 22 Oct 2009 at 2:04

Attachments:

• DispersalError.log
• res_time_deathC.xml

[GoogleCodeExporter]

Ben
I am unable to duplicate this issue using ann_soc_swap.  I do not have a
little_suitable social map to swap out with.  So I used not_suit instead.  Not 
sure
if that was the issue or not.  But then it would point to map issues not program
issues.  And this really sounds like program issues.  I first tried with using a
smaller release map to save time (Release_Sites2) and that passed.  So I thought
maybe I needed a full blown run and went back to 40mf, and still passed.  And 
all 40
died so maybe you need to send me the other map so some live and some die.?

Original comment by ran...@mwwb.net on 27 Oct 2009 at 11:30

[GoogleCodeExporter]

Sorry about that Bob. Stupid mistake on my part. I have attached the 
little_suitable.shp so you should be able to recreate the run now.

Original comment by bpa...@purdue.edu on 27 Oct 2009 at 2:51

Attachments:

• little_suitable.zip

[GoogleCodeExporter]

Original comment by bpa...@purdue.edu on 4 Nov 2009 at 2:51

• Added labels: Priority-Critical
• Removed labels: Priority-Medium

[GoogleCodeExporter]

I reran both of the above tests that caused the efail problem (ann_soc_swap and 
res_time_death). In the annual social swap run I no longer received the efail 
error 
(although I did get a bunch of errors related to tempMapManger.txt; see other 
issue 
for more detail). In resident death run, though, I again received the efail 
error. 
Again it seemed to be related to residents that died after the last disperser 
settled. I have attached the dispersal error log and used the xml posted above.

Original comment by bpa...@purdue.edu on 6 Nov 2009 at 6:05

Attachments:

• DispersalError.log

[GoogleCodeExporter]

Ok it is as I suspected last night.  This problem is closed.  

I did not suspect what the cause was though.  Now it appears that when we see 
the
problem using res_time_death xml it is related to the issue about the animals 
maps
not being correctly named.

Lesson to be learned is when you have an issue on one xml file it is one issue. 
 Just
because you think you see the same error using a different xml file, does not 
mean it
is the same.

That is why I thought it was different last night, it worked on one input file 
and
not the other.  So I am going to close this one down and start on the naming 
issue. 
I will use the res_time_death file to verify that is issue is fixed. 

Original comment by ran...@mwwb.net on 7 Nov 2009 at 4:32

• Changed state: Fixed

[GoogleCodeExporter]

seems to be fixed

Original comment by bpa...@purdue.edu on 8 Nov 2009 at 3:34

• Changed state: Verified