[postgis-users] ERROR: array size exceeds the maximum allowed (134217727)

[Mike Leahy]

Hello again,

It might be of interest to point out that substituting st_union() with 
st_memunion() seems to have worked around this.  I'm curious though, because 
there is not a great deal of data being processed, and I am running this on a 
fairly sturdy system that that has more capacity than some of the Fedora 
systems I'm running.

Mike

On Saturday 20 March 2010 05:17:50 Mike Leahy wrote:
> Hello list,
> 
> I have an interesting error that's coming up on an Ubuntu 64-bit system,
> running PostgreSQL 8.4.2, and PostGIS 1.5.1 (packaged by me) as well as
>  1.4.0 (from Ubuntu repos).  The query itself is somewhat complex (I'm sure
>  it could be optimized better - but it's been working well for quite some
>  time).  I've attached two files - one is a dummed-down version of the
>  query I am running (crashing.sql).  After discussing this with some folks
>  in the postgresql IRC channel, it seems that the st_geom(), or something
>  PostGIS-related is causing the error.
> 
> Running this query on various data will produce one of two results.  One is
> the error mentioned in the subject (ERROR:  array size exceeds the maximum
> allowed (134217727)).  I can find very little information on this error. 
>  The other outcome is that it often causes the PostgreSQL backend to
>  segfault (see gdboutput.txt).
> 
> I can attest that this query works fine on many installations of PostgreSQL
> 8.2+, and PostGIS 1.3+.  The only main difference I can see in this case
>  where I'm getting errors and segfaults is that I'm running this on Ubuntu,
>  whereas all of my successful cases are on Fedora (7 through 12).
> 
> If I replace the st_union() function in the query with accum(), it still
> produces the error/segfaults.
> 
> Also, strangely, if I remove one of the where conditions near the end of
>  the query (e.g., "  AND (cl.depth = 0)"), it suddenly works fine without
>  any issues (except that I actually need that condition).
> 
> Does this sort of problem look familiar to anyone?  I can try to put
>  together an example if that would be of any use, but since the same
>  data/query works on other Fedora systems that I am running, I'm not sure
>  how easy it is to replicate.
> 
> Regards,
> Mike
>