#POSTGIS IRC Log - 2008-08-21

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
00:45:38 CIA-6: mcayland * r2918 /trunk/ ( 3 files in 2 dirs ):
00:45:38 CIA-6: Alter the lwgeom Makefile so that liblwgeom.a is linked directly, rather than
00:45:38 CIA-6: using the -L...-l options on the command line. This is to prevent problems on
00:45:39 CIA-6: OSX where PGXSadds the PostgreSQL $libdir to the PostGIS link line, causing it
00:45:39 CIA-6: to link to liblwgeom.so produced by older versions of PostGIS rather than the
00:45:40 CIA-6: static liblwgeom.a.
00:45:42 CIA-6: Also change the PGXS CPP/LIB equivalent variables to use += so that extra compile/link options can be supplied. Per report from Robert Rainthorpe, with thanks to William Kyngesburye.
02:57:07 CIA-6: mcayland * r2919 /trunk/lwgeom/ ( lwgeom_gist.c lwpostgis.sql.in.c ):
02:57:07 CIA-6: Update PostGIS SVN trunk to use the new 8.4 GiST API ( where RECHECK is now
02:57:07 CIA-6: specified within the consistent function, rather than being supplied as part of
02:57:07 CIA-6: the operator class definition ). This allows PostGIS to compile and pass
02:57:07 CIA-6: regression tests on the latest PostgreSQL 8.4 CVS HEAD.
03:28:31 CIA-6: mcayland * r2920 /trunk/README.postgis: Update PostGIS README file to account for changes in SVN trunk.
06:23:31 ElAngelo: hi
06:24:26 ElAngelo: does anyone here have a clue on what the binary format esri dumps in postgresql, the so called "SDEBINARY", looks like? like... which compression did they use? is there any encryption envolved?
11:00:02 milo1: hi all... i have a prj for a co-ord system that isn't defined in the standard spatial_ref_sys table that comes with PostGIS... is there any special considerations involved in adding a new SRS?
11:01:54 bitner: milo1: you'll want to insert a new record similar to the "postgis spatial_ref_sys insert statement" option at http://spatialreference.org
11:01:55 sigq: Title: Home -- Spatial Reference ( at spatialreference.org )
11:03:24 milo1: ahh
11:04:23 milo1: just what i needed. thanks!
11:04:44 milo1: is that site mentioned in the PostGIS doc? i didnt see it
11:04:53 bitner: just make sure the srid is whatever you want it to be used as in postgis -- sr.org by default adds a 9 to the front of the srid so it doesn't step on the toes of anything already in the spatial_ref_sys table
11:05:07 bitner: milo1: not that I know of
11:05:59 milo1: should i remove the 9 then if it doesnt clash without it?
11:06:36 bitner: sure -- just make that number whatever you want to use in setsrid( ) and transform( ) and the like
11:06:49 milo1: yeah. is there a central authority that allocates the SRS IDs?
11:07:22 bitner: the epsg is the primary authority, but srid in postgis is really just an identifier
11:07:47 bitner: and postgis srids for convenience are mostly mapped to epsg identifiers
11:09:12 milo1: fair enough
16:28:56 Drago84: hi
16:29:02 Drago84: i have a ask
16:29:26 Drago84: is possible distribuite a database in more host?
16:30:30 mleslie: www.slony.info
16:31:13 Drago84: yes but slony duplicate databese true?
16:31:32 Drago84: its not distribuite database but is duplicate its in more host
16:31:50 mleslie: True. It distributes the load, not the tables.
16:32:01 mleslie: There's also plproxy, but I haven't found the link yet.
16:32:14 Drago84: :(
16:32:40 Drago84: but postgres have this function of distribuite data in more host
16:32:45 Drago84: for exemple
16:32:48 mleslie: http://pgfoundry.org/projects/plproxy/
16:32:49 sigq: Title: PgFoundry: PL/Proxy: Project Info ( at pgfoundry.org )
16:33:10 mleslie: Postgres doesn't do it out-of-the-box AFAIK.
16:33:22 Drago84: if i use tuhe copy from file is possibile distribuite the data in more host?
16:33:31 Drago84: :(
16:34:16 mleslie: Why do you need to distribute it?
16:34:54 Drago84: becouse my file is very very large
16:35:12 Drago84: ( sorry for my english )
16:35:35 Drago84: i use file of 12 tb size
16:35:37 Drago84: :|
16:35:57 Drago84: my idea is distribuite the database in 20 node
16:36:13 Drago84: is a bad idea?
16:37:06 iggy: you have 12TB of geographic data?
16:37:50 Drago84: yes
16:38:24 iggy: ffs, I thought I had a lot of data, what is it if you don't mind my asking
16:38:43 Drago84: i have in a CED 20 pc with 32 gb ram
16:39:56 Drago84: my idea is to use postgres for distribuite this record on all 20 node
16:40:44 Drago84: but i dont understand if postgres is able to do this
16:41:36 Drago84: in the documentations its not write if is possible distribuite database in mode host
16:43:51 iggy: it's not built into postgres
16:44:02 Drago84: :|
16:44:18 iggy: I don't know of anything that does that automatically
16:44:28 Drago84: oracle :P
16:44:50 iggy: you could split it up yourself and give your application some minimal logic for figuring out what is where
16:45:02 iggy: then you should probably stick with oracle
16:45:33 Drago84: :P
16:45:54 Drago84: yes i split its and give my nodes
16:46:33 Drago84: but than i have problem to syncronizaded the acces :P
16:46:38 Drago84: access*
16:47:37 iggy: what?
16:48:37 Drago84: my idea is to creat a principal node to organize access and the others node for immagazinate the datas
16:48:53 Drago84: ( uff my english is very very bad sorry )
16:49:08 iggy: yeah, immagazinate didn't translate well
16:49:59 tazle: Drago84: can you tell us what kind of data this 12 TB is?
16:50:20 Drago84: terabyte
16:50:39 bitner: Drago84: you can use tablespaces and table inheritance
16:50:43 tazle: Drago84: yes, I did get the unit, but Í'm curious about what the data actually is
16:51:57 bitner: http://www.postgresql.org/docs/8.3/interactive/ddl-partitioning.html
16:51:58 sigq: Title: PostgreSQL: Documentation: Manuals: PostgreSQL 8.3: Partitioning ( at www.postgresql.org )
16:52:31 Drago84: they are informations of mode system
16:52:35 Drago84: more*
16:52:48 bitner: http://www.postgresql.org/docs/8.3/interactive/manage-ag-tablespaces.html
16:52:49 sigq: Title: PostgreSQL: Documentation: Manuals: PostgreSQL 8.3: Tablespaces ( at www.postgresql.org )
16:53:18 bitner: Drago84: between the two methods I just linked to, you can do exactly what you are after completely within postgresql
16:54:35 : * iggy bookmarks for later
16:54:56 Drago84: can i send all log in one node?
16:56:00 Drago84: otherwise i have the same of logs then nodes
16:56:11 Drago84: the same number*
17:02:55 Drago84: uff uff
17:05:40 mleslie: Drago84: I'm not sure, but I would expect each postgresql instance to maintain it's own logs. You can always rsync them onto a single server as needed, or something similar.
17:12:33 Drago84: uff im stupid but i not understand with distribuite one instance in more servers :(