[Live-demo] [OSGeo] #991: chroot build method: mapfish does not start

OSGeo trac_osgeo at osgeo.org
Sat Aug 11 20:38:03 PDT 2012


#991: chroot build method: mapfish does not start
----------------------+-----------------------------------------------------
 Reporter:  kalxas    |       Owner:  live-demo@…              
     Type:  defect    |      Status:  new                      
 Priority:  critical  |   Milestone:  OSGeoLive6.0RC1          
Component:  LiveDVD   |    Keywords:  6.0 beta9                
----------------------+-----------------------------------------------------

Comment(by hamish):

 Hi,

 Brian and myself have been working on this one a bit, and we have a theory
 about what's happening.

 exec summary: the data base was not created correctly; it's not mapfish's
 fault. inchroot.sh is setting LC_ALL=C which is propagating to
 install_postgis.sh and the postgres templates are being created as
 SQL_ASCII encoding instead of UTF8 encoding. Mapfish was expecting it to
 be UTF8.

 solutions include: not having inchroot do that, having install_postgis.sh
 unset/reset the LANG variables at the beginning of its script, and/or
 explicitly stating the encoding type at each createdb step on the liveDVD.

 We'll wait to hear why inchroot was doing that before touching it, have
 added some debug messages to install_postgis.sh to get a clearer idea of
 what's happening there, and Brian's gone around and added "-E UTF8" in a
 number of places where it wasn't explicitly stated.


 ----

 from yesterday's 6.0-build-r8942 build log:
 {{{
 ...
 *************** PICKED VERSIONS ****************
 [versions]

 *************** /PICKED VERSIONS ***************
 createdb: database creation failed: ERROR:  new encoding (UTF8) is
 incompatible with the encoding of the template database (SQL_ASCII)
 HINT:  Use the same encoding as in the template database, or use template0
 as template.
 ...
 }}}


 ----

 in the 5.5 release all postgres tables were encoded as UTF8, same for
 6.0beta7 (the last of the remastersys line); we checked.

 For yesterday's 6.0-build-r8942 however:
 {{{
 user at osgeolive:~$ psql -l
                                  List of databases
        Name       |  Owner   | Encoding  | Collate | Ctype |   Access
 privileges
 ------------------+----------+-----------+---------+-------+-----------------------
  52nSOS           | postgres | UTF8      | C       | C     | =Tc/postgres
 +
                   |          |           |         |       |
 postgres=CTc/postgres+
                   |          |           |         |       |
 user=CTc/postgres
  RASBASE          | user     | SQL_ASCII | C       | C     |
  mapbender        | user     | UTF8      | C       | C     |
  natural_earth    | user     | UTF8      | C       | C     |
  osm_local        | user     | UTF8      | C       | C     |
  osm_local_smerc  | user     | UTF8      | C       | C     |
  petascopedb      | user     | SQL_ASCII | C       | C     |
  pgrouting        | user     | SQL_ASCII | C       | C     |
  postgres         | postgres | SQL_ASCII | C       | C     |
  sahana           | sahana   | SQL_ASCII | C       | C     |
  template0        | postgres | SQL_ASCII | C       | C     | =c/postgres
 +
                   |          |           |         |       |
 postgres=CTc/postgres
  template1        | postgres | SQL_ASCII | C       | C     | =c/postgres
 +
                   |          |           |         |       |
 postgres=CTc/postgres
  template_postgis | user     | UTF8      | C       | C     |
  user             | user     | SQL_ASCII | C       | C     |
 (14 rows)

 }}}


 cheers,
 Hamish

-- 
Ticket URL: <https://trac.osgeo.org/osgeo/ticket/991#comment:4>
OSGeo <http://www.osgeo.org/>
OSGeo committee and general foundation issue tracker.


More information about the Osgeolive mailing list