Setup problem: "Abnormal return code: 1"


I’m having trouble setting up scidb - apparently with communication with psql. Nothing in init-std*.logs, but extended debug output from “initall-force”:
Removing data directory /home/scidb/data/000/0 on server 0 (localhost), local instance 0
rm -rf /home/scidb/data/000/0/* on [0, ‘localhost’, 0] ( 0 )
mkdir -p /home/scidb/data/000 on [0, ‘localhost’, 0] ( 0 )
mkdir -p /home/scidb/data/000/0 on [0, ‘localhost’, 0] ( 0 )
rm -f SciDB-000-0-scidb on [0, ‘localhost’, 0] ( 0 )
ln -fs /opt/scidb/12.10/bin/scidb SciDB-000-0-scidb on [0, ‘localhost’, 0] ( 0 )
/opt/scidb/12.10/bin/scidb --register -p 1239 --initialize -i localhost -s /home/scidb/data/000/0/storage.cfg -l /opt/scidb/12.10/share/scidb/ --rle-chunk-format=True --chunk-segment-size=100485760 on [0, ‘localhost’, 0] ( 0 )
Abnormal return code: 1
error in command /opt/scidb/12.10/bin/scidb --register -p 1239 --initialize -i localhost -s /home/scidb/data/000/0/storage.cfg -l /opt/scidb/12.10/share/scidb/ --rle-chunk-format=True --chunk-segment-size=100485760 -c host=localhost port=5432 dbname=scidb user=scidb password=singleinstance
scidb@localhost logs in fine (certificate). User scidb owns scidb (db) in postgres, and logs-in with -W & password -or without password. Running Ubuntu 12.04/postgres 9.1/scidb 12.10 from source build.

I’m open to any ideas - Thanks!


Hmmmm …

  1. What’s your platform? Your last line indicated that you’re building SciDB from source … wondering how far off our reservation (Ubuntu 12.04) you’ve gotten.

  2. Can you log into postgres on your localhost?

  3. Did you “sudo -u postgres init_syscat < your_server_name >”? We’ve added this step to the install because we wanted to separate the only bit that requires sudo privs–getting to postgres to create the postgres role / system catalog database–from the parts of the init step that don’t need it.

  4. Postgres 9.1 might be a problem (depending on your platform, etc). We (still) build / test against PostgreSQL 8.4.


I had similar problem on new cluster when did not setup correctly postgres config. I could find correct error message in scidb.log. This log must be in instances data folders or you can just try to find it. So try to check not only init-std***.log but all *.log of instances.


I dropped the scidb table and user; downgraded to postgres 8.4, and called ‘ init_syscat scidb’ under user ‘postgres’. The ‘ initall’ script now returns without errors. Thanks!!


#include <beg.h>

If someone wants to help us out by doing the port to Postgres 9.X, we’d look upon your patch with great favor.


Paul, on my machine I have Postgres 9.1.6 and it works perfect. I guess the real fix was just call ‘ init_syscat scidb’ under user ‘postgres’.


I am facing this error too. Can you please tell me what change you made to your pg_hba.conf file? I get the error …

Removing data directory /home/scidb/data/002/1 on server 2 (qp2), local instance 1
Remote command exception:
cd /home/scidb/data/002/1;/opt/scidb/12.10/bin/scidb --register -p 1240 -i qp2 -s /home/scidb/data/002/1/storage.cfg -l /opt/scidb/12.10/share/scidb/ --rle-chunk-format=True --chunk-segment-size=100485760 -c ‘host=localhost port=5432 dbname=cluster user= password=’ 1> init-stdout.log 2> init-stderr.log
Abnormal return code: 1 stderr:

In the log files I can see the error “scidb::SCIDB_SE_SYSCAT::SCIDB_LE_CANT_CONNECT_PG”. Postgres is required to run only on coordinator, isn’t it? I have modifies the postgres.conf and added line mentioned lines so that Postgres can use ‘trust’ authentication for local and remote connections. I verified that the CIDR mask is correct.

Any help is highly appreciated.



I also have the same problem. It didn’t work even after I downgraded postgres to 8.4 :frowning:
For now, I installed scidb 12.3 on older ubuntu 11.04 and it works just fine :mrgreen: But I really want to get scidb 12.10 to work on ubuntu 12.04