Re: Dumping data using pg_dump after chrooting to a different partition

Thanks Craig.

After configuring to accept TCP connections on port 5432, I tried to specify the hostname as shown below and that didn't help. Is there anything else that needs to be configured?
pg_dump -h bldr-ccm36.cisco.com -p 5432 -a -U postgres
pg_dump: [archiver (db)] connection to database "postgres" failed: could not connect to server: Connection refused
    Is the server running on host "bldr-ccm36.cisco.com" and accepting
    TCP/IP connections on port 5432?

Radha

On 10/24/11 4:45 PM, Craig James wrote:

On 10/24/11 3:10 PM, Krishnamurthy Radhakrishnan wrote:

Hi,

I am new to PostgreSQL. We are using PostgreSQL 9.0.2 on our linux server. We have an instance of PostgreSQL 9.0 running using the primary partition on the server.

We want to use the pg_dump and psql programs to migrate the data during our software upgrade process. For upgrade, we plan to do the following:

• chroot to a secondary partition on the server.
• install the software RPMs including PostgreSQL RPMs
• start a secondary instance of PostgreSQL DB server using a different port and data directory.
• run pg_dump to dump the data from the primary instance to a file.
• run psql to import the data from the file into the secondary instance.

However when I tried to perform the pg_dump as mentioned above, I get the following error:
pg_dump: [archiver (db)] connection to database "TestDB" failed: could not connect to server: No such file or directory|<LVL::Debug>
Is the server running locally and accepting|<LVL::Debug>
connections on Unix domain socket "/tmp/.s.PGSQL.5432"?|<LVL::Debug>

I suspect the problem is that localhost sockets on Unix/Linux are actual implemented as file-system sockets rather than TCP/IP sockets.  If you do chroot, those files won't exist.  Try connecting using a "-h hostname (e.g. -h myserver.domain.com) rather than the default localhost.  You may have to reconfigure your server to listen on port 80.

Craig