It's typically recommended that you use the pg_dump of the same
version as the pg_restore. In your case, the v12 pg_dump will
successfully read from a 10.1 database.
On 12/19/19 9:24 AM, Sar wrote:
for pg_dump it is: PostgreSQL 10.1, compiled by Visual C++ build
1800, 64-bit
for pg_restore it is the same machine as well as newest
version of Postgresql 12, so I assume the error is in pg_dump.
selects work no problem, I even dumped one table with COPY
and it doesnt raise any issue. The other table however is
probably too big to ever do this.
schema for one table:
CREATE TABLE table1 (
foreign_id integer,
action character varying(10),
numarr smallint[]
);
and other table:
CREATE TABLE log (
id integer NOT NULL,
place character varying(50),
place_no integer,
"time" character varying(80),
action character varying(100),
"user" character varying(50),
place_definition character varying(10),
place_definition2 character varying(50),
place_definition3 character varying(30)
);
On 12/16/19
9:46 AM, Sar wrote:
> Hi,
> I have database on Windows Server, running postgresql
10 and I've run
> pg_dump with options: -Fc --verbose -d db_name --file
"E:\backup.sql"
> -T some_table.
> Now when I run pg_restore, also on windows server with
postgresql 12, or
> on the same original server with postgresql 10, using
options
> --data-only --verbose -j 8,
> or without jobs, but selecting some specific few tables
with -t option I
> get follwing error:
Which Postgres version of pg_dump/pg_restore are you using
to do above?
> error: pg_restore: error: found unexpected block ID (0)
when reading
> data -- expected 3142
> This error will only happen on very few tables, most
run fine.
What happens if you SELECT from these tables in the Postgres
10 instance?
Can you show the schema definition for one or more of the
problem tables?
> Currently, I'm running without jobs and without tables
selected with -t,
> however I haven't gotten to the specific table that
causes the trouble
> (the restore will take days), however I already fear
when it comes to
> that table the same error will trigger.
> I guess this could serve as a possible bug report as
well, but any
> advice as to what I can do is appreciated.
--
Adrian Klaver
adrian.klaver@xxxxxxxxxxx
--
Angular momentum makes the world go 'round.
|
