Thread: Upgrade 7.0.2 -> 7.1.2: Errors during Restore

Upgrade 7.0.2 -> 7.1.2: Errors during Restore

From
Jim Caley
Date:
On my Red Hat Linux 6.2 box, I just upgraded from 7.0.2 to 7.1.2.  I used the
7.0.2 pg_dumpall, and after moving the 7.1.2 installation into place, firing up
postmaster, and running initdb I restored my pg_dumpall output.

The results of the restore seem to be fine, but there were a number of errors
during it and I want to make sure they're harmless.  My guess is they are,
especially since they didn't take place within the areas of the dump file where
my data was located.  A number of them seem to be throwbacks to 7.0RC1, based on
the path given in the messages.

I've pulled all the error messages and put them below.  Many of them regard
failure to stat regress.so, and then there are a variety of other ones regarding
nonexistent functions, etc.

I'd just like to know if I need to be concerned or not.

Thanks,
Jim

--------------------------------
ERRORS DURING RESTORE OF 7.0.2 DUMP TO 7.1.2:

psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:142: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../regress.so': No such f
ile or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:143: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../regress.so': No such f
ile or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:144: ERROR:  TypeCreate: func
tion 'widget_in(opaque)' does not exist
.
.
.
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:850: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../../../../contrib/spi/r
efint.so': No such file or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:851: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../../../../contrib/spi/r
efint.so': No such file or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:852: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../../../../contrib/spi/a
utoinc.so': No such file or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:853: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../regress.so': No such f
ile or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:854: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../regress.so': No such f
ile or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:855: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../regress.so': No such f
ile or directory
.
.
.
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:863: ERROR:  ProcedureCreate:
 arg type 'widget' is not defined
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:864: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../regress.so': No such f
ile or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:865: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../regress.so': No such f
ile or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:866: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../regress.so': No such f
ile or directory
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:867: ERROR:  stat failed on f
ile '/usr/src/postgresql-7.0RC1/src/test/regress/input/../regress.so': No such f
ile or directory
.
.
.
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:1581: ERROR:  AggregateCreate
: function 'int4div(int4)' does not exist
CREATE
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:1583: NOTICE:  DefineAggregat
e: attribute "sfunc2" not recognized
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:1583: NOTICE:  DefineAggregat
e: attribute "stype2" not recognized
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:1583: NOTICE:  DefineAggregat
e: attribute "initcond2" not recognized
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:1583: ERROR:  Define: "stype"
 unspecified
CREATE
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:1592: ERROR:  OperatorGet: ri
ght type "widget" does not exist
.
.
.
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:118323: ERROR:  Function 'int
erpt_pp(path, path)' does not exist
        Unable to identify a function that satisfies the given argument types
        You may need to add explicit typecasts
.
.
.
psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:118341: ERROR:  Rule WHERE co
ndition may not contain references to other relations

END ERRORS

Re: Upgrade 7.0.2 -> 7.1.2: Errors during Restore

From
Tom Lane
Date:
Jim Caley <caley@chesco.com> writes:
> The results of the restore seem to be fine, but there were a number of errors
> during it and I want to make sure they're harmless.

Most of those look like you had a "regression" database hanging around
in your old installation; it's complaining because the CREATE FUNCTION
commands refer to shared libraries that you evidently don't have around
any more.

> psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:118341: ERROR:  Rule WHERE co
> ndition may not contain references to other relations

This one *might* be a real problem; you'll need to look at line 118341 of
your dump file and see if the failed rule is one of yours or not.  Since
the line number is so close to the others, I suspect it's also an
artifact of trying to reload the 7.0 regression database.

            regards, tom lane

Re: Upgrade 7.0.2 -> 7.1.2: Errors during Restore

From
Jim Caley
Date:
> Most of those look like you had a "regression" database hanging around
> in your old installation; it's complaining because the CREATE FUNCTION
> commands refer to shared libraries that you evidently don't have around
> any more.

Sure enough, okay I'll drop it.

> > psql:./pgdump.all-just.before.going.to.pgsql-7.1.2:118341: ERROR:  Rule WHERE co
> > ndition may not contain references to other relations
>
> This one *might* be a real problem; you'll need to look at line 118341 of
> your dump file and see if the failed rule is one of yours or not.  Since
> the line number is so close to the others, I suspect it's also an
> artifact of trying to reload the 7.0 regression database.

Yes, this rule is from the regression database.

Many thanks,
Jim