pg_dump failure, can't get data out...

Started by Larry Rosenmanover 24 years ago10 messages
#1Larry Rosenman
ler@lerctr.org

I had upgraded yesterday and *THOUGHT* all was fine.

Forgot to pg_dump, so I restored my $PGLIB and bin directory, now get
the following when I try to pg_dump:

$ pg_dump ler >z
Password:
pg_dump: query to get function name of oid - failed: ERROR: oidin:
error in "-": can't parse "-"
$

And the debug query log:

Aug 26 19:10:05 lerami pg-prod[3861]: [1] DEBUG: connection: host=[local] user=ler database=ler
Aug 26 19:10:05 lerami pg-prod[3861]: [2] DEBUG: InitPostgres
Aug 26 19:10:05 lerami pg-prod[3861]: [3] DEBUG: StartTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [4] DEBUG: query: select getdatabaseencoding()
Aug 26 19:10:05 lerami pg-prod[3861]: [5] DEBUG: ProcessQuery
Aug 26 19:10:05 lerami pg-prod[3861]: [6] DEBUG: CommitTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [7] DEBUG: StartTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [8] DEBUG: query: SELECT version();
Aug 26 19:10:05 lerami pg-prod[3861]: [9] DEBUG: ProcessQuery
Aug 26 19:10:05 lerami pg-prod[3861]: [10] DEBUG: CommitTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [11] DEBUG: StartTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [12] DEBUG: query: begin
Aug 26 19:10:05 lerami pg-prod[3861]: [13] DEBUG: ProcessUtility: begin
Aug 26 19:10:05 lerami pg-prod[3861]: [14] DEBUG: CommitTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [15] DEBUG: StartTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [16] DEBUG: query: set transaction isolation level serializable
Aug 26 19:10:05 lerami pg-prod[3861]: [17] DEBUG: ProcessUtility: set transaction isolation level serializable
Aug 26 19:10:05 lerami pg-prod[3861]: [18] DEBUG: CommitTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [19] DEBUG: StartTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [20] DEBUG: query: SELECT datlastsysoid from pg_database where datname = 'ler'
Aug 26 19:10:05 lerami pg-prod[3861]: [21] DEBUG: ProcessQuery
Aug 26 19:10:05 lerami pg-prod[3861]: [22] DEBUG: CommitTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [23] DEBUG: StartTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [24] DEBUG: query: select (select usename from pg_user where datdba = usesysid) as dba from pg_database where datname = 'ler'
Aug 26 19:10:05 lerami pg-prod[3861]: [25] DEBUG: ProcessQuery
Aug 26 19:10:05 lerami pg-prod[3861]: [26] DEBUG: CommitTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [27] DEBUG: StartTransactionCommand
Aug 26 19:10:05 lerami pg-prod[3861]: [28-1] DEBUG: query: SELECT pg_type.oid, typowner, typname, typlen, typprtlen, typinput, typoutput, typreceive, typsend, typelem,
Aug 26 19:10:05 lerami pg-prod[3861]: [28-2] typdelim, typdefault, typrelid, typalign, typstorage, typbyval, typisdefined, (select usename from pg_user where typowner =
Aug 26 19:10:05 lerami pg-prod[3861]: [28-3] usesysid) as usename, format_type(pg_type.oid, NULL) as typedefn from pg_type
Aug 26 19:10:05 lerami pg-prod[3861]: [29] DEBUG: ProcessQuery
Aug 26 19:10:06 lerami pg-prod[3861]: [30] DEBUG: CommitTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [31] DEBUG: StartTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [32-1] DEBUG: query: SELECT pg_proc.oid, proname, prolang, pronargs, prorettype, proretset, proargtypes, prosrc, probin, (select
Aug 26 19:10:06 lerami pg-prod[3861]: [32-2] usename from pg_user where proowner = usesysid) as usename, proiscachable, proisstrict from pg_proc where pg_proc.oid >
Aug 26 19:10:06 lerami pg-prod[3861]: [32-3] '16554'::oid
Aug 26 19:10:06 lerami pg-prod[3861]: [33] DEBUG: ProcessQuery
Aug 26 19:10:06 lerami pg-prod[3861]: [34] DEBUG: CommitTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [35] DEBUG: StartTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [36-1] DEBUG: query: SELECT pg_aggregate.oid, aggname, aggtransfn, aggfinalfn, aggtranstype, aggbasetype, agginitval, 't'::boolean as
Aug 26 19:10:06 lerami pg-prod[3861]: [36-2] convertok, (select usename from pg_user where aggowner = usesysid) as usename from pg_aggregate
Aug 26 19:10:06 lerami pg-prod[3861]: [37] DEBUG: ProcessQuery
Aug 26 19:10:06 lerami pg-prod[3861]: [38] DEBUG: CommitTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [39] DEBUG: StartTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [40-1] DEBUG: query: SELECT pg_operator.oid, oprname, oprkind, oprcode, oprleft, oprright, oprcom, oprnegate, oprrest, oprjoin,
Aug 26 19:10:06 lerami pg-prod[3861]: [40-2] oprcanhash, oprlsortop, oprrsortop, (select usename from pg_user where oprowner = usesysid) as usename from pg_operator
Aug 26 19:10:06 lerami pg-prod[3861]: [41] DEBUG: ProcessQuery
Aug 26 19:10:06 lerami pg-prod[3861]: [42] DEBUG: CommitTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [43] DEBUG: StartTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [44-1] DEBUG: query: SELECT pg_class.oid, relname, relacl, relkind, (select usename from pg_user where relowner = usesysid) as
Aug 26 19:10:06 lerami pg-prod[3861]: [44-2] usename, relchecks, reltriggers, relhasindex, relhasoids from pg_class where relname !~ '^pg_' and relkind in ('r', 'S', 'v')
Aug 26 19:10:06 lerami pg-prod[3861]: [44-3] order by oid
Aug 26 19:10:06 lerami pg-prod[3861]: [45] DEBUG: ProcessQuery
Aug 26 19:10:06 lerami pg-prod[3861]: [46] DEBUG: CommitTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [47] DEBUG: StartTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [48] DEBUG: query: SELECT indexrelid FROM pg_index i WHERE i.indisprimary AND i.indrelid = 20947
Aug 26 19:10:06 lerami pg-prod[3861]: [49] DEBUG: ProcessQuery
Aug 26 19:10:06 lerami pg-prod[3861]: [50] DEBUG: CommitTransactionCommand
Aug 26 19:10:06 lerami pg-prod[3861]: [51] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [448-2] relname = 'pg_class') and objsubid = 9
Aug 26 19:10:07 lerami pg-prod[3861]: [449] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [450] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [451] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [452-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20974 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [452-2] relname = 'pg_class') and objsubid = 10
Aug 26 19:10:07 lerami pg-prod[3861]: [453] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [454] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [455] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [456-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20974 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [456-2] relname = 'pg_class') and objsubid = 0
Aug 26 19:10:07 lerami pg-prod[3861]: [457] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [458] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [459] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [460-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [460-2] relname = 'pg_class') and objsubid = 1
Aug 26 19:10:07 lerami pg-prod[3861]: [461] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [462] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [463] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [464-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [464-2] relname = 'pg_class') and objsubid = 2
Aug 26 19:10:07 lerami pg-prod[3861]: [465] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [466] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [467] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [468-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [468-2] relname = 'pg_class') and objsubid = 3
Aug 26 19:10:07 lerami pg-prod[3861]: [469] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [470] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [471] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [472-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [472-2] relname = 'pg_class') and objsubid = 4
Aug 26 19:10:07 lerami pg-prod[3861]: [473] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [474] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [475] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [476-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [476-2] relname = 'pg_class') and objsubid = 5
Aug 26 19:10:07 lerami pg-prod[3861]: [477] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [478] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [479] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [480-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [480-2] relname = 'pg_class') and objsubid = 6
Aug 26 19:10:07 lerami pg-prod[3861]: [481] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [482] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [483] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [484-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [484-2] relname = 'pg_class') and objsubid = 7
Aug 26 19:10:07 lerami pg-prod[3861]: [485] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [486] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [487] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [488-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [488-2] relname = 'pg_class') and objsubid = 8
Aug 26 19:10:07 lerami pg-prod[3861]: [489] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [490] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [491] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [492-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [492-2] relname = 'pg_class') and objsubid = 9
Aug 26 19:10:07 lerami pg-prod[3861]: [493] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [494] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [495] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [496-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [496-2] relname = 'pg_class') and objsubid = 10
Aug 26 19:10:07 lerami pg-prod[3861]: [497] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [498] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [499] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [500-1] DEBUG: query: SELECT description FROM pg_description WHERE objoid = 20976 and classoid = (SELECT oid FROM pg_class where
Aug 26 19:10:07 lerami pg-prod[3861]: [500-2] relname = 'pg_class') and objsubid = 0
Aug 26 19:10:07 lerami pg-prod[3861]: [501] DEBUG: ProcessQuery
Aug 26 19:10:07 lerami pg-prod[3861]: [502] DEBUG: CommitTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [503] DEBUG: StartTransactionCommand
Aug 26 19:10:07 lerami pg-prod[3861]: [504] DEBUG: query: SELECT proname from pg_proc where pg_proc.oid = '-'::oid
Aug 26 19:10:07 lerami pg-prod[3861]: [505] ERROR: oidin: error in "-": can't parse "-"
Aug 26 19:10:07 lerami pg-prod[3861]: [506] DEBUG: AbortCurrentTransaction
Aug 26 19:10:07 lerami pg-prod[3861]: [507] DEBUG: proc_exit(0)
Aug 26 19:10:07 lerami pg-prod[3861]: [508] DEBUG: shmem_exit(0)
Aug 26 19:10:07 lerami pg-prod[3861]: [509] DEBUG: exit(0)

Any ideas on how I can get my data out?

LER

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

#2Tom Lane
tgl@sss.pgh.pa.us
In reply to: Larry Rosenman (#1)
Re: pg_dump failure, can't get data out...

Larry Rosenman <ler@lerctr.org> writes:

I had upgraded yesterday and *THOUGHT* all was fine.
Forgot to pg_dump, so I restored my $PGLIB and bin directory, now get
the following when I try to pg_dump:

Uh ... you did what exactly? Upgraded from what to what? And what's
this about "restoring the bin directory"? We've forced initdb often
enough in the past weeks that I'd not think the data directory would
be interchangeable across development versions more than a few days
apart...

regards, tom lane

#3Larry Rosenman
ler@lerctr.org
In reply to: Tom Lane (#2)
Re: pg_dump failure, can't get data out...

* Tom Lane <tgl@sss.pgh.pa.us> [010826 19:25]:

Larry Rosenman <ler@lerctr.org> writes:

I had upgraded yesterday and *THOUGHT* all was fine.
Forgot to pg_dump, so I restored my $PGLIB and bin directory, now get
the following when I try to pg_dump:

Uh ... you did what exactly? Upgraded from what to what? And what's
this about "restoring the bin directory"? We've forced initdb often
enough in the past weeks that I'd not think the data directory would
be interchangeable across development versions more than a few days
apart...

I had dumped and initdb'd to yesterdays sources. I then ran for a
while. My backup ran last nite/this am.

I then cvs update'd today, and compiled/installed. Got the initdb
warning (missing the initdb forced, my fault, and stupid).

I then restored my /usr/local/pgsql/bin and /usr/local/pgsql/lib
directory from the tape cut last nite.

I then get the posted error's.

:-(

LER

regards, tom lane

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

#4Tom Lane
tgl@sss.pgh.pa.us
In reply to: Larry Rosenman (#1)
Re: pg_dump failure, can't get data out...

Larry Rosenman <ler@lerctr.org> writes:

pg_dump: query to get function name of oid - failed: ERROR: oidin:
error in "-": can't parse "-"

Actually, I'm seeing it here too ... seems to have been a side-effect
of the change that I recently made to declare pg_index.indproc as
regproc instead of plain Oid. pg_dump is expecting "0" for empty
indproc but it's now getting "-". Hmm, maybe that change wasn't a
good idea; is it likely to break anything besides pg_dump?

Anyway, the quick-hack answer is to change line 4374 in pg_dump.c from

if (strcmp(indinfo[i].indproc, "0") == 0)

to

if (strcmp(indinfo[i].indproc, "-") == 0 ||
strcmp(indinfo[i].indproc, "0") == 0)

You shouldn't need to revert your sources for this --- apply the patch,
compile a new pg_dump, and use it against your yesterday's server.

BTW, I have to congratulate you on your bravery. I sure wouldn't keep
data I cared about in CVS tip ;-)

regards, tom lane

#5Larry Rosenman
ler@lerctr.org
In reply to: Tom Lane (#4)
Re: pg_dump failure, can't get data out...

* Tom Lane <tgl@sss.pgh.pa.us> [010826 19:39]:

Larry Rosenman <ler@lerctr.org> writes:

pg_dump: query to get function name of oid - failed: ERROR: oidin:
error in "-": can't parse "-"

Actually, I'm seeing it here too ... seems to have been a side-effect
of the change that I recently made to declare pg_index.indproc as
regproc instead of plain Oid. pg_dump is expecting "0" for empty
indproc but it's now getting "-". Hmm, maybe that change wasn't a
good idea; is it likely to break anything besides pg_dump?

Anyway, the quick-hack answer is to change line 4374 in pg_dump.c from

if (strcmp(indinfo[i].indproc, "0") == 0)

to

if (strcmp(indinfo[i].indproc, "-") == 0 ||
strcmp(indinfo[i].indproc, "0") == 0)

You shouldn't need to revert your sources for this --- apply the patch,
compile a new pg_dump, and use it against your yesterday's server.

BTW, I have to congratulate you on your bravery. I sure wouldn't keep
data I cared about in CVS tip ;-)

It's not totally irreplaceable, but...

Hey, I find bugs this way.

This fix gets the data out.

NOW, do we need a regression test for pg_dump?

LER

regards, tom lane

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

#6Tom Lane
tgl@sss.pgh.pa.us
In reply to: Larry Rosenman (#5)
Re: pg_dump failure, can't get data out...

Larry Rosenman <ler@lerctr.org> writes:

NOW, do we need a regression test for pg_dump?

We always have and still do (isn't it on the TODO list?). Some Great
Bridge people have been poking at the problem but haven't yet come up
with a clean answer AFAIK. The obvious approach of dumping and
reloading the regression database does not currently work well...

regards, tom lane

#7Larry Rosenman
ler@lerctr.org
In reply to: Tom Lane (#6)
Re: pg_dump failure, can't get data out...

* Tom Lane <tgl@sss.pgh.pa.us> [010826 19:52]:

Larry Rosenman <ler@lerctr.org> writes:

NOW, do we need a regression test for pg_dump?

We always have and still do (isn't it on the TODO list?). Some Great
Bridge people have been poking at the problem but haven't yet come up
with a clean answer AFAIK. The obvious approach of dumping and
reloading the regression database does not currently work well...

I understand. Thanks for the quick looksee and finding this one. :-)

I'll be more careful with making sure pg_dump works before losing
updates and saving the pg_dump output.

regards, tom lane

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

#8Peter Eisentraut
peter_e@gmx.net
In reply to: Larry Rosenman (#1)
Re: pg_dump failure, can't get data out...

Larry Rosenman writes:

pg_dump: query to get function name of oid - failed: ERROR: oidin:
error in "-": can't parse "-"

It's trying to dump a functional index but the function appears to be
absent. (Therefore the name comes out as '-'.) It's hard to tell which
index this is because the query pg_dump uses does not ORDER BY, but if in
doubt you can try the query in pg_dump.c:getIndexes() yourself -- it will
be the first index. Offhand I don't see a related change in pg_dump in
recent times, so it probably isn't necessarily an upgrade related issue,
it might be an inconsistent schema.

It might be easiest to remove the index in question before dumping.

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

#9Larry Rosenman
ler@lerctr.org
In reply to: Peter Eisentraut (#8)
Re: pg_dump failure, can't get data out...

* Peter Eisentraut <peter_e@gmx.net> [010826 20:23]:

Larry Rosenman writes:

pg_dump: query to get function name of oid - failed: ERROR: oidin:
error in "-": can't parse "-"

It's trying to dump a functional index but the function appears to be
absent. (Therefore the name comes out as '-'.) It's hard to tell which
index this is because the query pg_dump uses does not ORDER BY, but if in
doubt you can try the query in pg_dump.c:getIndexes() yourself -- it will
be the first index. Offhand I don't see a related change in pg_dump in
recent times, so it probably isn't necessarily an upgrade related issue,
it might be an inconsistent schema.

It might be easiest to remove the index in question before dumping.

seems to happen on ANY table/index.

I.E. I couldn't dump ANY user DB's.

LER

--
Peter Eisentraut peter_e@gmx.net http://funkturm.homeip.net/~peter

--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749

#10Tom Lane
tgl@sss.pgh.pa.us
In reply to: Peter Eisentraut (#8)
Re: pg_dump failure, can't get data out...

Peter Eisentraut <peter_e@gmx.net> writes:

Offhand I don't see a related change in pg_dump in
recent times, so it probably isn't necessarily an upgrade related issue,
it might be an inconsistent schema.

The proximate cause is that I changed pg_index.indproc from plain "oid"
to "regproc" a week ago, so the output format is different. It seemed
like a good idea at the time ...

I have hacked pg_dump to force the output format back to oid, but I'm
sorta thinking that the schema change was ill-advised because it's
likely to break other applications that look at indproc. Should we
change the column back to oid?

regards, tom lane