nope ... still get it ... I can do a 'select' from the -O -P backend (what I refer to as single user mode), but not from a normal start up (ie. not pg_dumpable) ...
# select * from email; ERROR: catalog is missing 3 attribute(s) for relid 5773277
'k, in pg_attribute, I find the following for the 'original toast table':
restore=# select * from pg_attribute where attrelid = 8709051;
attrelid | attname | atttypid | attstattarget | attlen | attnum | attndims | attcacheoff | atttypmod | attbyval | attstorage | attisset | attalign | attnotnull | atthasdef | attisdropped | attislocal | attinhcount ----------+------------+----------+---------------+--------+--------+----------+-------------+-----------+----------+------------+----------+----------+------------+-----------+--------------+------------+-------------
8709051 | tableoid | 26 | 0 | 4 | -7 | 0 | -1 | -1 | t | p | f | i | t | f | f | t | 0
8709051 | cmax | 29 | 0 | 4 | -6 | 0 | -1 | -1 | t | p | f | i | t | f | f | t | 0
8709051 | xmax | 28 | 0 | 4 | -5 | 0 | -1 | -1 | t | p | f | i | t | f | f | t | 0
8709051 | cmin | 29 | 0 | 4 | -4 | 0 | -1 | -1 | t | p | f | i | t | f | f | t | 0
8709051 | xmin | 28 | 0 | 4 | -3 | 0 | -1 | -1 | t | p | f | i | t | f | f | t | 0
8709051 | ctid | 27 | 0 | 6 | -1 | 0 | -1 | -1 | f | p | f | i | t | f | f | t | 0
8709051 | chunk_id | 26 | -1 | 4 | 1 | 0 | -1 | -1 | t | p | f | i | f | f | f | t | 0
8709051 | chunk_seq | 23 | -1 | 4 | 2 | 0 | -1 | -1 | t | p | f | i | f | f | f | t | 0
8709051 | chunk_data | 17 | -1 | -1 | 3 | 0 | -1 | -1 | f | p | f | i | f | f | f | t | 0
(9 rows)
which there are more then 3 attributes ... but, of course, there is nothing for the one I just manually added to pg_class ... should this just be duplicated for 5773277?
On Sat, 15 Jan 2005, Marc G. Fournier wrote:
belay that one ... my error, wasn't doin ghte reindex right ... fixed the pg_class issue, but still have one with the email table itself in multi-user:
restore=# select * from email; ERROR: catalog is missing 3 attribute(s) for relid 5773277
but, have an idea I'm going to try one that one ...
On Sat, 15 Jan 2005, Marc G. Fournier wrote:
On Sat, 15 Jan 2005, Tom Lane wrote:
"Marc G. Fournier" <[EMAIL PROTECTED]> writes:Hrmmm ... how about if I pg_dump --oids pg_class, make the modes and then reload it as opg_class, shut down the database server and move opg_class over pg_class, after making the required modifications to opg_class?
Good idea. Give it a shot. Don't forget you'll need to REINDEX pg_class.
Doesn't seem to like that ...
backend> reindex pg_class; ERROR: syntax error at or near "pg_class" at character 9
funny thing is, I can do a 'select * from pg_class' ... and I can do a 'select * from email' now too ... but this is all in single user mode ... do it from 'multi user', and I get:
restore=# select * from email; ERROR: relation "email" does not exist
and:
restore=# select * from pg_class; ERROR: "category_id_remap_key" is an index
seems *damn* close though ... thoughts?
---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: [EMAIL PROTECTED] Yahoo!: yscrappy ICQ: 7615664
---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: [EMAIL PROTECTED] Yahoo!: yscrappy ICQ: 7615664
---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: [EMAIL PROTECTED] Yahoo!: yscrappy ICQ: 7615664
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
