Re: table not shown

Hi,

Just wondering if we can bring up PG after we stop it and
reindex database in the case DD corrupted?

That is our worry.

Thanks,

-----Original Message-----
From: Lee Wu
Sent: Tuesday, June 15, 2004 2:25 PM
To: 'Tom Lane'
Cc: pgsql-admin@postgresql.org
Subject: RE: [ADMIN] table not shown

I think data dictionary got corrupted:

mxl=# \d mxl_user_wbl
                               Table "public.mxl_user_wbl"
 Column  |           Type           |                     Modifiers
---------+--------------------------+-----------------------------------
-----------------
 user_id | integer                  | not null
 sender  | character varying(128)   | not null
 action  | integer                  |
 created | timestamp with time zone | default
('now'::text)::timestamp(6) with time zone
Indexes: mxl_user_wbl_pkey primary key btree (user_wbl_id),
         mxl_user_wbl_uid_action_idx btree (user_id, "action")
Foreign Key constraints: mxl_user_wbl_uid_fkey FOREIGN KEY (user_id)
REFERENCES mxl_user(user_id) ON UPDATE NO ACTION ON DELETE CASCADE
Triggers: mxl_user_wbl_u_trg

mxl=# select * from mxl_user_wbl where 1=2;
 user_wbl_id | user_id | sender | action | created
-------------+---------+--------+--------+---------
(0 rows)

Note: user_wbl_id is our primary key as from the scripts:
"user_wbl_id" integer DEFAULT nextval('mxl_quar_id_seq'::text) PRIMARY
KEY,

From this database's pg_attribute:
select a.attname from pg_attribute a, pg_class b
where b.relname = 'mxl_user_wbl' and a.attrelid = b.oid;
attname
----------
 tableoid
 cmax
 xmax
 cmin
 xmin
 oid
 ctid
 user_id
 sender
 action
 created
(11 rows)

Tom, I will reindex database to see what happens when time permits.

Thanks,


-----Original Message-----
From: Tom Lane [mailto:tgl@sss.pgh.pa.us]
Sent: Tuesday, June 15, 2004 1:22 PM
To: Lee Wu
Cc: pgsql-admin@postgresql.org
Subject: Re: [ADMIN] table not shown

"Lee Wu" <Lwu@mxlogic.com> writes:
> Here is result:
> mxl=# select * from pg_catalog.pg_class where relname =
> 'mxl_quar_process';
> [ still no rows ]

Okay.  I was wondering about bizarre ideas like a non-system table named
pg_class, but that seems ruled out now.  I think you're down to the
REINDEX.  Good luck!

            regards, tom lane