pg.ddx.io  pgsql-admin@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Oliver Duke-Williams <o.w.duke-williams@leeds.ac.uk>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-admin@postgresql.org
Subject: Re: Unable to drop index
Date: Tue, 03 May 2005 12:03:28 +0100
Message-ID: <42775A80.3090103@leeds.ac.uk> (raw)
In-Reply-To: <1935.1114877970@sss.pgh.pa.us>
References: <426E3057.6010804@leeds.ac.uk>
	<1935.1114877970@sss.pgh.pa.us>

Tom Lane wrote:
> 
>>The table on which the index was built is no longer present (it should 
>>have been dropped by my application after being used), but it would 
>>appear that the index was not dropped at the same time as the table:
> 
> 
> That's fairly strange; it should be impossible, in fact, because of the
> pg_depend mechanism.
> 
> 
>>drop index iext_41827440000_24 ;
>>ERROR:  could not open relation with OID 147483811
> 
> 
>>Looking in pg_class for an entry with this OID (147483811) returns 0 
>>rows; an entry exists in pg_class for the index itself.
> 
> 
> What is the OID of the index itself?  Are there any rows in pg_depend
> matching either the index OID or 147483811 in either objid or refobjid?

The index has OID 149592817; there are two rows in pg_depend relating to 
the index as a object dependent on the table:

zzwicid=# select * from pg_depend where objid =  149592817 or refobjid = 
147483811;
classid|objid|objsubid|refclassid|refobjid|refobjsubid|deptype
1259|149592817|0|1259|147483811|1|a
1259|149592817|0|1259|147483811|3|a

The index was on two fields, so this is as expected.

> How about pg_index (see indexrelid and indrelid respectively)?

Yes

>  How about pg_attribute (see attrelid)?  

Yes - two rows corresponding to the index.

> Is the underlying file (named by pg_class.relfilenode) still there?

Yes
> 
>>How can I safely remove the index?  The old copy of the index is 
>>preventing the creation of a new table and associated index of the same 
>>names.
> 
> 
> You can probably just "rm" the underlying file and DELETE the pg_class
> row plus any other rows you found above.  However I wonder what other
> corruption may have occurred in whatever event produced this situation.
> I'm thinking you might have lost a whole page of pg_class, for example.
> It'd likely be a good idea to see if you can pg_dump and reload the
> database.

OK - will dump / reload the database and then try dropping the index 
manually.

Thanks for your reply,

Oliver





view thread (3+ messages)

Message-ID: <42775A80.3090103@leeds.ac.uk>
Permalink:  ../42775A80.3090103@leeds.ac.uk/
Also on:    postgresql.org/message-id/42775A80.3090103@leeds.ac.uk

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients from your mail client; the link fills in
  the headers below: reply via email
  Some clients, webmail especially, drop In-Reply-To and References from
  a link, and the reply then starts a new thread: use the mbox below.

  To: pgsql-admin@postgresql.org
  Cc: o.w.duke-williams@leeds.ac.uk, tgl@sss.pgh.pa.us
  Subject: Re: Unable to drop index
  In-Reply-To: <42775A80.3090103@leeds.ac.uk>
  References: <426E3057.6010804@leeds.ac.uk> <1935.1114877970@sss.pgh.pa.us> <42775A80.3090103@leeds.ac.uk>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox