Thread: Re: BUG #3245: PANIC: failed to re-find shared loc k ob je ct

Re: BUG #3245: PANIC: failed to re-find shared loc k ob je ct

From
"Dorochevsky,Michel"
Date:
> Von: Tom Lane [mailto:tgl@sss.pgh.pa.us]
> So, which of your tables has OID 417227?  Easiest way to check is
>    select 417227::regclass
>
Tom,

Currently a table with name
 requireditemtypes_mill7b7c7621
has the object id 417227.

I did restart the DB server in the meantime several times.
I assume that will not change OIDs of existing tables.

I am not 100 percent sure whether I dropped and recreated all tables in the
meantime. Would dropping and recreating the tables reuse old OIDs that
became free? If so the OID might got assigned to another table.

I certainly did not drop the schema in the meantime.

Shall we send you the database?

Best Regards
--  Michel

Re: BUG #3245: PANIC: failed to re-find shared loc k ob je ct

From
Tom Lane
Date:
"Dorochevsky,Michel" <michel.dorochevsky@softcon.de> writes:
>> Von: Tom Lane [mailto:tgl@sss.pgh.pa.us]
>> So, which of your tables has OID 417227?  Easiest way to check is
>> select 417227::regclass

> Currently a table with name
>  requireditemtypes_mill7b7c7621
> has the object id 417227.

Okay, and that table has been referenced within the failing transaction,
so we can conclude that there's nothing corrupt about the locktag.  It's
just been selected from, AFAICS, so we can also eliminate the idea that
there's something weird about the way you are using that particular
table.  Which leaves us with ... hmmm ... not a lot of ideas.  Perhaps
the LOCK_DEBUG trace will suggest something.

            regards, tom lane