vacuumdb: vacuuming of database "xy" failed: PANIC: corrupted item pointer: 19227

2009/11/18 Tech 2010 <tch072@gmail.com>:

Hello!

How do I location of this pointer and how do I zero it so I can access
the rest of the data?

"zero_damaged_pages = true" did not help in this case, because I
always get same numbers being zeroed. This is with 8.4.0 and 8.4.1.

Thanks.

You probably just need to reindex a table. Try vacuuming each table
individually until you get this error, or just check the vacuum log
output to identify it, then just run a REINDEX on it.

Regards

Thom

On 18 nov., 13:40, thombr...@gmail.com (Thom Brown) wrote:

2009/11/18 Tech 2010 <tch...@gmail.com>:

Hello!

How do I location of this pointer and how do I zero it so I can access
the rest of the data?

"zero_damaged_pages = true" did not help in this case, because I
always get same numbers being zeroed. This is with 8.4.0 and 8.4.1.

Thanks.

You probably just need to reindex a table.  Try vacuuming each table
individually until you get this error, or just check the vacuum log
output to identify it, then just run a REINDEX on it.

xy=# reindex table xy_data;
ERROR: concurrent insert in progress

On 18 nov., 16:48, Tech 2010 <tch...@gmail.com> wrote:

On 18 nov., 13:40, thombr...@gmail.com (Thom Brown) wrote:> 2009/11/18 Tech 2010 <tch...@gmail.com>:

Hello!

How do I location of this pointer and how do I zero it so I can access
the rest of the data?

"zero_damaged_pages = true" did not help in this case, because I
always get same numbers being zeroed. This is with 8.4.0 and 8.4.1.

Thanks.

You probably just need to reindex a table.  Try vacuuming each table
individually until you get this error, or just check the vacuum log
output to identify it, then just run a REINDEX on it.

xy=# reindex table xy_data;
ERROR:  concurrent insert in progress

Should I drop and recreate indexes?

2009/11/19 Tech 2010 <tch072@gmail.com>:

On 18 nov., 16:48, Tech 2010 <tch...@gmail.com> wrote:

On 18 nov., 13:40, thombr...@gmail.com (Thom Brown) wrote:> 2009/11/18 Tech 2010 <tch...@gmail.com>:

Hello!

How do I location of this pointer and how do I zero it so I can access
the rest of the data?

"zero_damaged_pages = true" did not help in this case, because I
always get same numbers being zeroed. This is with 8.4.0 and 8.4.1.

Thanks.

You probably just need to reindex a table.  Try vacuuming each table
individually until you get this error, or just check the vacuum log
output to identify it, then just run a REINDEX on it.

xy=# reindex table xy_data;
ERROR:  concurrent insert in progress

Should I drop and recreate indexes?

That might be a good idea. Use the CREATE INDEX CONCURRENTLY command
to prevent that previous error message from appearing.

Regards

Thom

Tech 2010 <tch072@gmail.com> writes:

xy=# reindex table xy_data;
ERROR: �concurrent insert in progress

Should I drop and recreate indexes?

I think that database is pretty well hosed. (What happened to it to
cause all this, anyway?) A dump/initdb/reload is probably indicated
at this point.

regards, tom lane

On 19 nov., 16:36, t...@sss.pgh.pa.us (Tom Lane) wrote:

Tech 2010 <tch...@gmail.com> writes:

xy=# reindex table xy_data;
ERROR:  concurrent insert in progress

Should I drop and recreate indexes?

I think that database is pretty well hosed.  (What happened to it to
cause all this, anyway?)  A dump/initdb/reload is probably indicated
at this point.

reindex did not help. pg_dump crashes and select * into new_tables
crashes as well. Any idea?

Thanks.

What is this?

PANIC: corrupted item lengths: total 8192, available space 8068

On 29 nov., 07:02, Tech 2010 <tch...@gmail.com> wrote:

What is this?

PANIC:  corrupted item lengths: total 8192, available space 8068

for the others... find the selected row and replace/delete it....