How to troubleshoot: ERROR: cache lookup failed for type?

Started by Marcin Barczynskiover 1 year ago2 messagesgeneral
Jump to latest
#1Marcin Barczynski
mbarczynski@starfishstorage.com

Hello!

When reindexing a table, PostgreSQL returns the following errors for
two of its indexes:

Index foo_idx1 on foo failed with error: failed to find parent tuple
for heap-only tuple at (1162372,1) in table "foo"
Index foo_idx2 on foo failed with error: failed to find parent tuple
for heap-only tuple at (1162372,1) in table "foo"

When trying to show the tuple, there is another error message:

SET enable_indexscan = OFF;
SET enable_tidscan = OFF;
SELECT ctid, xmin, xmax, * FROM sf.dir_current_part_8 WHERE ctid =
'(1162372,1)';

ERROR: cache lookup failed for type 1769235301

Searching for oid=1769235301 in pg_class, pg_type, pg_namespace and
pg_proc returned no results.

It appears to be data corruption.
It's PostgreSQL 13 and unfortunately, data_checksums are turned off.

I would appreciate any suggestions for troubleshooting this issue.

--
Marcin Barczyński

#2Adrian Klaver
adrian.klaver@aklaver.com
In reply to: Marcin Barczynski (#1)
Re: How to troubleshoot: ERROR: cache lookup failed for type?

On 10/4/24 05:09, Marcin Barczyński wrote:

Hello!

When reindexing a table, PostgreSQL returns the following errors for
two of its indexes:

Index foo_idx1 on foo failed with error: failed to find parent tuple
for heap-only tuple at (1162372,1) in table "foo"
Index foo_idx2 on foo failed with error: failed to find parent tuple
for heap-only tuple at (1162372,1) in table "foo"

When trying to show the tuple, there is another error message:

SET enable_indexscan = OFF;
SET enable_tidscan = OFF;
SELECT ctid, xmin, xmax, * FROM sf.dir_current_part_8 WHERE ctid =
'(1162372,1)';

ERROR: cache lookup failed for type 1769235301

Searching for oid=1769235301 in pg_class, pg_type, pg_namespace and
pg_proc returned no results.

It appears to be data corruption.
It's PostgreSQL 13 and unfortunately, data_checksums are turned off.

What is the complete version number e.g. 13.x?

What prompted the reindex of the table?

Have there been any recent hardware/software issues on the machine?

Does the Postgres log have any relevant information?

Can you pg_dump -t sf.dir_current_part_8?

What is the schema definition for sf.dir_current_part_8?

I would appreciate any suggestions for troubleshooting this issue.

--
Marcin Barczyński

--
Adrian Klaver
adrian.klaver@aklaver.com