| From: | Andres Freund <andres(at)anarazel(dot)de> |
|---|---|
| To: | Laurenz Albe <laurenz(dot)albe(at)cybertec(dot)at> |
| Cc: | Rambabu V <ram(dot)wissen(at)gmail(dot)com>, pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: Bug in PostgreSQL |
| Date: | 2018-06-28 16:13:16 |
| Message-ID: | 20180628161316.jbzzo76kzcmczh2r@alap3.anarazel.de |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
Hi,
On 2018-06-28 08:55:21 +0200, Laurenz Albe wrote:
> Rambabu V wrote:
> > While taking pgdump we are getting error message cache lookup failed for function 7418447.
> > While trying select * from pg_proc where oid=7418447 returns zero rows. Please help us on this.
>
> That means that some catalog data are corrupted.
It does *NOT* have mean that. You can get such reports e.g. because
there was concurrent DDL. Most things are protected against via locks,
but there's enough of a window between getting the list of objects and
locking to cause issues for tables. And for functions it's fairly easy
to get into trouble because there's a mismatch between the snapshot
pg_dump uses (a normal transactional snapshot) and the snapshot used to
deparse expressions etc (a fresh catalog snapshot that takes into
account concurrent commits).
Greetings,
Andres Freund
| From | Date | Subject | |
|---|---|---|---|
| Next Message | AJG | 2018-07-01 09:31:24 | Trigger overhead/performance and alternatives? |
| Previous Message | Laurenz Albe | 2018-06-28 06:55:21 | Re: Bug in PostgreSQL |