Hi,
> > Is there any other option to increase the speed of vacuum?
>
> For autovacuum, decrease "autovacuum_vacuum_cost_delay".
He mentioned in the original message that his problems was with a
global VACUUM after upgrading, so cost_delay should be zero by
default.
Regard
taking more than 5 hours and still running.
>
> Any suggestions for making the process(vacuum analyze) faster are welcome.
Yes, upgrade to PG13.
Kind regards, Martin,
--
Martín Marqués
It’s not that I have something to hide,
it’s that I have nothing I want you to see
/lib
>
> >nm libpq.a | grep sprint
>
> U sprintf
>
> U sprintf
>
> U sprintf
>
> U sprintf
>
> U sprintf
>
> U sprintf
>
>
>
> %:/postgres/12.3/lib >nm libpq.a | grep sprint
>
> U pg_sprintf
>
> U pg_sprintf
>
> U pg_sprintf
here.
Having the output from logs of autovacuum for these tables would give
some insights on where the problem might reside.
--
Martín Marqués
It’s not that I have something to hide,
it’s that I have nothing I want you to see
El 9/1/19 a las 17:38, Ron escribió:
> On 1/9/19 12:19 PM, Martín Marqués wrote:
>> El 9/1/19 a las 14:58, Steve Clark escribió:
>>> Hi List,
>>>
>>> Is there a sure fire way to tell if postgres server is up an
>>> operational. I was testing to see if
nsert it into our
> elasticsearch cluster directly as we get it.
I realy doubt that would work. Aurora doesn't have WALs, so how would
you be able to decode the transactions?
AFAIU, you can't use logical decoding on Aurora. Maybe you should be
asking at the Aurora support channel
e the
> socket was there but my script couldn't read from my database yet.
Use `check_postgres`
--
Martín Marquéshttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
to manage, reason
why that was removed in later versions
Cheers,
--
Martín Marquéshttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
e system.
>
I would start by running amcheck to see if there is index corruption
somewhere, as that can lead to data corruption (in particular if the
corrupted index is a unique or primary key index)
Regards,
--
Martín Marqués
It’s not that I have something to hide,
it’s that I have nothing I want you to see
there, and the primary doesn't have
information of index_scans on other nodes of the cluster.
Regards,
--
Martín Marqués
It’s not that I have something to hide,
it’s that I have nothing I want you to see
yum clean all
yum update postgresql10-server
Regards,
--
Martín Marqués
It’s not that I have something to hide,
it’s that I have nothing I want you to see
in WAL files
BTW, could you point us to the blog you read this?
Maybe you got REDO and UNDO mixed up.
Saludos,
--
Martín Marquéshttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
he first BDR group.
Keep in mind you'll need more slots and replication connections, so
check the values of max_replication_slots, max_worker_processes and
max_wal_senders.
Regards,
--
Martín Marquéshttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
s
nothing to do. The connection has already been assigned to the client
and the waiting is happening on the database server, not the pooler.
Regards,
--
Martín Marquéshttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
you are going to first jump to 9.4, I'd recommend using pglogical
after getting to 9.4 and upgrade straight from 9.4 to 10 (always after
testing your application against 10)
Regards,
--
Martín Marquéshttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
last caveat is a stopper. If the active node is *active* (receiving
writes statements) you'll lose all those changes.
I would instead suggest using pglogical and the
pglogical_create_subscriber tool to create the subscriber from a basebackup.
Kind Regards,
--
Martín Marquéshttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
ll supported versions will work (and some of
the unsupported ones will as well, although it's not recommended to
still be on versions prior to 9.3).
--
Martín Marquéshttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
quot;
Still, the best option, and the most secure, is to use a .pgpass file.
The psql invocation with it's password will very likely end in
.bash_history and alike.
Regards,
--
Martín Marquéshttp://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
18 matches
Mail list logo
