We have a PostgreSQL DB on Heroku, without warning a few days ago it started logging all data changes as if the log_statement setting was set to mod which is having a massive impact on performance when we do our nightly batch updates.
I've checked the settings in the DB via the pg_settings view and using the Heroku CLI and it very clearly says the the log_statement setting is set to ddl.
We haven't made any changes to the DB or its settings. Does anyone know why this might be happening and how we can reset the log level?
Related
I have inherited a PostgreSQL 13 cluster which was setup using log shipping. I had to make a few changes and would to confirm everything is working well.
Is there a command to make pgsql report whether it is in standby mode or active mode?
Is there a command to make pgsql report up to which WAL file it has applied changes?
I assume that the only way to test the wal shipping is working is to manually modify the active db server and watch for the change on the standby?
I know there are better ways to setup a cluster, but for now I just need to ensure the system remains operational as setup.
From the doc:
select pg_is_in_recovery();
select pg_walfile_name(pg_last_wal_replay_lsn());
Either that or select pg_current_wal_lsn(); on the primary, then select pg_wal_lsn_diff(pg_current_wal_lsn__taken_from_primary,pg_last_wal_replay_lsn()); on standby as commented by #Laurenz Albe.
I'm working with a PostgreSQL server version 10.5 (can't upgrade it at the moment) running in Docker. XL Deploy is connected to it and I upload a new archive that is 232MB. I get the error FATAL: terminating connection due to idle-in-transaction timeout.
In the config, I saw that idle_in_transaction_session_timeout was set to 11000ms. I updated the value to 600000ms and reload the configuration. I see the message LOG: parameter "idle_in_transaction_session_timeout" changed to "600000" in the logs, so I know the setting has taken.
However, when I upload the archive again, I still get the same FATAL timeout message. It's as if the setting is having no effect.
As it turns out, the issue was in Postgres, but not in the config file. It seems you can update the value of idle_in_transaction_session_timeout at various levels. In my case, it was at the ROLE level. Issuing this SQL statement fixed the timeout.
ALTER ROLE role_abc SET idle_in_transaction_session_timeout = '10min';
A slave database was set up some time ago for the purpose of backing up or replicating a remote database. However I can no longer write to the database using a Delphi based ETL (the ETL works for another database pair, but to date has never been used for this particular pair). The replication database was setup by somebody else who has since left the company. I am reasonably sure this has been setup as a replication database, however the employee who has since left told me that replication never worked for unrelated reasons. Using the ETL we can (using SQL queries) read from the one database, and write back to the replication database, Or should be able to, as it is currently read only.
I have tried:
Maintenance such as VACUUM
Attempt to drop tables and the entire database
Restore a full backup from the master database
None of these work, and I am told the database is read-only.
I have looked at postgresql.conf and see that hot_standby is checked, so I think (but am not 100% certain) that the database is in some sort of replication mode (I've never touched replication as supported by Postgres, so I wouldn't know).
I have checked permissions in pg_hba.conf and see there are some credentials in there for replication. I am not sure whether this activates "replication mode" for the database, or simply means these credentials are for replication only.
I have been through months worth of log files (This has not been working since our IT department upgraded the entire network about 5 months ago). I see the log file contents seen below, repeated over and over with nothing else for months. Note the IP address shown below is listed in the pg_hba.conf file, so credentials are valid.
The database is in recovery mode, as I have found by using:
select pg_is_in_recovery();
This explains to me why it's read only, but why can I not restore databases, or just simply dump the entire database and start again (it's a backup so losing/restoring it is not an issue)?
I was tempted to try modifying the recovery.conf file (which exists) but I read/believe that once recovery has been initiated (which in my case it has) modifying the file will have no effect.
I'm using a legacy version of Postgres: 9.2.9
Any help here would be greatly appreciated, as I have been working solidly on this for more than a day now.
Log File entry (sample):
FATAL: could not connect to the primary server:
FATAL: no pg_hba.conf entry for replication connection from host "192.168.20.2", user "postgres", SSL off
FATAL: could not connect to the primary server: server closed the connection unexpectedly
This probably means the server terminated abnormally before or while processing the request.
A couple of options would work for me:
Convert the database from being a read-only replication database, to a standard read/write database or
Dump/drop the entire database so I can create a new one with write capabilities.
It looks like the two database clusters have been set up for replication, but configuration changes on one of the machines broke the replication (changed pg_hba.conf on the primary, changed IP addresses, …).
Here is the way to your desired solutions:
Bringing the standby out of recovery mode: Run
/path/to/pg_ctl promote -D /path/to/data/directory
on the standby as operating system user postgres.
Nuking the standby: Remove the data directory on the standby with rm -rf (or the equivalent on your operating system). Kill all PostgreSQL processes.
Then use initdb to create a new database cluster in the same location.
I am trying to use sails.js on Heroku. When I push my changes, the application starts, but when I try to run MyModel.find() methods, I receive an E_NOTFOUND error. When I log into the database with psql, I see that the tables have not been created automatically. I have the policy in models.js set to migrate: 'drop', so shouldn't I at least get empty tables made when I launch the application? Is there something going on on Heroku that it doesn't like me running sails?
edit:
I had previously been putting settings in development.js and production.js that were different settings (heroku postgresql settings in production). I took that out, and put the settings in connections.js, and it seems like I am able to do the queries and such on the Models, however, when I do heroku pg:psql to connect to my database, I don't see the tables if I do "select * mytable;" it tells me no relation is found.
I believe the problem was that even though I had separate settings for development or production databases, when NODE_ENV was set to production, sails defaults to "safe" and will not let you change it to alter or drop - silently. Changing NODE_ENV to something else allowed the tables to be created.
My company's website uses a PostgreSQL database. In our data center we have a master DB and a few read-only slave DB's, and we use Londiste for continuous replication between them.
I would like to setup another read-only slave DB for reporting purposes, and I'd like this slave to be in a remote location (outside the data center). This slave doesn't need to be 100% up-to-date. If it's up to 24 hours old, that's fine. Also, I'd like to minimize the load I'm putting on the master DB. Since our master DB is busy during the day and idle at night, I figure a good idea (if possible) is to get the reporting slave caught up once each night.
I'm thinking about using log shipping for this, as described on
http://www.postgresql.org/docs/8.4/static/continuous-archiving.html
My plan is:
Setup WAL archiving on the master DB
Produce a full DB snapshot and copy it to the remote location
Restore the DB and get it caught up
Go into steady state where:
DAYTIME -- the DB falls behind but people can query it
NIGHT -- I copy over the day's worth of WAL files and get the DB caught up
Note: the key here is that I only need to copy a full DB snapshot one time. Thereafter I should only have to copy a day's worth of WAL files in order to get the remote slave caught up again.
Since I haven't done log-shipping before I'd like some feedback / advice.
Will this work? Does PostgreSQL support this kind of repeated recovery?
Do you have other suggestions for how to set up a remote semi-fresh read-only slave?
thanks!
--S
Your plan should work.
As Charles says, warm standby is another possible solution. It's supported since 8.2 and has relatively low performance impact on the primary server.
Warm Standby is documented in the Manual: PostgreSQL 8.4 Warm Standby
The short procedure for configuring a
standby server is as follows. For full
details of each step, refer to
previous sections as noted.
Set up primary and standby systems as near identically as possible,
including two identical copies of
PostgreSQL at the same release level.
Set up continuous archiving from the primary to a WAL archive located
in a directory on the standby server.
Ensure that archive_mode,
archive_command and archive_timeout
are set appropriately on the primary
(see Section 24.3.1).
Make a base backup of the primary server (see Section 24.3.2), and load
this data onto the standby.
Begin recovery on the standby server from the local WAL archive,
using a recovery.conf that specifies a
restore_command that waits as
described previously (see Section
24.3.3).
To achieve only nightly syncs, your archive_command should exit with a non-zero exit status during daytime.
Additional Informations:
Postgres Wiki about Warm Standby
Blog Post Warm Standby Setup
9.0's built-in WAL streaming replication is designed to accomplish something that should meet your goals -- a warm or hot standby that can accept read-only queries. Have you considered using it, or are you stuck on 8.4 for now?
(Also, the upcoming 9.1 release is expected to include an updated/rewritten version of pg_basebackup, a tool for creating the initial backup point for a fresh slave.)
Update: PostgreSQL 9.1 will include the ability to pause and resume streaming replication with a simple function call on the slave.