https://github.com/dgapitts/vagrant-pg96-slony

Last synced: over 1 year ago
JSON representation
Host: GitHub
URL: https://github.com/dgapitts/vagrant-pg96-slony
Owner: dgapitts
License: gpl-3.0
Created: 2021-01-21T10:48:20.000Z (over 5 years ago)
Default Branch: main
Last Pushed: 2021-01-23T20:16:45.000Z (over 5 years ago)
Last Synced: 2025-01-08T22:38:52.931Z (over 1 year ago)
Language: Shell
Size: 42 KB
Stars: 0
Watchers: 3
Forks: 0
Open Issues: 0
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project

README

          ## Summary

Training environment for learning how to setup slony replication.

*I've got this setup and working now*, after reading the slony overview document from dalibo (https://public.dalibo.com/exports/conferences/_archives/_2011/201110_slony/pgconfeu_slony.pdf).

The last two things I've done

* set listen_addresses='*' in postgres.conf

* added final setup_slony_master_and_slave.sh (before i.e. below, I was only slarting the master slon process - hence the connectivity problems I was getting below )

```

~/projects/vagrant-pg96-slony $ cat setup_slony_master_and_slave.sh

. ~/.bash_profile

mkdir /tmp/master

cd /tmp/master

/usr/pgsql-9.6/bin/slonik /vagrant/init_master.slonik &

sleep 2

mkdir /tmp/slave

cd /tmp/slave

/usr/pgsql-9.6/bin/slonik /vagrant/init_slave.slonik &

sleep 2

cd /tmp/master

nohup slon slony_example "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" &

sleep 2

cd /tmp/slave

nohup slon slony_example "dbname=pgbenchslave  host=localhost port=5432 user=slonyrep password=changeme" &

```

Note: this is 99% scripted i.e. you should just need to type vagrant up, although I did need to start the master and slave slon processes manually i.e. for some reason the called embedded within the setup_slony_master_and_slave.sh are not working (maybe sort sort of missing env parameter with shell scripts call shell scripts) 

```

~/projects/vagrant-pg96-slony $ vagrant ssh

Last login: Fri Jan 22 17:17:25 2021

[pg96slony:vagrant:~] # pg

Last login: Fri Jan 22 17:17:39 UTC 2021

[pg96slony:postgres:~] #  PGPASSWORD=changeme psql -U slonyrep -h localhost -p 5432 -d pgbenchslave -c "select count(*) from pgbench_branches;"

 count

-------

     0

(1 row)

[pg96slony:postgres:~] # ps -ef|grep slon

postgres  3944  3895  0 17:19 pts/0    00:00:00 grep --color=auto slon

[pg96slony:postgres:~] # cat /tmp/setup_slony_master_and_slave.sh

. ~/.bash_profile

mkdir /tmp/master

cd /tmp/master

/usr/pgsql-9.6/bin/slonik /vagrant/init_master.slonik &

sleep 2

mkdir /tmp/slave

cd /tmp/slave

/usr/pgsql-9.6/bin/slonik /vagrant/init_slave.slonik &

sleep 2

cd /tmp/master

nohup slon slony_example "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" &

sleep 2

cd /tmp/slave

nohup slon slony_example "dbname=pgbenchslave  host=localhost port=5432 user=slonyrep password=changeme" &

[pg96slony:postgres:~] # cd /tmp/master

[pg96slony:postgres:/tmp/master] # lt

total 0

[pg96slony:postgres:/tmp/master] # cd /tmp/master

[pg96slony:postgres:/tmp/master] # nohup slon slony_example "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" &

[1] 3968

[pg96slony:postgres:/tmp/master] # nohup: ignoring input and appending output to ‘nohup.out’

[pg96slony:postgres:/tmp/master] # cd /tmp/slave

[pg96slony:postgres:/tmp/slave] # nohup slon slony_example "dbname=pgbenchslave  host=localhost port=5432 user=slonyrep password=changeme" &

[2] 3985

[pg96slony:postgres:/tmp/slave] # nohup: ignoring input and appending output to ‘nohup.out’

[pg96slony:postgres:/tmp/slave] #  PGPASSWORD=changeme psql -U slonyrep -h localhost -p 5432 -d pgbenchslave -c "select count(*) from pgbench_branches;"

 count

-------

     1

(1 row)

[pg96slony:postgres:/tmp/slave] 
postgres  3968  3895  0 17:20 pts/0 
postgres  3969  3968  0 17:20 pts/0 
postgres  3973  3801  0 17:20 ? 
postgres  3979  3801  0 17:20 ? 
postgres  3980  3801  0 17:20 ? 
postgres  3981  3801  0 17:20 ? 
postgres  3982  3801  0 17:20 ? 
postgres  3983  3801  0 17:20 ? 
postgres  3985  3895  0 17:20 pts/0 
postgres  3986  3985  0 17:20 pts/0 
postgres  3990  3801  0 17:20 ? 
postgres  3996  3801  0 17:20 ? 
postgres  3997  3801  0 17:20 ? 
postgres  3998  3801  0 17:20 ? 
postgres  3999  3801  0 17:20 ? 
postgres  4000  3801  0 17:20 ? 
postgres  4002  3801  0 17:20 ? 
postgres  4003  3801  0 17:20 ? 
postgres  4021  3895  0 17:21 pts/0 
[pg96slony:postgres:/tmp/slave] 
 count 
------- 
     1 
(1 row) 
```

# ps -ef|grep slon 00:00:00 slon slony_example dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme 00:00:00 slon slony_example dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme 00:00:00 postgres: slonyrep pgbench ::1(47480) idle 00:00:00 postgres: slonyrep pgbenchslave ::1(47482) idle 00:00:00 postgres: slonyrep pgbench ::1(47484) idle 00:00:00 postgres: slonyrep pgbench ::1(47486) idle 00:00:00 postgres: slonyrep pgbench ::1(47488) idle 00:00:00 postgres: slonyrep pgbench ::1(47490) idle 00:00:00 slon slony_example dbname=pgbenchslave  host=localhost port=5432 user=slonyrep password=changeme 00:00:00 slon slony_example dbname=pgbenchslave  host=localhost port=5432 user=slonyrep password=changeme 00:00:00 postgres: slonyrep pgbenchslave ::1(47494) idle 00:00:00 postgres: slonyrep pgbench ::1(47496) idle 00:00:00 postgres: slonyrep pgbenchslave ::1(47498) idle 00:00:00 postgres: slonyrep pgbenchslave ::1(47500) idle 00:00:00 postgres: slonyrep pgbenchslave ::1(47502) idle 00:00:00 postgres: slonyrep pgbenchslave ::1(47504) idle 00:00:00 postgres: slonyrep pgbench ::1(47508) idle 00:00:00 postgres: slonyrep pgbenchslave ::1(47510) idle 00:00:00 grep --color=auto slon # psql -U slonyrep -h localhost -p 5432 -d pgbenchslave -c "select count(*) from pgbench_branches;"

## Useful documentation for slony (reference)

* https://www.slony.info/documentation/2.1/tutorial.html

* https://www.slony.info/documentation/stmtsetaddtable.html

* https://sites.google.com/site/pgsqldoc/Home/slony

* https://hub.packtpub.com/replication-solutions-postgresql/

* https://severalnines.com/blog/experts-guide-slony-replication-postgresql

* https://serverfault.com/questions/198002/postgresql-what-does-grant-all-privileges-on-database-d

* https://public.dalibo.com/exports/conferences/_archives/_2011/201110_slony/pgconfeu_slony.pdf 

## Earlier issues - setup seems almost complete although still hitting connectivity issues between master and slave 

To install slony we first need to install perl

```

  yum install -y perl

  yum install -y slony1-96

```

Initial DB setup ... 

```

  su -c "/usr/pgsql-12/bin/pgbench -i -s 1" -s /bin/sh vagrant

  su -c "pg_dump -s -p 5432 -h localhost pgbench | psql -h localhost -p 5432 pgbenchslave" -s /bin/sh vagrant

```

I'm struggling to get slony up and running, but making notes of all the challenges/steps I'm working through, this is a learning exercise...

* I've added md5 .pgpass entries as per https://www.slony.info/documentation/security.html and https://gazelle.ihe.net/content/adding-slave

* I also updated pg_hba.conf entries

* However the initiallisation following commanads only partially work

```

First  Session:

[pg96slony:vagrant:~] # pg

Last login: Thu Jan 21 21:02:58 UTC 2021

[pg96slony:postgres:~] # /usr/pgsql-9.6/bin/slonik /vagrant/init_master.slonik

[pg96slony:postgres:~] # /usr/pgsql-9.6/bin/slonik /vagrant/init_slave.slonik

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

Second Session

[pg96slony:postgres:~] #  nohup slon slony_example "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" &

[1] 4100

[pg96slony:postgres:~] # nohup: ignoring input and appending output to ‘nohup.out’

[pg96slony:postgres:~] #

[pg96slony:postgres:~] # tail -f nohup.out

2021-01-21 21:13:18 UTC CONFIG version for "dbname=pgbenchslave host=localhost port=5432 user=slonyrep password=changeme" is 90620

2021-01-21 21:13:18 UTC CONFIG version for "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" is 90620

2021-01-21 21:13:18 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 21:13:18 UTC CONFIG cleanupThread: bias = 60

2021-01-21 21:13:18 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 21:13:18 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 21:13:20 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 21:13:20 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 21:13:20 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 21:13:20 UTC CONFIG remoteWorkerThread_2: update provider configuration

^C

[pg96slony:postgres:~] # tail -1000f nohup.out

2021-01-21 21:13:18 UTC CONFIG main: slon version 2.2.10 starting up

2021-01-21 21:13:18 UTC INFO   slon: watchdog process started

2021-01-21 21:13:18 UTC CONFIG slon: watchdog ready - pid = 4100

2021-01-21 21:13:18 UTC CONFIG slon: worker process created - pid = 4101

2021-01-21 21:13:18 UTC CONFIG main: Integer option vac_frequency = 3

2021-01-21 21:13:18 UTC CONFIG main: Integer option log_level = 0

2021-01-21 21:13:18 UTC CONFIG main: Integer option sync_interval = 2000

2021-01-21 21:13:18 UTC CONFIG main: Integer option sync_interval_timeout = 10000

2021-01-21 21:13:18 UTC CONFIG main: Integer option sync_group_maxsize = 20

2021-01-21 21:13:18 UTC CONFIG main: Integer option syslog = 0

2021-01-21 21:13:18 UTC CONFIG main: Integer option quit_sync_provider = 0

2021-01-21 21:13:18 UTC CONFIG main: Integer option remote_listen_timeout = 300

2021-01-21 21:13:18 UTC CONFIG main: Integer option monitor_interval = 500

2021-01-21 21:13:18 UTC CONFIG main: Integer option explain_interval = 0

2021-01-21 21:13:18 UTC CONFIG main: Integer option tcp_keepalive_idle = 0

2021-01-21 21:13:18 UTC CONFIG main: Integer option tcp_keepalive_interval = 0

2021-01-21 21:13:18 UTC CONFIG main: Integer option tcp_keepalive_count = 0

2021-01-21 21:13:18 UTC CONFIG main: Integer option apply_cache_size = 100

2021-01-21 21:13:18 UTC CONFIG main: Boolean option log_pid = 0

2021-01-21 21:13:18 UTC CONFIG main: Boolean option log_timestamp = 1

2021-01-21 21:13:18 UTC CONFIG main: Boolean option tcp_keepalive = 1

2021-01-21 21:13:18 UTC CONFIG main: Boolean option monitor_threads = 1

2021-01-21 21:13:18 UTC CONFIG main: Boolean option enable_version_check = 1

2021-01-21 21:13:18 UTC CONFIG main: Boolean option remote_listen_serializable_transactions = 1

2021-01-21 21:13:18 UTC CONFIG main: Real option real_placeholder = 0.000000

2021-01-21 21:13:18 UTC CONFIG main: String option cluster_name = slony_example

2021-01-21 21:13:18 UTC CONFIG main: String option conn_info = dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme

2021-01-21 21:13:18 UTC CONFIG main: String option pid_file = [NULL]

2021-01-21 21:13:18 UTC CONFIG main: String option log_timestamp_format = %Y-%m-%d %H:%M:%S %Z

2021-01-21 21:13:18 UTC CONFIG main: String option archive_dir = [NULL]

2021-01-21 21:13:18 UTC CONFIG main: String option sql_on_connection = [NULL]

2021-01-21 21:13:18 UTC CONFIG main: String option lag_interval = [NULL]

2021-01-21 21:13:18 UTC CONFIG main: String option command_on_logarchive = [NULL]

2021-01-21 21:13:18 UTC CONFIG main: String option syslog_facility = LOCAL0

2021-01-21 21:13:18 UTC CONFIG main: String option syslog_ident = slon

2021-01-21 21:13:18 UTC CONFIG main: String option cleanup_interval = 10 minutes

2021-01-21 21:13:18 UTC CONFIG main: local node id = 1

2021-01-21 21:13:18 UTC INFO   main: main process started

2021-01-21 21:13:18 UTC CONFIG main: launching sched_start_mainloop

2021-01-21 21:13:18 UTC CONFIG main: loading current cluster configuration

2021-01-21 21:13:18 UTC CONFIG storeNode: no_id=2 no_comment='Slave node'

2021-01-21 21:13:18 UTC CONFIG storePath: pa_server=2 pa_client=1 pa_conninfo="dbname=pgbenchslave host=localhost port=5432 user=slonyrep password=changeme" pa_connretry=10

2021-01-21 21:13:18 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 21:13:18 UTC CONFIG storeSet: set_id=1 set_origin=1 set_comment='All pgbench tables'

2021-01-21 21:13:18 UTC CONFIG main: last local event sequence = 5000000013

2021-01-21 21:13:18 UTC CONFIG main: configuration complete - starting threads

2021-01-21 21:13:18 UTC INFO   localListenThread: thread starts

2021-01-21 21:13:18 UTC CONFIG version for "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" is 90620

2021-01-21 21:13:18 UTC CONFIG enableNode: no_id=2

2021-01-21 21:13:18 UTC INFO   main: running scheduler mainloop

2021-01-21 21:13:18 UTC INFO   remoteListenThread_2: thread starts

2021-01-21 21:13:18 UTC CONFIG cleanupThread: thread starts

2021-01-21 21:13:18 UTC INFO   syncThread: thread starts

2021-01-21 21:13:18 UTC INFO   monitorThread: thread starts

2021-01-21 21:13:18 UTC INFO   remoteWorkerThread_2: thread starts

2021-01-21 21:13:18 UTC CONFIG version for "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" is 90620

2021-01-21 21:13:18 UTC CONFIG version for "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" is 90620

2021-01-21 21:13:18 UTC CONFIG version for "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" is 90620

2021-01-21 21:13:18 UTC CONFIG version for "dbname=pgbenchslave host=localhost port=5432 user=slonyrep password=changeme" is 90620

2021-01-21 21:13:18 UTC CONFIG version for "dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme" is 90620

2021-01-21 21:13:18 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 21:13:18 UTC CONFIG cleanupThread: bias = 60

2021-01-21 21:13:18 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 21:13:18 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 21:13:20 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 21:13:20 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 21:13:20 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 21:13:20 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 21:15:40 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 21:15:40 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 21:19:26 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

```

* Partially work in that they setup the _slony_example schema and added triggers

```

[pg96slony:postgres:~] # psql -d pgbench

psql (9.6.20)

pgbench=# \d+ pgbench_accounts

                       Table "public.pgbench_accounts"

  Column  |     Type      | Modifiers | Storage  | Stats target | Description

----------+---------------+-----------+----------+--------------+-------------

 aid      | integer       | not null  | plain    |              |

 bid      | integer       |           | plain    |              |

 abalance | integer       |           | plain    |              |

 filler   | character(84) |           | extended |              |

Indexes:

    "pgbench_accounts_pkey" PRIMARY KEY, btree (aid)

Triggers:

    _slony_example_logtrigger AFTER INSERT OR DELETE OR UPDATE ON pgbench_accounts FOR EACH ROW EXECUTE PROCEDURE _slony_example.logtrigger('_slony_example', '1', 'k')

    _slony_example_truncatetrigger BEFORE TRUNCATE ON pgbench_accounts FOR EACH STATEMENT EXECUTE PROCEDURE _slony_example.log_truncate('1')

Disabled user triggers:

    _slony_example_denyaccess BEFORE INSERT OR DELETE OR UPDATE ON pgbench_accounts FOR EACH ROW EXECUTE PROCEDURE _slony_example.denyaccess('_slony_example')

    _slony_example_truncatedeny BEFORE TRUNCATE ON pgbench_accounts FOR EACH STATEMENT EXECUTE PROCEDURE _slony_example.deny_truncate()

Options: fillfactor=100

[pg96slony:postgres:~] # psql -U slonyrep -h localhost -p 5432 -d pgbench

psql (9.6.20)

Type "help" for help.

pgbench=# \dn

      List of schemas

      Name      |  Owner

----------------+----------

 _slony_example | slonyrep

 public         | postgres

(2 rows)

pgbench=# SET search_path TO _slony_example, public;

SET

pgbench=# \x

Expanded display is on.

pgbench=# select * from sl_table;

-[ RECORD 1 ]----------------------

tab_id      | 1

tab_reloid  | 16393

tab_relname | pgbench_accounts

tab_nspname | public

tab_set     | 1

tab_idxname | pgbench_accounts_pkey

tab_altered | f

tab_comment | accounts table

-[ RECORD 2 ]----------------------

tab_id      | 2

tab_reloid  | 16396

tab_relname | pgbench_branches

tab_nspname | public

tab_set     | 1

tab_idxname | pgbench_branches_pkey

tab_altered | f

tab_comment | branches table

-[ RECORD 3 ]----------------------

tab_id      | 3

tab_reloid  | 16390

tab_relname | pgbench_tellers

tab_nspname | public

tab_set     | 1

tab_idxname | pgbench_tellers_pkey

tab_altered | f

tab_comment | tellers table

-[ RECORD 4 ]----------------------

tab_id      | 4

tab_reloid  | 16387

tab_relname | pgbench_history

tab_nspname | public

tab_set     | 1

tab_idxname | pgbench_history_pkey

tab_altered | f

tab_comment | history table

```

However the actual replication seems to be hanging 

```

[pg96slony:postgres:~] # PGPASSWORD=changeme psql -U slonyrep -h localhost -p 5432 -d pgbenchslave -c "select count(*) from pgbench_branches;"

 count

-------

     0

(1 row)

[pg96slony:postgres:~] # PGPASSWORD=changeme psql -U slonyrep -h localhost -p 5432 -d pgbench -c "select count(*) from pgbench_branches;"

 count

-------

     1

(1 row)

```

I suspect connectivity issues

```

postgres=# \du

                                   List of roles

 Role name |                         Attributes                         | Member of

-----------+------------------------------------------------------------+-----------

 postgres  | Superuser, Create role, Create DB, Replication, Bypass RLS | {}

 slonyrep  | Superuser                                                  | {}

 vagrant   |                                                            | {}

```

NB I also trying to work through this example (https://hub.packtpub.com/replication-solutions-postgresql/#more) and using the postgres user

```

[pg96slony:postgres:~] # createdb test1

[pg96slony:postgres:~] # createdb test2

[pg96slony:postgres:~] # psql -d test1

psql (9.6.20)

Type "help" for help.

test1=# create table t_test (id numeric primary key, name

test1(#   varchar);

CREATE TABLE

test1=# insert into t_test values(1,'A'),(2,'B'), (3,'C');

INSERT 0 3

test1=#

test1=# \q

[pg96slony:postgres:~] # psql

psql (9.6.20)

Type "help" for help.

postgres=# alter user postgres with password 'changeme';

ALTER ROLE

postgres=# \q

[pg96slony:postgres:~] # vi ~/.pgpass

[pg96slony:postgres:~] # pg_dump -s -p 5432 -h localhost test1 | psql -h localhost -p 5432 test2

SET

..

CREATE EXTENSION

COMMENT

SET

SET

CREATE TABLE

ALTER TABLE

ALTER TABLE

[pg96slony:postgres:~] # psql -d  test2

psql (9.6.20)

Type "help" for help.

test2=# select * from t_test;

 id | name

----+------

(0 rows)

test2=# \q

[pg96slony:postgres:~] # head -20 init_*

==> init_master2.slonik <==

#! /bin/slonik

  cluster name = mycluster;

  node 1 admin conninfo = 'dbname=test1 host=localhost

port=5432 user=postgres password=changeme';

  node 2 admin conninfo = 'dbname=test2 host=localhost

port=5432 user=postgres password=changeme';

  init cluster ( id=1);

  create set (id=1, origin=1);

  set add table(set id=1, origin=1, id=1, fully qualified

name = 'public.t_test');

  store node (id=2, event node = 1);

  store path (server=1, client=2, conninfo='dbname=test1

host=localhost port=5432 user=postgres password=changeme');

  store path (server=2, client=1, conninfo='dbname=test2

host=localhost port=5432 user=postgres password=changeme');

  store listen (origin=1, provider = 1, receiver = 2);

  store listen (origin=2, provider = 2, receiver = 1);

==> init_slave2.slonik <==

#! /bin/slonik

  cluster name = mycluster;

  node 1 admin conninfo = 'dbname=test1 host=localhost

port=5432 user=postgres password=changeme';

  node 2 admin conninfo = 'dbname=test2 host=localhost

port=5432 user=postgres password=changeme';

subscribe set ( id = 1, provider = 1, receiver = 2,

  forward = no);

```  

but similar results i.e. the triggers and _mycluster (for slony metadata) are added but no data is replicated

```

[pg96slony:postgres:~] # slonik init_master2.slonik

init_master2.slonik:16: waiting for event (1,5000000007) to be confirmed on node 2

init_master2.slonik:16: waiting for event (1,5000000007) to be confirmed on node 2

init_master2.slonik:16: waiting for event (1,5000000007) to be confirmed on node 2

init_master2.slonik:16: waiting for event (1,5000000007) to be confirmed on node 2

[pg96slony:postgres:~] # slonik init_slave2.slonik

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

waiting for events  (2,5000000001) only at (2,0) to be confirmed on node 1

[pg96slony:postgres:~] # nohup slon mycluster "dbname=test1 host=localhost port=5432

>   user=postgres password=changeme" &

[1] 8057

[pg96slony:postgres:~] # nohup: ignoring input and appending output to ‘nohup.out’

[pg96slony:postgres:~] # tail -1000f nohup.out

2021-01-21 23:37:05 UTC CONFIG main: slon version 2.2.10 starting up

2021-01-21 23:37:05 UTC INFO   slon: watchdog process started

2021-01-21 23:37:05 UTC CONFIG slon: watchdog ready - pid = 8057

2021-01-21 23:37:05 UTC CONFIG slon: worker process created - pid = 8058

2021-01-21 23:37:05 UTC CONFIG main: Integer option vac_frequency = 3

2021-01-21 23:37:05 UTC CONFIG main: Integer option log_level = 0

2021-01-21 23:37:05 UTC CONFIG main: Integer option sync_interval = 2000

2021-01-21 23:37:05 UTC CONFIG main: Integer option sync_interval_timeout = 10000

2021-01-21 23:37:05 UTC CONFIG main: Integer option sync_group_maxsize = 20

2021-01-21 23:37:05 UTC CONFIG main: Integer option syslog = 0

2021-01-21 23:37:05 UTC CONFIG main: Integer option quit_sync_provider = 0

2021-01-21 23:37:05 UTC CONFIG main: Integer option remote_listen_timeout = 300

2021-01-21 23:37:05 UTC CONFIG main: Integer option monitor_interval = 500

2021-01-21 23:37:05 UTC CONFIG main: Integer option explain_interval = 0

2021-01-21 23:37:05 UTC CONFIG main: Integer option tcp_keepalive_idle = 0

2021-01-21 23:37:05 UTC CONFIG main: Integer option tcp_keepalive_interval = 0

2021-01-21 23:37:05 UTC CONFIG main: Integer option tcp_keepalive_count = 0

2021-01-21 23:37:05 UTC CONFIG main: Integer option apply_cache_size = 100

2021-01-21 23:37:05 UTC CONFIG main: Boolean option log_pid = 0

2021-01-21 23:37:05 UTC CONFIG main: Boolean option log_timestamp = 1

2021-01-21 23:37:05 UTC CONFIG main: Boolean option tcp_keepalive = 1

2021-01-21 23:37:05 UTC CONFIG main: Boolean option monitor_threads = 1

2021-01-21 23:37:05 UTC CONFIG main: Boolean option enable_version_check = 1

2021-01-21 23:37:05 UTC CONFIG main: Boolean option remote_listen_serializable_transactions = 1

2021-01-21 23:37:05 UTC CONFIG main: Real option real_placeholder = 0.000000

2021-01-21 23:37:05 UTC CONFIG main: String option cluster_name = mycluster

2021-01-21 23:37:05 UTC CONFIG main: String option conn_info = dbname=test1 host=localhost port=5432

  user=postgres password=changeme

2021-01-21 23:37:05 UTC CONFIG main: String option pid_file = [NULL]

2021-01-21 23:37:05 UTC CONFIG main: String option log_timestamp_format = %Y-%m-%d %H:%M:%S %Z

2021-01-21 23:37:05 UTC CONFIG main: String option archive_dir = [NULL]

2021-01-21 23:37:05 UTC CONFIG main: String option sql_on_connection = [NULL]

2021-01-21 23:37:05 UTC CONFIG main: String option lag_interval = [NULL]

2021-01-21 23:37:05 UTC CONFIG main: String option command_on_logarchive = [NULL]

2021-01-21 23:37:05 UTC CONFIG main: String option syslog_facility = LOCAL0

2021-01-21 23:37:05 UTC CONFIG main: String option syslog_ident = slon

2021-01-21 23:37:05 UTC CONFIG main: String option cleanup_interval = 10 minutes

2021-01-21 23:37:05 UTC CONFIG main: local node id = 1

2021-01-21 23:37:05 UTC INFO   main: main process started

2021-01-21 23:37:05 UTC CONFIG main: launching sched_start_mainloop

2021-01-21 23:37:05 UTC CONFIG main: loading current cluster configuration

2021-01-21 23:37:05 UTC CONFIG storeNode: no_id=2 no_comment=''

2021-01-21 23:37:05 UTC CONFIG storePath: pa_server=2 pa_client=1 pa_conninfo="dbname=test2

host=localhost port=5432 user=postgres password=changeme" pa_connretry=10

2021-01-21 23:37:05 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 23:37:05 UTC CONFIG storeSet: set_id=1 set_origin=1 set_comment='A replication set so boring no one thought to give it a name'

2021-01-21 23:37:05 UTC CONFIG main: last local event sequence = 5000000007

2021-01-21 23:37:05 UTC CONFIG main: configuration complete - starting threads

2021-01-21 23:37:05 UTC INFO   localListenThread: thread starts

2021-01-21 23:37:05 UTC CONFIG version for "dbname=test1 host=localhost port=5432

  user=postgres password=changeme" is 90620

2021-01-21 23:37:05 UTC CONFIG enableNode: no_id=2

2021-01-21 23:37:05 UTC INFO   main: running scheduler mainloop

2021-01-21 23:37:05 UTC INFO   remoteListenThread_2: thread starts

2021-01-21 23:37:05 UTC CONFIG cleanupThread: thread starts

2021-01-21 23:37:05 UTC INFO   syncThread: thread starts

2021-01-21 23:37:05 UTC INFO   monitorThread: thread starts

2021-01-21 23:37:05 UTC INFO   remoteWorkerThread_2: thread starts

2021-01-21 23:37:05 UTC CONFIG version for "dbname=test1 host=localhost port=5432

  user=postgres password=changeme" is 90620

2021-01-21 23:37:05 UTC CONFIG version for "dbname=test1 host=localhost port=5432

  user=postgres password=changeme" is 90620

2021-01-21 23:37:05 UTC CONFIG version for "dbname=test2

host=localhost port=5432 user=postgres password=changeme" is 90620

2021-01-21 23:37:05 UTC CONFIG version for "dbname=test1 host=localhost port=5432

  user=postgres password=changeme" is 90620

2021-01-21 23:37:05 UTC CONFIG version for "dbname=test1 host=localhost port=5432

  user=postgres password=changeme" is 90620

2021-01-21 23:37:05 UTC CONFIG cleanupThread: bias = 60

2021-01-21 23:37:05 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 23:37:05 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 23:37:05 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 23:37:07 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 23:37:07 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 23:37:07 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 23:37:07 UTC CONFIG remoteWorkerThread_2: update provider configuration

2021-01-21 23:37:27 UTC CONFIG storeListen: li_origin=2 li_receiver=1 li_provider=2

2021-01-21 23:37:27 UTC CONFIG remoteWorkerThread_2: update provider configuration

NOTICE:  Slony-I: Logswitch to sl_log_2 initiated

2021-01-21 23:47:06 UTC INFO   cleanupThread:    0.024 seconds for cleanupEvent()

NOTICE:  Slony-I: log switch to sl_log_2 still in progress - sl_log_1 not truncated

2021-01-21 23:57:06 UTC INFO   cleanupThread:    0.004 seconds for cleanupEvent()

NOTICE:  Slony-I: log switch to sl_log_2 still in progress - sl_log_1 not truncated

2021-01-22 00:07:06 UTC INFO   cleanupThread:    0.009 seconds for cleanupEvent()

2021-01-22 00:07:06 UTC INFO   cleanupThread:    0.033 seconds for vacuuming

NOTICE:  Slony-I: log switch to sl_log_2 still in progress - sl_log_1 not truncated

2021-01-22 00:17:07 UTC INFO   cleanupThread:    0.004 seconds for cleanupEvent()

``` 

### Setting up slonyrep user with md5 

I'm not sure if this is the really the right approach, being a (slony newbie) but here is what I setup

```

~/projects/vagrant-pg96-slony $ cat setup_slonyrep_dbuser.sh

# to be run as postgres

createuser -S slonyrep

psql -c "create user slonyrep with password 'changeme'";

psql -d pgbench -c "GRANT ALL PRIVILEGES ON DATABASE pgbench TO slonyrep;";

psql -d pgbench -c "GRANT ALL PRIVILEGES ON ALL TABLES IN SCHEMA public TO slonyrep;"

psql -d pgbenchslave -c "GRANT ALL PRIVILEGES ON DATABASE pgbenchslave TO slonyrep;";

psql -d pgbenchslave -c "GRANT ALL PRIVILEGES ON ALL TABLES IN SCHEMA public TO slonyrep;"

```

and finally checking that I can 

```

[pg96slony:postgres:~] # PGPASSWORD=changeme psql -U slonyrep -d pgbenchslave -c "select count(*) from pgbench_branches;"

 count

-------

     0

```

### Debuging - Ident authentication failed

#### problem - Ident authentication failed for user "slonyrep

This was my rough startin point 

```

[pg96slony:vagrant:/vagrant] # cat init_master.slonik

#! /usr/pgsql-9.6/bin/slonik

  cluster name = mycluster;

  node 1 admin conninfo = 'dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme';

  node 2 admin conninfo = 'dbname=pgbenchslave host=localhost port=5432 user=slonyrep password=changeme'';

  init cluster ( id=1);

  create set (id=1, origin=1);

  set add table(set id=1, origin=1, id=1, fully qualified

name = 'public.t_test');

  store node (id=2, event node = 1);

  store path (server=1, client=2, conninfo='dbname=pgbench  host=localhost port=5432 user=slonyrep password=changeme');

  store path (server=2, client=1, conninfo='dbname=pgbenchslave host=localhost port=5432 user=slonyrep password=changeme');

  store listen (origin=1, provider = 1, receiver = 2);

  store listen (origin=2, provider = 2, receiver = 1);

```

and while this works

```

[pg96slony:vagrant:/vagrant] # PGPASSWORD=changeme psql -d pgbench  -p 5432 -U slonyrep

psql (9.6.20)

Type "help" for help.

pgbench=> \q

```

this doesn't

```

[pg96slony:vagrant:/vagrant] # PGPASSWORD=changeme psql -d pgbench -h 127.0.0.1  -p 5432 -U slonyrep

psql: FATAL:  Ident authentication failed for user "slonyrep"

```

I needed to update my pg_hba.conf

```

host    pgbench         slonyrep        127.0.0.1/32            md5

host    pgbench         slonyrep        localhost               md5

host    pgbenchslave    slonyrep        127.0.0.1/32            md5

host    pgbenchslave    slonyrep        localhost               md5

```

### Debug permission denied

#### problem - permission denied to set parameter "session_replication_role"

```

[pg96slony:postgres:/vagrant] # /usr/pgsql-9.6/bin/slonik init_master.slonik

init_master.slonik:5: PGRES_FATAL_ERROR SET datestyle TO 'ISO'; SET session_replication_role TO local;  - ERROR:  permission denied to set parameter "session_replication_role"

Unable to set session configuration parameters

init_master.slonik:5: PGRES_FATAL_ERROR SET datestyle TO 'ISO'; SET session_replication_role TO local;  - ERROR:  permission denied to set parameter "session_replication_role"

Unable to set session configuration parameters

init_master.slonik:5: PGRES_FATAL_ERROR select 1 from "pg_catalog".pg_namespace N 	where N.nspname = '_mycluster'; - ERROR:  current transaction is aborted, commands ignored until end of transaction block

init_master.slonik:5: Error: cannot determine  if namespace "_mycluster" exists in node 1

[pg96slony:postgres:/vagrant] #

```

#### solution add superuser role to replication user

```

[pg96slony:postgres:/vagrant] # psql -c "alter user slonyrep superuser"

ALTER ROLE

```

### Debug duplicate keys

#### problem - duplicate key value violates unique constraint sl_set-pkey

```

[pg96slony:postgres:~] # /usr/pgsql-9.6/bin/slonik /vagrant/init_master.slonik

/vagrant/init_master.slonik:7: PGRES_FATAL_ERROR lock table "_mycluster".sl_event_lock, "_mycluster".sl_config_lock;select "_mycluster".storeSet(1, 'All pgbench tables');  - ERROR:  duplicate key value violates unique constraint "sl_set-pkey"

DETAIL:  Key (set_id)=(1) already exists.

CONTEXT:  SQL statement "insert into "_mycluster".sl_set

			(set_id, set_origin, set_comment) values

			(p_set_id, v_local_node_id, p_set_comment)"

```

#### solution rebuild VM

The problem here was around try to initialize the master node multiple nodes:

* it is hard to unpick the slony internal schema

* in the real world I guess you should be careful, maybe you can backup slony schema before making changes

* in my case I rebuilt the VM (vagrant destroy;vagrant up ...)

### Diggging around in the _slony_example schema

```

[pg96slony:postgres:~] # psql -U slonyrep -h localhost -p 5432 -d pgbenchslave

psql (9.6.20)

Type "help" for help.

pgbenchslave=# select count(*) from pgbench_accounts;

 count

-------

     0

(1 row)

pgbenchslave=# \dn

      List of schemas

      Name      |  Owner

----------------+----------

 _slony_example | slonyrep

 public         | postgres

(2 rows)

pgbenchslave=# SET search_path TO _slony_example, public;

SET

pgbenchslave=# select * from sl_table;

 tab_id | tab_reloid | tab_relname | tab_nspname | tab_set | tab_idxname | tab_altered | tab_comment

--------+------------+-------------+-------------+---------+-------------+-------------+-------------

(0 rows)

```

Although I was seeing data is the master node

### Legacy step

The tutorial documentation (https://www.slony.info/documentation/2.1/tutorial.html) is a bit vauge, 

```

Because Slony-I depends on the databases having the pl/pgSQL procedural language installed, we better install it now. 

It is possible that you have installed pl/pgSQL into the template1 database in which case you can skip this step because it's already installed into the $MASTERDBNAME.

createlang -h $MASTERHOST plpgsql $MASTERDBNAME

```

but this step appears to legacy for my setup  

```

[pg96slony:postgres:~] # createlang plpgsql pgbench

createlang: language "plpgsql" is already installed in database "pgbench"

[pg96slony:postgres:~] # createlang plpgsql pgbenchslave

createlang: language "plpgsql" is already installed in database "pgbenchslave"

```
ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Awesome

https://github.com/dgapitts/vagrant-pg96-slony

Awesome Lists containing this project

README
