Re: Swap on postgres master server - Mailing list pgsql-general

Hello,

we are running Postgres 10.5 with master slave replication.

These are our custom params

archive_command = 'pgbackrest --stanza=cmdprod archive-push %p' # command to use to archive a logfile segment

archive_mode = on # enables archiving; off, on, or always

checkpoint_completion_target = 0.7 # checkpoint target duration, 0.0 - 1.0

default_text_search_config = 'pg_catalog.english' #

datestyle = 'iso, mdy' #

effective_cache_size = 120GB #

hot_standby = on                        # "on" allows queries during recovery

lc_messages = 'en_US.UTF-8'                      # locale for system error message

lc_monetary = 'en_US.UTF-8'                      # locale for monetary formatting

lc_numeric = 'en_US.UTF-8'                      # locale for number formatting

lc_time = 'en_US.UTF-8'                          # locale for time formatting

listen_addresses = '*' # defaults to 'localhost', '*' = all

log_autovacuum_min_duration = 1000ms        # -1 disables, 0 logs all actions and

log_checkpoints = on #

log_line_prefix = '%t [%r] [%p]: [%l-1] db=%d,user=%u ' #

log_lock_waits = on                      # log lock waits >= deadlock_timeout

log_min_duration_statement = 1000ms        # -1 is disabled, 0 logs all statements

log_statement = 'ddl'                  # none, ddl, mod, all

log_temp_files = 1024kB                      # log temporary files equal or larger

maintenance_work_mem = 2GB #

max_connections = 220 #

max_parallel_workers_per_gather = 8        # taken from max_worker_processes

max_wal_size = 2GB #

min_wal_size = 1GB #

pg_stat_statements.max = 10000 #

pg_stat_statements.track = all #

port = 5432                            # port number which Postgres listen

shared_buffers = 10GB #

shared_preload_libraries = 'pg_stat_statements'          # (change requires restart)

synchronous_standby_names = '1 ( "usnyh2" )' # comment out during upgrade

track_activity_query_size = 16384        # (change requires restart)

track_io_timing = on #

wal_buffers = 16MB #

wal_keep_segments = 100 #

wal_level = replica                    #  minimal, replica, or logical

work_mem = 600MB #

This server is on Centos 7 and the strange thing is that we see a lot of swap usage :

[root@usnyh-cmd1 ~]# free -m

              total        used        free      shared  buff/cache   available

Mem:         257652        7555        5559       12804      244536      236036

Swap:         16383        7326        9057

7GB used.

But can't see it from any of the commands like top etc.

I am sure it is postgres because it is the only service running on that machine.

Is there anything we can do?

On the sync slave, the usage is just 400MB.

Any trick?

Thanks a lot,

Nicola