Managing an Advanced Server installation v13

Unless otherwise noted, the commands and paths noted in the following section assume that you have performed an installation using the native packages.

Starting and stopping Advanced Server and supporting components

A service is a program that runs in the background and requires no user interaction (in fact, a service provides no user interface); a service can be configured to start at boot time, or manually on demand. Services are best controlled using the platform-specific operating system service control utility. Many of the Advanced Server supporting components are services.

The following table lists the names of the services that control Advanced Server and services that control Advanced Server supporting components:

You can use the Linux command line to control Advanced Server's database server and the services of Advanced Server's supporting components. The commands that control the Advanced Server service on a Linux platform are host specific.

Controlling a service on RHEL/Rocky Linux/AlmaLinux 8.x

If your installation of Advanced Server resides on RHEL/Rocky Linux/AlmaLinux 8.x, you must use the systemctl command to control the Advanced Server service and supporting components.

The systemctl command must be in your search path and must be invoked with superuser privileges. To use the command, open a command line, and enter:

systemctl <action> <service_name>

Where:

service_name specifies the name of the service.

action specifies the action taken by the service command. Specify:

• start to start the service.
• stop to stop the service.
• restart to stop and then start the service.
• status to discover the current status of the service.

Controlling a service on Ubuntu 20.04

If your installation of Advanced Server resides on version 20.04 of Ubuntu, assume superuser privileges and invoke the following commands (using bundled scripts) to manage the service. Use the following commands to:

• Discover the current status of a service:

  → WrapCopy

  /usr/edb/as13/bin/epas_ctlcluster 13 main status

• Stop a service:

  → WrapCopy

  /usr/edb/as13/bin/epas_ctlcluster 13 main stop

• Restart a service:

  → WrapCopy

  /usr/edb/as13/bin/epas_ctlcluster 13 main restart

• Reload a service:

  → WrapCopy

  /usr/edb/as13/bin/epas_ctlcluster 13 main reload

• Control the component services:

  → WrapCopy

  systemctl restart edb-as@13-main

Using pg_ctl to control Advanced Server

You can use the pg_ctl utility to control an Advanced Server service from the command line on any platform. pg_ctl allows you to start, stop, or restart the Advanced Server database server, reload the configuration parameters, or display the status of a running server. To invoke the utility, assume the identity of the cluster owner, navigate into the home directory of Advanced Server, and issue the command:

./bin/pg_ctl -D <data_directory> <action>

data_directory is the location of the data controlled by the Advanced Server cluster.

action specifies the action taken by the pg_ctl utility. Specify:

• start to start the service.
• stop to stop the service.
• restart to stop and then start the service.
• reload sends the server a SIGHUP signal, reloading configuration parameters
• status to discover the current status of the service.

For more information about using the pg_ctl utility or the command line options available, see the official PostgreSQL Core Documentation.

Choosing Between pg_ctl and the service Command

You can use the pg_ctl utility to manage the status of an Advanced Server cluster, but it is important to note that pg_ctl does not alert the operating system service controller to changes in the status of a server, so it is beneficial to use the service command whenever possible.

Configuring component services to AutoStart at System Reboot

After installing, configuring, and starting the services of Advanced Server supporting components on a Linux system, you must manually configure your system to autostart the service when your system reboots. To configure a service to autostart on a Linux system, open a command line, assume superuser privileges, and enter the following command.

On a Redhat-compatible Linux system, enter:

/sbin/chkconfig <service_name> on

Where service_name specifies the name of the service.

Connecting to Advanced Server with edb-psql

edb-psql is a command line client application that allows you to execute SQL commands and view the results. To open the edb-psql client, the client must be in your search path. The executable resides in the bin directory, under your Advanced Server installation.

Use the following command and options to start the edb-psql client:

psql -d edb -U enterprisedb

Where:

-d specifies the database to which edb-psql will connect.

-U specifies the identity of the database user that will be used for the session.

edb-psql is a symbolic link to a binary called psql, a modified version of the PostgreSQL community psql, with added support for Advanced Server features. For more information about using the command line client, see the PostgreSQL Core Documentation.

Configuring a package installation

The packages that install the database server component create service startup scripts.

Creating a database cluster and starting the service

The PostgreSQL initdb command creates a database cluster; when installing Advanced Server with an RPM package, the initdb executable is in /usr/edb/asx.x/bin. After installing Advanced Server, you must manually configure the service and invoke initdb to create your cluster. When invoking initdb, you can:

• Specify environment options on the command line.
• Include the systemd service manager on RHEL/Rocky Linux/AlmaLinux 8.x and use a service configuration file to configure the environment.

For more information, see the initdb documentation.

After specifying any options in the service configuration file, you can create the database cluster and start the service; these steps are platform specific.

On RHEL/Rocky Linux/AlmaLinux 8.x

To invoke initdb on a RHEL/Rocky Linux/AlmaLinux 8.x system, with the options specified in the service configuration file, assume the identity of the operating system superuser:

su - root

To initialize a cluster with the non-default values, you can use the PGSETUP_INITDB_OPTIONS environment variable by invoking the edb-as-13-setup cluster initialization script that resides under EPAS_Home/bin.

To invoke initdb export the PGSETUP_INITDB_OPTIONS environment variable with the following command:

PGSETUP_INITDB_OPTIONS="-E UTF-8" /usr/edb/as13/bin/edb-as-13-setup initdb

After creating the cluster, use systemctl to start, stop, or restart the service:

systemctl { start | stop | restart } edb-as-13

On Debian 10.x | 11 or Ubuntu 18.04 | 20.04

You can initialize multiple clusters using the bundled scripts. To create a new cluster, assume root privileges, and invoke the bundled script:

/usr/bin/epas_createcluster 13 main2

To start a new cluster, use the following command:

/usr/bin/epas_ctlcluster 13 main2 start  

To list all the available clusters, use the following command:

/usr/bin/epas_lsclusters

Specifying Cluster Options with INITDBOPTS

You can use the INITDBOPTS variable to specify your cluster configuration preferences. By default, the INITDBOPTS variable is commented out in the service configuration file; unless modified, when you run the service startup script, the new cluster will be created in a mode compatible with Oracle databases. Clusters created in this mode will contain a database named edb, and have a database superuser named enterprisedb.

Initializing the Cluster in Oracle Mode

If you initialize the database using Oracle compatibility mode, the installation includes:

• Data dictionary views compatible with Oracle databases.
• Oracle data type conversions.
• Date values displayed in a format compatible with Oracle syntax.
• Support for Oracle-styled concatenation rules (if you concatenate a string value with a NULL value, the returned value is the value of the string).
• Support for the following Oracle built-in packages.

Initializing the Cluster in Postgres Mode

Clusters created in PostgreSQL mode do not include compatibility features. To create a new cluster in PostgreSQL mode, remove the pound sign (#) in front of the INITDBOPTS variable, enabling the "--no-redwood-compat" option. Clusters created in PostgreSQL mode will contain a database named postgres and have a database superuser named postgres.

You may also specify multiple initdb options. For example, the following statement:

INITDBOPTS="--no-redwood-compat -U alice --locale=en_US.UTF-8"

Creates a database cluster (without compatibility features for Oracle) that contains a database named postgres that is owned by a user named alice; the cluster uses UTF-8 encoding.

If you initialize the database using "--no-redwood-compat" mode, the installation includes the following package:

In addition to the cluster configuration options documented in the PostgreSQL core documentation, Advanced Server supports the following initdb options:

--no-redwood-compat

Include the --no-redwood-compat keywords to instruct the server to create the cluster in PostgreSQL mode. When the cluster is created in PostgreSQL mode, the name of the database superuser will be postgres and the name of the default database will be postgres. The few Advanced Server’s features compatible with Oracle databases will be available with this mode. However, we recommend using the Advanced server in redwood compatibility mode to use all its features.

--redwood-like

Include the --redwood-like keywords to instruct the server to use an escape character (an empty string ('')) following the LIKE (or PostgreSQL-compatible ILIKE) operator in a SQL statement that is compatible with Oracle syntax.

--icu-short-form

Include the --icu-short-form keywords to create a cluster that uses a default ICU (International Components for Unicode) collation for all databases in the cluster. For more information about Unicode collations, refer to the EDB Postgres Advanced Server Guide available at:

https://www.enterprisedb.com/docs

For more information about using initdb, and the available cluster configuration options, see the PostgreSQL Core Documentation.

You can also view online help for initdb by assuming superuser privileges and entering:

/path_to_initdb_installation_directory/initdb --help

Where path_to_initdb_installation_directory specifies the location of the initdb binary file.

Modifying the Data Directory Location

on RHEL or Rocky Linux 8.x

On a RHEL/Rocky Linux/AlmaLinux 8.x host, the unit file is named edb-as-13.service and resides in /usr/lib/systemd/system. The unit file contains references to the location of the Advanced Server data directory. You should avoid making any modifications directly to the unit file because it may be overwritten during package upgrades.

By default, data files reside under /var/lib/edb/as13/data directory. To use a data directory that resides in a non-default location, perform the following steps:

• Create a copy of the unit file under the /etc directory:

  → WrapCopy

  cp /usr/lib/systemd/system/edb-as-13.service /etc/systemd/system/

• After copying the unit file to the new location, create the service file /etc/systemd/system/edb-as-13.service.

• Update the following values with new location of data directory in the /lib/systemd/system/edb-as-13.service file:

  → WrapCopy

  Environment=PGDATA=/var/lib/edb/as13/data
  PIDFile=/var/lib/edb/as13/data/postmaster.pid

• Delete the entire content of /etc/systemd/system/edb-as-13.service file, except the following line:

  → WrapCopy

  .include /lib/systemd/system/edb-as-13.service

• Run the following command to initialize the cluster at the new location:

  → WrapCopy

  PGSETUP_INITDB_OPTIONS="-E UTF-8" /usr/edb/as13/bin/edb-as-13-setup initdb

• Use the following command to reload systemd, updating the modified service scripts:

  → WrapCopy

  systemctl daemon-reload

• Start the Advanced Server service with the following command:

  → WrapCopy

  systemctl start edb-as-13

Configuring SELinux Policy to Change the Data Directory Location on RHEL/Rocky Linux/AlmaLinux 8.x

By default, the data files resides under /var/lib/edb/as13/data directory. To change the default data directory location depending on individual environment preferences, you must configure the SELinux policy and perform the following steps:

• Stop the server using the following command:

  → WrapCopy

  systemctl stop edb-as-13

• Check the status of SELinux using the getenforce or sestatus command:

  → WrapCopy

  # getenforce
  Enforcing
  
  # sestatus
  SELinux status:                 enabled
  SELinuxfs mount:                /sys/fs/selinux
  SELinux root directory:         /etc/selinux
  Loaded policy name:             targeted
  Current mode:                   enforcing
  Mode from config file:          enforcing
  Policy MLS status:              enabled
  Policy deny_unknown status:     allowed
  Max kernel policy version:      31

• Use the following command to view the SELinux context of the default database location:

  → WrapCopy

  ls -lZ /var/lib/edb/as13/data
  drwx------. enterprisedb enterprisedb unconfined_u:object_r:var_lib_t:s0 log

• Create a new directory for a new location of the database using the following command:

  → WrapCopy

  mkdir /opt/edb

• Use the following command to move the data directory to /opt/edb:

  → WrapCopy

  mv /var/lib/edb/as13/data /opt/edb/

• Create a file edb-as-13.service under /etc/systemd/system directory to include the location of a new data directory:

  → WrapCopy

  .include /lib/systemd/system/edb-as-13.service
  [Service]
  Environment=PGDATA=/opt/edb/data
  PIDFile=/opt/edb/data/postmaster.pid

• Use the semanage utility to set the context mapping for /opt/edb/. The mapping is written to /etc/selinux/targeted/contexts/files/file.contexts.local file.

  → WrapCopy

  semanage fcontext --add --equal /var/lib/edb/as13/data /opt/edb

• Apply the context mapping using restorecon utility:

  → WrapCopy

  restorecon -rv /opt/edb/

• Reload systemd to modify the service script using the following command:

  → WrapCopy

  systemctl daemon-reload

• Now, the /opt/edb location has been labeled correctly with the context, use the following command to start the service:

  → WrapCopy

  systemctl start edb-as-13

Enabling Core Dumps

You can use core dumps to diagnose or debug errors. A core dump is a file containing a process's address space (memory) when the process terminates unexpectedly. Core dumps may be produced on-demand (such as by a debugger) or automatically upon termination.

Enabling Core Dumps on a RHEL/Rocky Linux/AlmaLinux 8 Host

On a RHEL/Rocky Linux/AlmaLinux 8.x, core file creation is disabled by default. To enable the core file generation, follow the following commands:

• Identify the system's current limit using the ulimit -c or ulimit -a command. 0 indicates that core file generation is disabled.

  → WrapCopy

  # ulimit -c
  0
  
  # ulimit -a
  core file size          (blocks, -c) 0
  data seg size           (kbytes, -d) unlimited
  scheduling priority             (-e) 0
  file size               (blocks, -f) unlimited
  pending signals                 (-i) 3756
  max locked memory       (kbytes, -l) 64
  max memory size         (kbytes, -m) unlimited
  open files                      (-n) 1024
  pipe size            (512 bytes, -p) 8
  POSIX message queues     (bytes, -q) 819200
  real-time priority              (-r) 0
  stack size              (kbytes, -s) 8192
  cpu time               (seconds, -t) unlimited
  max user processes              (-u) 3756
  virtual memory          (kbytes, -v) unlimited
  file locks                      (-x) unlimited

• Create a new directory to store the core dumps and modify kernel.core_pattern to store the dumps in a specified directory:

  → WrapCopy

  mkdir -p /var/coredumps
  chmod a+w /var/coredumps
  
  sysctl kernel.core_pattern=/var/coredumps/core-%e-%p
  kernel.core_pattern = /var/coredumps/core-%e-%p

• Use the following command to persist the kernel.core_pattern setting across reboots:

  → WrapCopy

  echo 'kernel.core_pattern=/var/coredumps/core-%e-%p' >> /etc/sysctl.conf

• Enable core dumps in /etc/security/limits.conf to allow a user to create core files. Each line describes a limit for a user in the following form:

  → WrapCopy

  <domain>  <type>  <item>  <value>
     *       soft    core    unlimited

  Use * to enable the core dump size to unlimited.

• Set the limit of core file size to UNLIMITED by using the following command:

  → WrapCopy

  ulimit -c unlimited
  
  ulimit -c
  unlimited

• To set a core limit for the services, add the following setting in /usr/lib/systemd/system/edb-as-13.service.

  → WrapCopy

  [Service]
  LimitCore=Infinity

• Reload the service configuration:

  → WrapCopy

  systemctl daemon-reload

• Modify the global default limit using systemd, add the following setting in /etc/systemd/system.conf.

  → WrapCopy

  DefaultLimitCORE=Infinity

• Restart the systemd:

  → WrapCopy

  systemctl daemon-reexec

• Stop and then start Advanced Server:

  → WrapCopy

  systemctl stop edb-as-13
  systemctl start edb-as-13

• Now, the core dumps are enabled, install the gdb tool and debug packages using the following command:

  → WrapCopy

  yum install gdb
  debuginfo-install edb-as13 edb-as13-server-contrib edb-as13-server edb-as13-libs

• Replace the path to a core dump file before proceeding to get a backtrace using the bt command to analyze output:

  → WrapCopy

  gdb /usr/edb/as13/bin /var/coredumps/core-edb-postgres-65499
  (gdb) bt full

Enabling Core Dumps on a Debian or Ubuntu Host

On Debian 10 or Ubuntu 18, 20, core file creation is disabled by default. To enable the core file generation, follow the following commands:

• Identify the system's current limit using the ulimit -c or ulimit -a command. 0 indicates that core file generation is disabled.

  → WrapCopy

  # ulimit -c
  0
  
  # ulimit -a
  core file size          (blocks, -c) 0
  data seg size           (kbytes, -d) unlimited
  scheduling priority             (-e) 0
  file size               (blocks, -f) unlimited
  pending signals                 (-i) 7617
  max locked memory       (kbytes, -l) 65536
  max memory size         (kbytes, -m) unlimited
  open files                      (-n) 1024
  pipe size            (512 bytes, -p) 8
  POSIX message queues     (bytes, -q) 819200
  real-time priority              (-r) 0
  stack size              (kbytes, -s) 8192
  cpu time               (seconds, -t) unlimited
  max user processes              (-u) 7617
  virtual memory          (kbytes, -v) unlimited
  file locks                      (-x) unlimited

• Create a new directory to store the core dumps and modify kernel.core_pattern to store the dumps in a specified directory:

  → WrapCopy

  mkdir -p /var/coredumps
  chmod a+w /var/coredumps
  
  sysctl kernel.core_pattern=/var/coredumps/core-%e-%p
  kernel.core_pattern = /var/coredumps/core-%e-%p

• Use the following command to persist the kernel.core_pattern setting across reboots:

  → WrapCopy

  echo 'kernel.core_pattern=/var/coredumps/core-%e-%p' >> /etc/sysctl.conf

• Enable core dumps in /etc/security/limits.conf to allow a user to create core files. Each line describes a limit for a user in the following form:

  → WrapCopy

  <domain>  <type>  <item>  <value>
     *       soft    core    unlimited

  Use * to enable the core dump size to unlimited.

• Set the limit of core file size to UNLIMITED by using the following command:

  → WrapCopy

  ulimit -c unlimited
  
  ulimit -c
  unlimited

• To set a core limit for the services, add the following setting in /lib/systemd/system/edb-as@.service.

  → WrapCopy

  [Service]
  LimitCore=Infinity

• Reload the service configuration:

  → WrapCopy

  systemctl daemon-reload

• Modify the global default limit using systemd, add the following setting in /etc/systemd/system.conf.

  → WrapCopy

  DefaultLimitCORE=Infinity

• Restart the systemd:

  → WrapCopy

  systemctl daemon-reexec

• Stop and then start Advanced Server:

  → WrapCopy

  systemctl stop edb-as@13.service
  systemctl start edb-as@13.service

• Now, the core dumps are enabled, install the gdb tool and debug symbols using the following command:

  → WrapCopy

  apt-get install gdb
  apt-get install edb-as13 edb-as-contrib edb-as13-server edb-debugger-dbgsym

• Replace the path to a core dump file before proceeding to get a backtrace using the bt command to analyze output:

  → WrapCopy

  gdb /usr/lib/edb-as/13/bin /var/coredumps/core-edb-postgres-21638
  (gdb) bt full

Starting Multiple Postmasters with Different Clusters

You can configure Advanced Server to use multiple postmasters, each with its own database cluster. The steps required are version specific to the Linux host.

On RHEL/Rocky Linux/AlmaLinux 8.x

The edb-as13-server-core RPM for version 8.x contains a unit file that starts the Advanced Server instance. The file allows you to start multiple services, with unique data directories and that monitor different ports. You must have root access to invoke or modify the script.

The example that follows creates an Advanced Server installation with two instances; the secondary instance is named secondary:

• Make a copy of the default file with the new name. As noted at the top of the file, all modifications must reside under /etc. You must pick a name that is not already used in /etc/systemd/system.

  → WrapCopy

  cp /usr/lib/systemd/system/edb-as-13.service /etc/systemd/system/secondary-edb-as-13.service

• Edit the file, changing PGDATA to point to the new data directory that you will create the cluster against.

• Create the target PGDATA with user enterprisedb.

• Run initdb, specifying the setup script:

  → WrapCopy

  /usr/edb/as13/bin/edb-as-13-setup initdb secondary-edb-as-13

• Edit the postgresql.conf file for the new instance, specifying the port, the IP address, TCP/IP settings, etc.

• Make sure that new cluster runs after a reboot:

  → WrapCopy

  systemctl enable secondary-edb-as-13

• Start the second cluster with the following command:

  → WrapCopy

  systemctl start secondary-edb-as-13

Creating an Advanced Server Repository on an Isolated Network

You can create a local repository to act as a host for the Advanced Server RPM packages if the server on which you wish to install Advanced Server (or supporting components) cannot directly access the EDB repository. Please note that this is a high-level listing of the steps requires; you will need to modify the process for your individual network.

To create and use a local repository, you must:

• Use dnf to install the epel-release, yum-utils, and createrepo packages.

  On RHEL or Rocky Linux or AlmaLinux 8.x:

  → WrapCopy

  dnf install epel-release
  dnf install yum-utils
  dnf install createrepo

• Create a directory in which to store the repository:

  → WrapCopy

  mkdir /srv/repos

• Copy the RPM installation packages to your local repository. You can download the individual packages or use a tarball to populate the repository. The packages are available from the EDB repository at https://repos.enterprisedb.com/.

• Sync the RPM packages and create the repository.

  → WrapCopy

  reposync -r edbas13 -p /srv/repos
  createrepo /srv/repos

• Install your preferred webserver on the host that will act as your local repository, and ensure that the repository directory is accessible to the other servers on your network.

• On each isolated database server, configure yum or dnf to pull updates from the mirrored repository on your local network. For example, you might create a repository configuration file called /etc/yum.repos.d/edb-repo with connection information that specifies:

  → WrapCopy

  [edbas13]
  name=EnterpriseDB Advanced Server 13
  baseurl=https://yum.your_domain.com/edbas13
  enabled=1
  gpgcheck=0

After specifying the location and connection information for your local repository, you can use dnf to install Advanced Server and its supporting components on the isolated servers. For example:

• On RHEL or Rocky Linux or AlmaLinux 8:

  → WrapCopy

  dnf -y install edb-as13-server