Database: DB2 v9.5 on AIX
Scenario: I have 2 instances- db2inst1 and db2inst2. I created db2inst1 first. I went ahead and created db2inst2 next and I do 'db2set -all' on db2inst2 instance I see this:
[g] DB2INSTDEF=db2inst1
I would like to reset DB2INSTDEF and make it not show up when I do 'db2set -all' (other than dropping and recreating db2inst1).
Thanks.
DB2INSTDEF does not apply on AIX (it's only used on Windows systems), so it doesn't really matter if (or what) it is set to. Why is this a problem?
If you really want to eliminate the variable, then run (as root):
/opt/IBM/db2/V9.5/adm/db2set -g DB2INSTDEF=
(The path /opt/IBM/db2/V9.5 may be slightly different on your system, depending on whether you installed DB2 into the default location or not).
Related
Running PostgreSQL on MANJARO, the actual /var/lib/postgres/data/postgresql.conf defines my Data-Directory:
data_directory = '/mnt/Data/.../PostgreSQL'
Now I'm facing the upgrade from PostgreSQL 13 to 14 and I'm thinking to use pg_upgrade, as described in the Arch-Wiki.
The Step 4.) described there is pretty clear, but is it also sufficient for my data_directory = '/mnt/Data/.../PostgreSQL' or do I need to take additional actions?
Furthermore, I'm thinking to have versioned directories, e.g.
/var/lib/postgres/data/14/
/mnt/Data/.../PostgreSQL/14/
What should I consider to realize that?
Thank you!
When running pg_upgrade, point the --old-datadir and --new-datadir options to the directories that contain postgresql.conf (not to the actual data directories). I admit that this is confusing; however, the documentation says:
-d configdir
--old-datadir=configdir
the old database cluster configuration directory; environment variable PGDATAOLD
Note that the wording is “cluster configuration directory”, not “data directory”.
I've been having a bit of trouble trying to install PostgreSQL 14 for the first time.
I would like to apologize in advance if this question has been asked in the manner that I am about to ask it, but I do not think it has. If it has been, please direct me to the appropriate location!
I've done a fair amount of Googling on the matter, and all the information that I find seems to be rather fragmented, or I end up following a spaghetti trail of hyperlinks (a la-do-this-and-follow-this-other-link-with-more-information-than-you-need-to-understand-this-other-required-portion).
Personally, I don't want to jump around to 50 different locations on the web to try and conjure up a piecemeal solution that I believe works, only to be proven wrong later. I want to know what to do and why it works. I've tried reading the documentation, and have given up on it, because to me, it seems to assume that the server has already been set up by a database administrator.
Instead of articulating my problem directly (as I seem to be having more trouble than I would like by trying to do so), I believe it would be easier to articulate my problem indirectly by stating what my expectations would be after installing PostgreSQL for the first time.
So to start, I will mention that I'm running Ubuntu 18.04.6 LTS, and am installing PostgreSQL 14.1 with the following command:
sudo apt install postgresql-14
Before continuing, I would like to add a side note in advance, that I do not want suggestions for an alternative OS or install method. I just want to be able to get "up and running" in a common-sense fashion from this exact point.
Moving on, I know that the aforementioned command creates a *nix user called postgres.
From here, I can now indirectly state my problem using an outline of what my goals and expectations are immediately after installing the software via that command.
After installing PostgreSQL via apt, these are my expected goals:
I want any client to be able to connect to the database server from any computer where a route exists from the client to the server.
For the sake of simplicity with these stated goals, when it is directly or implicitly stated that I am trying to connect a client to the database server, I am making the assumption that the client is able to, at a minimum, ping the machine that the server is running on, and vice versa.
For now, I'm not completely worried about the database being accessible from the public Internet.
I expect to be able to access the database from any computer on my LAN, whether it is an actual LAN, or some sort of logical LAN (like a WAN or a VPN).
If I change the PostgreSQL password of the postgres user, I expect that any client logging into the database server via the postgres user will require the password.
This means if I want to change the password to some_password via \password postgres or ALTER USER postgres WITH PASSWORD 'some_password'; (I am assuming this is how you change the login password of a PostgreSQL user), then...
I expect running psql [-h host] -U postgres -W from any host...
That when I am prompted to enter the password...
I can only log in by entering the exact password of some_password.
Entering any other arbitrary text for the password should not allow me to log in.
I am adding this as a requirement because previous install attempts have shown me that this is NOT the case.
I expect to be able to create a PostgreSQL user account other than postgres (e.g. db_user) with a password and have it be subject to the same requirements as the postgres user.
i.e. once the new account is given permission to log in, the same common-sense login requirements to log in must be imposed, i.e. you can't get in if you don't have the correct username/password combination.
If the process to achieve the aforementioned can be explained in such a way that it can be understood with minimal mental friction, I would be extremely grateful.
Feel free to assume that my knowledge is on par with that of a undergraduate CS student who just completed their first year of university, who also understands Linux filesystems and basic computer networking. I just want the answer to be as accessible to as many people as possible, as I am sure I'm not the only person who has struggled with installing PostgreSQL, in spite of having a power user's level of computer literacy.
sudo apt install postgresql
sudo -u postgres psql
Set a password for this user with \password or the other method you mention
sudo vi /etc/postgresql/10/main/pg_hba.conf
Make the only uncommented nonblank line in this file be host all all all md5
sudo vi /etc/postgresql/10/main/postgresql.conf
uncomment listen_addresses line and set it to '*'
sudo service postgresql restart
When you make a new user, you should also make a new database which has the same spelling as the user does. Otherwise you will need to specify the database name when you try to log in with psql -U, such as psql -U newname -d postgres -h[hhh]. Should you actually be running 14 not 10, then you will need to change the paths of the config files you need to edit accordingly.
Two days ago, i started learning postgresql. Most tutorials I followed online were either old or the codes wont just work on my mac. I followed a lot of tutorials that did a lot of totally different things.
When i switched on my system today. I noticed Postgresql created a user on my mac. I don't know what this is or maybe i used the wrong CLI code.
When I tried viewing the user, I saw this
should I delete this user or it has a function?
postgres user account
Creating a user account specifically for Postgres, commonly named postgres, is a normal part of a Postgres installation. Your installer app likely prompted you for a password to assign to this new user account.
One reason for this is security: The database’s data files and security configuration files are stored in folders owned by the postgres user. So if your main user account is hijacked, the intruder does not yet have access to the database (often the most valuable thing in storage). The intruder must jump through more hoops to compromise Postgres. Also, the separate ownership prevents other apps from inadvertently stomping on the Postgres files.
You will find Postgres is much more enterprise-oriented than other products such as MySQL. This means locking-down for security. Another example: Postgres by default is configured to not accept connections over the network. To enable connections from other computers, you must change the configuration. Inconvenient for the beginner, but more secure. Like a bar on your car steering wheel and deadbolts on your doors, more security always means more steps to take and more annoyance.
Use a virtual machine
Installing the postgres user account is one of the things that makes Postgres a rather heavyweight installation. I suggest to those learning Postgres to use a virtual machine for Postgres. Something like:
Parallels or Fusion or VirtualBox on your own computer
Cloud server such as FreeBSD on DigitalOcean.com.
To remove Postgres, simply discard the vm.
Postgres.app for macOS
Another option for a Mac user is Postgres.app, created by the person who built one of the first Postgres-as-a-Service implementations (on Heroku). I have not used Postgres.app, but I understand it wraps Postgres, so it does not install the postgres user account. Also, Postgres starts and stops when launch and quit the app, rather than running in the background all the time.
Be aware: you may have conflicts with Postgres.app on a Mac where you already have a conventional installation. I suggest you first carefully remove the conventional Postgres from your Mac before installing Postgres.app. Uninstalling involves finding and deleting various files and folders in various places.
Database-as-a-Service (DBaaS)
Another option to avoid local installation is the increasing choices for running Postgres as a service. This is sometimes referred to as “managed Postgres” because the vendor maintains the installation of Postgres on your behalf. You simply use Postgres to create your database, but you do not fully control Postgres in such a service.
Some examples:
Heroku
Digital Ocean
Azure
Amazon Web Services (AWS)
ElephantSQL
My experience
Personally, I often install Postgres on a Mac using the installer by EnterpriseDB.com. That company sells added-value versions of Postgres, but kindly provides an installer for plain-vanilla Postgres, as a service to the community.
I have also used that same installer from EnterpriseDB.com to install onto a Parallels VM running macOS as the guest OS within the VM on a MacBook Pro running macOS as the host OS. You can easily configure the VM to share the host Mac’s IP address on the network, or you can give the VM its own network address which might be handy for demo/dev/test work.
Thirdly, I have installed Postgres on FreeBSD on DigitalOcean.com.
All three of these options have worked quite well for me. Which is preferable depends on the scenario. For example, the DigitalOcean.com approach is good if I want colleagues to be able to reach the database 24x7 without my own MacBook being available.
This discussion is for development work. For mission-critical deployment, I strongly recommend using heavy-duty server equipment with error-correcting memory and redundant storage such as RAID or ZFS pool. Postgres is extremely reliable but depends, of course, on reliable hardware.
Your tag says Postgres 9.1. That version is quite old now. I suggest using the latest version. By the way, the version numbering system has changed for postgres. The first number is now the roughly-annual release number likely requiring you to dump and reload data to upgrade, and the second number is compatible updates.
As pointed out by #basil-bourque, the account is required for several reasons.
That said, if it annoys you to have the PostgreSQL showing up in the login screen -as it did me-, you can remove it as long as you have admin user rights in MacOS.
Apple Support gives the following command to hide a user from the login screen:
$ sudo dscl . create /Users/hiddenuser IsHidden 1
However, since the PostgreSQL user was not included listed in the login window at installation, that command will yield no result -at least in Catalina, which is my OS.
You should use the following two commands instead, as suggested by josemarluedke:
## add postgres to the list of hidden users on login screen
$ sudo defaults write /Library/Preferences/com.apple.loginwindow HiddenUsersList -array-add 'postgres'
## instruct not to show any hidden accounts at login
$ sudo defaults write /Library/Preferences/com.apple.loginwindow SHOWOTHERUSERS_MANAGED -bool FALSE
Worked for me!
I have Postgresql 9.5 installed on Ubuntu and need to use dblink. However I am using port 5433 (set in postgresql.conf) but dblink always defaults to port 5432. I know you can specify the port in the dblink parameters, but this is massive redundancy when I want to use the default port of my instance.
There is a way of telling dblink to use a specific port by default. I know this because I have successfully changed it in the past - but can't for the life of me remember how I did it. I have tried setting the PGPORT environment variable and rebooting. I have tried recreating the extension. I have tried reinstalling postgresql-contrib. None of this has made a difference. I have a feeling that there is some obscure file I edited but can't remember what, where or how.
Any suggestions of how to change the default connection parameters for dblink?
Got it!
It IS done by changing the environment variable, but this is not an environment variable in the sense that a Windows user would understand it (so changing /etc/environment does not work and neither does any other normal nix-style setting of global environment variables).
To change the default connection parameters for dblink you need to add PGPORT=5433 (in my case) to /etc/postgresql/9.5/main/environment. The annoying thing with Ubuntu install of Postgres is that I have at least three versions of all these sorts of files (including pg_HBA.conf and so on). The ones in /etc/... are mostly not read apart from in this case. It's the ones in my data directory (which for space reasons is in a non-default location). So, for anybody else having the same issue check /etc, /var, /your-data-directory.
I want to make a script that will run postgres in-memory without durability.
I read this page: http://www.postgresql.org/docs/9.1/static/non-durability.html
But I didn't understand how I can set this parameters in script. Could you please, help me?
Thanks for help!
Most of those parameters, like fsync, can only be set in postgresql.conf. Changes are applied by re-starting PostgreSQL. They apply to the whole database cluster - all the databases in that PostgreSQL install. That's because the databases all share a single postmaster, write-ahead log, and set of shared system tables.
The only parameter listed there that you can set at the SQL level in a script is synchronous_commit. By setting synchronous_commit = 'off' you can say "it's OK to lose this transaction if the database crashes in the next few seconds, just make sure it still applies atomically".
I wrote more on this topic in a previous answer, Optimise PostgreSQL for fast testing.
If you want to set the other params with a script you can do so but you have to do it by opening and modifying postgresql.conf using the script, then re-starting PostgreSQL. Text-processing tools like sed make this kind of job easier.
If you're running a debian based linux distro, you can just do something like:
pg_createcluster -d /dev/shm/mypgcluster 8.4 ramcluster
to create a ram based cluster. Note that you'll have to do:
pg_drop cluster 8.4 ramcluster
and recreate it on reboot etc.