How to delete all lines above and below the current line in the text editor vi

In vi (or vim), you can delete all lines above the current line with this one simple command:

:1,.−1d

To delete all rows above the current lin, press d (delete) followed by capital G (end of document):

d + G

Convert all contents in a file to lower case using vim

Open your file in vi

vi myfile

Press the escape key to enter interactive mode, then type

:%s/.*/\L&/

Press escape, then save + exit the editor:

:wq

How to list the postgres processes running on your server and what their jobs are

A neet little trick I learned from the series "Mirage of memory, part 2: PSS"

ps -fu postgres | grep $(pgrep -f postmaster)
postgres    1913       1  0 Jul17 ?        00:02:51 /usr/pgsql-15/bin/postmaster -D /var/lib/pgsql/15/data/
postgres    1982    1913  0 Jul17 ?        00:00:24 postgres: logger
postgres    2063    1913  0 Jul17 ?        00:00:18 postgres: checkpointer
postgres    2064    1913  0 Jul17 ?        00:00:03 postgres: background writer
postgres    2119    1913  0 Jul17 ?        00:00:07 postgres: walwriter
postgres    2120    1913  0 Jul17 ?        00:00:14 postgres: autovacuum launcher
postgres    2121    1913  0 Jul17 ?        00:00:01 postgres: archiver
postgres    2122    1913  0 Jul17 ?        00:00:00 postgres: logical replication launcher
postgres    2657    1913  0 Jul17 ?        00:02:05 postgres: postgres_exporter postgres localhost(48674) idle

The command lists all processes by user "postgres" and and greps for the output of the command "pgrep -f postmaster", which returns a list of process IDs

You could also count the number of processes by user "postgres" by simply typing

pgrep -c postmaster
9

How to list all subdirectories and exclude certain subdirectories in a tree-like fashion

In this example I want to see all subdirectories that starts with "dir", but also to exclude the directory "dirdat":

cd $ORACLE_GG
tree -L 2 --noreport $(ls -d dir* | grep -v '^dirdat$')

ls -d dir*: lists all dir* entries (only directories if they match)

grep -v '^dirdat$': removes dirdat from the list

$(...): expands the filtered list as arguments to tree

How to use rsync between two servers

For example, to syncronize two servers so they get identical file structures: 

• Source server: server1.oric.no 
• Destination server: server2.oric.no 
• Execute as: root

target=server2.oric.no && rsync -vaz --delete /u01/app/myfiles ebsuser@$target:/u01/app

This will sync the entire "/u01/app/myfiles" folder + all subfolders.

The --delete flag in rsync ensures that files in the destination directory (ebsuser@$target:/u01/app) that are not present in the source directory (/u01/app/inst) are deleted. This helps keep the destination directory synchronized with the source directory.

How to list the contents of a folder in a tree-like fashion in Linux

Use the tree command:

cd /sw/oracle/admin
tree -L 2
.
├── cdb
│   ├── adump
│   ├── dpdump
│   ├── log
│   ├── pfile
│   └── xdb_wallet
├── sales
│   ├── adump
│   ├── dpdump
│   ├── pfile
│   ├── sql
│   └── xdb_wallet
└── hr
    ├── adump
    ├── dpdump
    ├── log
    ├── pfile
    ├── scripts
    ├── sql
    └── xdb_wallet

20 directories, 0 files

The -L flag indicates the number of levels you want to display. In my case, if I change the value from 2 to 3, I get the output below instead (abbreviated):

.
├── cdb
│   ├── adump
│   │   ├── FE78BD1F8E6730CDE0536709D10AC9C0
│   │   └── FE7BD04D2DFBE569E0536709D10A3AF0
│   ├── dpdump
│   │   ├── dp.log
│   │   ├── FE78BD1F8E6730CDE0536709D10AC9C0
│   │   ├── FE7AF28B415262F7E0536709D10A8B2E
│   │   └── FE7BD04D2DFBE569E0536709D10A3AF0
│   ├── log
│   │   ├── 2023-07-11_cdb.2431565
│   │   ├── 2023-07-12_cdb.2489915
│   │   ├── 2023-07-12_cdb.2576176
├── sales
│   ├── adump
│   ├── dpdump
│   │   └── dp.log
│   ├── pfile
│   │   └── init.ora.5192023145034
│   ├── sql
│   │   ├── analyze_sales.sh
│   │   ├── cfgtoollogs
│   │   ├── config_sales.txt
│   │   ├── cre_db.sh
│   │   └── deploy_sales.sh

26 directories, 181 files

The command is not installed by default but is avaible both for RHEL / CentOS / Fedora Linux as well as Debian based Linux distributions like Ubuntu.

How to list the content of a folder and get total size summary of the parent folder at the end

To list the size of all folders and get a summery at the end, use the --max-depth=1 switch with the du command.

In the listing below, I am following the examples of a recent blog post by Mike Dietrich, to measure the size of the subfolders in the .patch_storage folder, which is maintained by OPatch in case the DBA needs to rollback the patches later.

cd $ORACLE_HOME/.patch_storage

du -h --max-depth=1
100K    ./29585399_Apr_9_2019_19_12_47
284K    ./NApply
4.0K    ./oracle-home-1681903491076105
434M    ./34786990_Dec_6_2022_13_24_50
20K     ./NRollback
436M    ./35050341_Mar_17_2023_04_11_10
2.0G    ./35042068_Apr_6_2023_15_25_04
91M     ./backup_delete_inactive
2.9G    .

How to get verbose output in a call to mailx

Use the -vvv flag to debug your call to mailx:

root # echo "Test message" | mailx -S smtp="smtp.oric.no" -vvv -s "$(hostname) is up" monitoring@oric.no

How to identified active files/executables when using opatch to deinstall software

During a opatch rollback operation, we saw the following error message in the opatch logfile:

[Feb 19, 2023 2:29:51 PM] [INFO]    Prerequisite check "CheckActiveFilesAndExecutables" failed.
                                    The details are:


                                    Following active files/executables/libs are used by ORACLE_HOME :/sw/oracle/product/19.18
                                    /sw/oracle/product/19.18/lib/libclntsh.so.19.1

Reason:
Some processes are still using the /sw/oracle/product/19.18/lib/libclntsh.so.19.1.
Use the fuser utility with verbose output to find the process:

fuser -v /sw/oracle/product/19.18/lib/libclntsh.so.19.1
                     USER        PID ACCESS COMMAND
/sw/oracle/product/19.18/lib/libclntsh.so.19.1:
                     oracle    48439 ....m prometheus_orac
                     oracle    595787 ....m ggsci

There were two open processes using the file libclntsh.so.19.1: 1. the Golden Gate Manager 2. a utility called prometheus_oracle_exporter

Solution:

log in as the Golden Gate software owner
ggsci --> info all --> list all processes
stop mgr !

One of two processes quit its handler on the file:

 fuser -v /sw/oracle/product/19.18/lib/libclntsh.so.19.1
                     USER        PID ACCESS COMMAND
/sw/oracle/product/19.18/lib/libclntsh.so.19.1:
                     oracle    48439 ....m prometheus_orac

For the prometheus agent, we simply kill the agent, and the output from fuser now reveals that no file handlers are open:

kill 48439
fuser -v /sw/oracle/product/19.18/lib/libclntsh.so.19.1

How to prevent sed to destroy your symlinks

Because of EBS, we have a subdirectory in our $TNS_ADMIN which has the following name

PDB1_test1.oric.no

In other words, the files listener.ora, sqlnet.ora and tnsnames.ora are symlinks pointing to this subdirectory:

# ls -altr
lrwxrwxrwx  1 oracle dba   34 Feb 10 13:29 sqlnet.ora -> PDB1_test1.oric.no/sqlnet.ora
lrwxrwxrwx  1 oracle dba   36 Feb 10 13:29 tnsnames.ora -> PDB1_test1.oric.no/tnsnames.ora
lrwxrwxrwx  1 oracle dba   36 Feb 18 12:17 listener.ora -> PDB1_test1.oric.no/listener.ora
drwxr-xr-x  2 oracle dba 4096 Feb 18 12:18 PDB1_test1.oric.no
drwxr-xr-x  5 oracle dba 4096 Feb 18 12:19 .

If you are to exchange a string in these files, for example when switching to a new version of the Oracle software, make sure you use the directive --follow-symlinks to preserve your symlinks:

export PRE_ORACLE_VERSION=19.17
export NEW_ORACLE_VERSION=19.18
cd $TNS_ADMIN
sed --follow-symlinks -i "s/$PRE_ORACLE_VERSION/$NEW_ORACLE_VERSION/g" listener.ora

If you don't, the listener.ora will be placed directly in $TNS_ADMIN, and the symlinks will be dropped.

Error fetching the value from a shell script executed by root, when doing "su - oracle"

Problem:
Inside a new bash shell session, executed as user "oracle", like this:

su - oracle <<!

instructions here

!

I was not able to save the value of a string extracted from an xml file in a session variable.
I tried all possible variants (I thought), but was still unable to get my shell script to hold the value in a variable and use it locally within the oracle user session.

Solution:
Use backslash around the entire command, marked in yellow below:

su - oracle <<!
export CNTXT_FILE_PERL_VERSION=\`grep PERL5LIB \$CONTEXT_FILE | awk -F "site_perl/" '{print \$2}' | cut -f 1 -d":"\`
!

Credit goes to an experienced and clever colleague of mine, who shall remain nameless unless he actively wants his name listed on my blog ;-)

How to solve script error $'\r': command not found syntax error: unexpected end of file

when trying to execute a script from root, like the following:

 su - oracle -c "/tmp/set_params.sh"

You may see errors like the following:

set_params.sh: line 6: $'\r': command not found
set_params.sh: line 42: syntax error: unexpected end of file

Root cause:

You have windows-style formatting in your file

Solution:

dos2unix set_params.sh
dos2unix: converting file flexpod_params.sh to Unix format...

When executed again, the above errors should be gone!

How to log entire output from a shell script

I have recently used exec to send all output from a bash script to a designated logfile.

For example, if you have a complex script which performs multiple tasks and even execute tasks under a different user along the way, all the output will go into the global logfile indicated by the $GLOGFILE variable:

#!/bin/bash
export SCRIPT_NAME=`basename $0`
export TS=`date +\%m.\%d.\%y\_%H_%M_%S`
export GLOGFILE=/u01/mylogfile_${TS}.log
touch ${GLOGFILE}
chmod 666 ${GLOGFILE}
exec 1> ${GLOGFILE} 2>&1
echo "Starting job at " `date`

your shell instructions here

echo "Ending job at " `date`
exit

Good sources:

Linuxfordevices.com

Geeksforgeeks.org

How to list all nfs mountpoints

findmnt -t nfs

TARGET SOURCE                                               FSTYPE OPTIONS
/u01   pzl2ora1:/Oracle/software/pzh0oric/u01 nfs    rw,nodiratime,relatime,vers=3,rsize=65536,wsize=65536,namlen=255,acregmin=300,acregmax=300,acdirmin=300,ac
/u02   pzl2ora1:/Oracle/pzh0oric/u02          nfs    rw,nodiratime,relatime,vers=3,rsize=65536,wsize=65536,namlen=255,acregmin=300,acregmax=300,acdirmin=300,ac
/u03   pzl2ora1:/Oracle/pzh0oric/u03          nfs    rw,nodiratime,relatime,vers=3,rsize=65536,wsize=65536,namlen=255,acregmin=300,acregmax=300,acdirmin=300,ac
/u04   pzl2ora1:/Oracle/pzh0oric/u04          nfs    rw,nodiratime,relatime,vers=3,rsize=65536,wsize=65536,namlen=255,acregmin=300,acregmax=300,acdirmin=300,ac

How to work around error [INS-08101] Unexpected error while executing the action at state: ‘supportedOSCheck’

Thanks so much to Martin Berger for his blog post showing how to work around an error that shows up when you attempt to install Oracle 19c software on a RHEL8 distribution.

I wanted to install Oracle software on a RH8.5 Linux server:

cat /etc/redhat-release
Red Hat Enterprise Linux release 8.5 (Ootpa)
  

This is the error I received:

    

In my case, the only thing I had to do was to add another environmental variable:

export CV_ASSUME_DISTID=OEL7.8

Execute the installer again and you will see a different screen:

./runInstaller

How to view the contens of a zip file without extracting it

Method 1:

zip -sf myzipfile.zip

Method 2:

zipinfo myzipfile.zip

How to select a specific line number using awk

The file clonetext.txt contains:

  Clone Volume: true
                      Clone Parent server Name: myserver1-cluster49
                        SnapClone Parent Volume: myvolume_mirror1
                        

You would like to get the name of the volume, which is the 4th word on the 3rd line of the file. When grepping for the keyword "Clone" all lines are returned:

 cat clonetext.txt | grep Clone
  Clone Volume: true
                      Clone Parent server Name: myserver1-cluster49
                        SnapClone Parent Volume: myvolume_mirror1

Grepping for the 4th column helps, but you still get two rows, not just the single one you're interested in:

oracle@oric-db01:[mydb01]# cat clonetext.txt | grep Clone | awk -F' ' '{print $4}'

Name:
myvolume_mirror1

Adding the NR flag to your command solves the problem:

cat clonetext.txt | grep Clone | awk -F' ' 'NR==3 {print $4}'
myvolume_mirror1

How to find offending process when getting error umount.nfs: /data1: device is busy

As root, use lsof ("list open files"):

[root@myserver /home/oracle]# lsof | grep '/data1'
extract    14041        oracle   25r      REG               0,23      2896 1235955379 /data1/goldengate/dirdat/et000000127 (storage1:/Oracle/myserver/data1)

You will see a list of processes. Terminate these with the kill-command:

kill 13859 14041 

After this is done, you can unmount the nfs file system:

 umount -f /data1

How to debug an ssh connection

ssh -X -v myserver.mydomain.com

You can also use -vv and -vvv for more detailed debug information.

How to display Linux kernel and version information on debian platforms

Use the lsb (Linux Standard Base) utility:

lsb_release -a

Output:

No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 18.04.1 LTS
Release: 18.04
Codename: bionic

Replace the -a switch with the -d switch to limit the output to the version only.

or view the file /etc/os-release:

cat /etc/os-release:

NAME="Ubuntu"
VERSION="18.04.1 LTS (Bionic Beaver)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 18.04.1 LTS"
VERSION_ID="18.04"
HOME_URL="https://www.ubuntu.com/"
SUPPORT_URL="https://help.ubuntu.com/"
BUG_REPORT_URL="https://bugs.launchpad.net/ubuntu/"
PRIVACY_POLICY_URL="https://www.ubuntu.com/legal/terms-and-policies/privacy-policy"
VERSION_CODENAME=bionic
UBUNTU_CODENAME=bionic