DBCollect/Install: Difference between revisions

From Dirty Cache Wiki
Jump to navigation Jump to search
← Older edit
VisualWikitext
m Bart moved page DBCollect/Iinstall to DBCollect/Install without leaving a redirect
No edit summary
 
(52 intermediate revisions by the same user not shown)
Line 1: Line 1:
[[Category:Tools]]
[[Category:Tools]]
[[Category:DBCollect]]
{{PrevNext|DBCollect/Considerations|DBCollect/Usage}}


= DBcollect Install =
= Summary =


== Summary ==
Installing DBCollect is pretty simple: download the latest dbcollect zipapp, put it somewhere in $PATH, and make it executable.
 
On Linux this usually just works, on UNIX it may require some additional steps.
 
{{ PageBreak }}
 
= Requirements =
 
== Operating Systems ==


Installing DBCollect is pretty simple: download the latest dbcollect zipapp, put it somewhere in $PATH, and make it executable.
* Linux
: Enterprise Linux 6 (EL6) or higher (RHEL, OEL, CentOS, Rocky Linux)
* SPARC/Solaris
* IBM AIX


== Downloading DBcollect ==
== Python ==


The latest release is always available on the releases page:
DBCollect is a Python ZIP app. This means Python needs to be available in the environment to run <code>dbcollect</code>. Some systems do not have this by default.
[https://github.com/outrunnl/dbcollect/releases/latest DBcollect package]


On the releases page, go to the latest release, open Assets and download dbcollect.
* Requires Python (2.6+ or 3.x)
** On Linux (EL6+) works out of the box
** On AIX/HP-UX, need to install Python first, Windows not supported
** On Solaris (and AIX) Python may be a very old version that is not supported


== Scripted install ==
A Python ZipApp is a standard ZIP file with an additional "shebang" prepended: <code>#!/usr/bin/env python</code>


To make life a bit easier, the github page provides a simple script that downloads the latest dbcollect release, puts it in /usr/local/bin and makes it executable.
* This causes dbcollect to be called with whatever ‘python’ version is in $PATH
* If this does not work, manually run python: <code>/usr/bin/python3 /usr/local/bin/dbcollect <options></code>
* Executable can be inspected with <code>unzip –v /usr/local/bin/dbcollect</code>
* Run simply as root or as ‘oracle’ user/owner


This requires the host on which you run this to have internet access. If you run this as root, all works as expected. If you run this as other user (i.e., oracle), you need to move it yourself to the right location.
On UNIX (AIX, Solaris) you may have to create a symbolic link or run {{dbcollect}} like this:


<syntaxhighlight lang=bash>
<syntaxhighlight lang=bash>
# This requires internet access via https
python3 /usr/local/bin/dbcollect
# Inspect downloader (optional)
</syntaxhighlight>
curl https://raw.githubusercontent.com/outrunnl/dbcollect/master/scripts/download | less


# Download using downloader
If all else fails, you can run {{dbcollect}} from python directly:
curl https://raw.githubusercontent.com/outrunnl/dbcollect/master/scripts/download | python


# Move it to /usr/local/bin
<syntaxhighlight lang=bash>
sudo mv dbcollect /usr/local/bin
python /usr/local/bin/dbcollect
# or
/usr/bin/python /usr/local/bin/dbcollect
# or
python2 /usr/local/bin/dbcollect
</syntaxhighlight>


# Or move it to the current user's $HOME/bin (if it exists, usually user 'oracle'):
== Oracle ==
mv dbcollect $HOME/bin/
</syntaxhighlight>


'''WARNING'''
* Oracle RDBMS 11g or higher (optional)
: If the system is not an Oracle database server, only OS config and performance data will be collected
* Diagnostics pack license OR statspack configured on the database(s)
* Access to the 'oracle' special user or dba privileges ('root' not required)
* Database instances up and running (opened read/write required for AWR/Statspack)
* SYS credentials (hence the 'oracle' or dba user requirement)


Do NOT put dbcollect in non-standard locations such as /tmp, /root etc. The user dbcollect runs as (usually oracle) needs to have access to the dbcollect file (hence it does not work in /root).
== Other Prerequisites ==
It also may not work on mounted paths that have execute disabled (/tmp, /var/tmp, ...)


Do NOT clone the git repository and run it from there (it will result in an error). This to prevent tampering, and weird problems (the zipapp package ensures all required files are contained in a predictable location).
* Some free space in /tmp (or elsewhere, use --tempdir)


= Other requirements =
== Linux ==


== Python ==
On Linux (EL7+) both Python 2 and 3 may be installed. To set the preferred version globally on the system, use <code>alternatives</code>


The dbcollect app runs using a shebang (<code>#!/usr/bin/env python</code>). This means Python needs to be available in the environment. Some systems do not have this by default. Enable it using
<syntaxhighlight lang=bash>
<syntaxhighlight lang=bash>
alternatives --set python /usr/bin/python3
alternatives --set python /usr/bin/python3
</syntaxhighlight>
</syntaxhighlight>
== AIX ==
On IBM AIX, you need to install Python first. You can get python for AIX from
[https://www.ibm.com/support/pages/aix-toolbox-linux-applications-overview AIX Toolbox (IBM)]
Make sure it is executable and available in $PATH.
Simply running <code>python</code> should work.
SAR/Sysstat is usually available on AIX but not enabled by default. In that case there will be no OS performance (SAR) data. {{dbcollect}} can use NMON data as well with some caveats, see TBD.
== Solaris ==
On Solaris, Python should be already available. It may be an older version.
Make sure you can run <code>python</code> from the command line.
SAR/Sysstat is usually available on Solaris (like AIX) but not enabled by default. In that case there will be no OS performance (SAR) data. {{dbcollect}} can use NMON data as well with some caveats, see TBD.
== HP-UX ==
HP-UX has experimental support. Some OS commands require sudo access, run <code>dbcollect --sudoers</code> to install the sudoers file
before proceeding. This will allow {{dbcollect}} to pick up required stuff like CPU model, disk capacities and more.
<code>sudo</code> must be installed for this to work.
== Windows ==
{{dbcollect}} does not yet support Windows. Let me know if you need it.
{{ PageBreak }}
= Installation =
== Downloading DBcollect ==
* Via [https://loadmaster.dirty-cache.com/dbcollect Quick Install Guide]
* Direct download: [https://loadmaster.dirty-cache.com/download DBCollect Download] (application page, not always working) is a direct link to the latest version.
* Via Github Releases: [https://github.com/outrunnl/dbcollect/releases/latest DBcollect releases (GitHub)] On the Github releases page, go to the latest release, open Assets and download dbcollect.
The latest release is always available on the releases page.
== Easy Install ==
If your system has internet (https) access, see [https://loadmaster.dirty-cache.com/dbcollect dbcollect install]
for an easy way to install {{dbcollect}}
== Warning ==
* DBCollect location
: Do NOT put dbcollect in non-standard locations such as /tmp, /root etc. The user dbcollect runs as (usually oracle) needs to have access to the dbcollect file (hence it does not work in /root). It also may not work on mounted paths that have execute disabled (/tmp, /var/tmp, ...)
* Cloning Git Repository
: Do NOT clone the git repository and run it from there (it will result in an error). This to prevent tampering, and weird problems (the zipapp package ensures all required files are contained in a predictable location).
= Test working =
* Run <code>dbcollect -h</code> to see the help page
: If this works, dbcollect is installed correctly
* Run <code>dbcollect --version</code> to show version information
* Run <code>dbcollect --update</code> to update to the latest version
: Optional, requires internet access

Latest revision as of 09:55, 20 November 2024


Page


Summary

Installing DBCollect is pretty simple: download the latest dbcollect zipapp, put it somewhere in $PATH, and make it executable.

On Linux this usually just works, on UNIX it may require some additional steps.

Requirements

Operating Systems

  • Linux
Enterprise Linux 6 (EL6) or higher (RHEL, OEL, CentOS, Rocky Linux)
  • SPARC/Solaris
  • IBM AIX

Python

DBCollect is a Python ZIP app. This means Python needs to be available in the environment to run dbcollect. Some systems do not have this by default.

  • Requires Python (2.6+ or 3.x)
    • On Linux (EL6+) works out of the box
    • On AIX/HP-UX, need to install Python first, Windows not supported
    • On Solaris (and AIX) Python may be a very old version that is not supported

A Python ZipApp is a standard ZIP file with an additional "shebang" prepended: #!/usr/bin/env python

  • This causes dbcollect to be called with whatever ‘python’ version is in $PATH
  • If this does not work, manually run python: /usr/bin/python3 /usr/local/bin/dbcollect <options>
  • Executable can be inspected with unzip –v /usr/local/bin/dbcollect
  • Run simply as root or as ‘oracle’ user/owner

On UNIX (AIX, Solaris) you may have to create a symbolic link or run dbcollect like this: 

python3 /usr/local/bin/dbcollect

If all else fails, you can run dbcollect from python directly: 

python /usr/local/bin/dbcollect
# or
/usr/bin/python /usr/local/bin/dbcollect
# or
python2 /usr/local/bin/dbcollect

Oracle

  • Oracle RDBMS 11g or higher (optional)
If the system is not an Oracle database server, only OS config and performance data will be collected
  • Diagnostics pack license OR statspack configured on the database(s)
  • Access to the 'oracle' special user or dba privileges ('root' not required)
  • Database instances up and running (opened read/write required for AWR/Statspack)
  • SYS credentials (hence the 'oracle' or dba user requirement)

Other Prerequisites

  • Some free space in /tmp (or elsewhere, use --tempdir)

Linux

On Linux (EL7+) both Python 2 and 3 may be installed. To set the preferred version globally on the system, use alternatives 

alternatives --set python /usr/bin/python3

AIX

On IBM AIX, you need to install Python first. You can get python for AIX from AIX Toolbox (IBM)

Make sure it is executable and available in $PATH.

Simply running python should work.

SAR/Sysstat is usually available on AIX but not enabled by default. In that case there will be no OS performance (SAR) data. dbcollect can use NMON data as well with some caveats, see TBD.

Solaris

On Solaris, Python should be already available. It may be an older version. Make sure you can run python from the command line.

SAR/Sysstat is usually available on Solaris (like AIX) but not enabled by default. In that case there will be no OS performance (SAR) data. dbcollect can use NMON data as well with some caveats, see TBD.

HP-UX

HP-UX has experimental support. Some OS commands require sudo access, run dbcollect --sudoers to install the sudoers file before proceeding. This will allow dbcollect to pick up required stuff like CPU model, disk capacities and more. sudo must be installed for this to work.

Windows

dbcollect does not yet support Windows. Let me know if you need it.

Installation

Downloading DBcollect

The latest release is always available on the releases page.

Easy Install

If your system has internet (https) access, see dbcollect install for an easy way to install dbcollect

Warning

  • DBCollect location
Do NOT put dbcollect in non-standard locations such as /tmp, /root etc. The user dbcollect runs as (usually oracle) needs to have access to the dbcollect file (hence it does not work in /root). It also may not work on mounted paths that have execute disabled (/tmp, /var/tmp, ...)
  • Cloning Git Repository
Do NOT clone the git repository and run it from there (it will result in an error). This to prevent tampering, and weird problems (the zipapp package ensures all required files are contained in a predictable location).

Test working

  • Run dbcollect -h to see the help page
If this works, dbcollect is installed correctly
  • Run dbcollect --version to show version information
  • Run dbcollect --update to update to the latest version
Optional, requires internet access
Retrieved from "https://wiki.dirty-cache.com/index.php?title=DBCollect/Install&oldid=618"