Oracle Patch Downloader

The Oracle Patch Downloader can be used to download a list of patches from My Oracle Support (MOS).

It is written in Java and based on HtmlUnit, which has the big advantage that it simulates a browser very closely, including execution of JavaScript. This should (hopefully) allow for more stable web scraping, even if Oracle should decide to change anything in the logon or download procedure. However, using HtmlUnit comes with the disadvantage that this tool requires the inclusion of quite some dependencies, resulting in a relatively big executable jar.

There are other similar tools available, for example oracle_quarter_patch_downloader or getMOSPatch.

The Oracle Quarter Patch Downloader provides additional logic to download important database and GI patches based on the Automated Release Update (ARU) catalog. However, it is controlled by input files, while this tool is completely command-line-driven and, hence, slightly more scripting-friendly.

Prerequisites

• My Oracle Support Account
• JRE version >= 1.8
• JDK version >= 1.8 and Maven version >= 3.5.4 (only if you plan to build the Oracle Patch Downloader yourself from its sources)

Usage

Download the latest release of the Oracle Patch Downloader jar from its GitHub release page. Then start it as

java -jar OraclePatchDownloader-1.1.1.jar

which should bring up the short usage description shown below.

You can use this tool to download one or more patches from MOS. Specify the patch list and other lists on the command line either as comma-separated lists (-x p26749785,6880880) or through multiple repeated options (-x p26749785 -x 6880880) or through a combination of both.

See Section Specifying Patch Platforms and Other Selection Criteria on how to select the patch files to be downloaded.

 -h : --help        help text

 -D : --debug       debug mode

 -Q : --quiet       quiet mode

 -d : --directory   output folder, default user home

 -x : --patches     list of patches (e.g. "26749785,6880880")

 -f : --patchfile   file containing list of patches, one patch per line
                    (e.g. "p26749785", "26749785", "# comment")

 -q : --query |     list of platforms, releases, or languages
 -t : --platforms   (e.g. "226P" for Linux x86-64, "600000000063735R"
                    for OPatch 12.2.0.1.0, or "4L" for German (D))

 -r : --regex       regex for file filter, multiple possible
                    (e.g. "192[23]")

      --authmeth    MOS authentication method, one or more of "Basic",
                    "Legacy", or "IDCS", default "Basic,IDCS"

 -u : --user        email/userid

 -p : --password    password ("env:ENV_VAR" to use password from env)

 -T : --temp        temporary directory

Example

export MOS_PASSWORD="my secret MOS password"
java -jar oraclePatchDownloader-1.1.1.jar -u [email protected] -p env:MOS_PASSWORD \
     -x 26749785 -q 226P -r 1919 -d $HOME/Downloads

Specifying Passwords

Specifying passwords on the command line is inherently unsafe, in particular on multi-user systems. You should better enter the password interactively in the console, which is the default if you do not specify option --password.

Or you can store the password in an environment variable and pass a reference to that variable in the argument to option --password, as shown in the example above.

Specifying Patch Platforms and Other Selection Criteria

Disclaimer: This section is written mainly from the perspective of a user who downloads database patches from MOS. It is not clear how far this section applies to non-database patches as well.

You can use the synonymous command line options -q or -t to specify selection criteria for all patches that you download in one invocation with the Oracle Patch Downloader. At a second stage, you can filter the resulting patch files by patch file name with command line option -r.

Specifying Option -q

The arguments to command line option -q (or -t) are query items consisting of a numeric ID and a one-letter query term, like this:

Query Item	Query ID	Term	Purpose
226P	226	P	select by platform
600000000063735R	600000000063735	R	select by release
4L	4	L	select by language

To calculate the overall patch selection criterion, all query IDs specified per query term are logically ORed, and all different query terms are ANDed. So the following command line options:

-q 226P,23P -q 600000000063735R -q 0L -q 4L

would select the patch files to process according to the following pseudo expression:

(file has platform 226 OR file has platform 23) AND
(file has release 600000000063735) AND
(file has language 0 OR file has language 4)

As a special case, you can specify a query item ALL to select all patch files per patch number for processing:

-q ALL

If you specify -q ALL, then the downloader ignores all other, more specific selection criteria that you also may have specified with option -q.

Specifying Option -r

The arguments to command line option -r are patch file name filters specified as Java regular expressions. If you do not specify any file name filter, then the Oracle Patch downloader processes all patch files previously selected with option -q. Otherwise, the downloader processes only those patch files where it can find at least one of the filter patterns in the file name.

For example, with the following command line options:

-x 26749785 -q ALL -r 1919 -r '192[23]'

the downloader would process the following patch files:

p26749785_1919000DBRU_Generic.zip
p26749785_191900230418WINDBBP_Generic.zip
p26749785_1922000DBRU_Generic.zip
p26749785_192200240116WINDBBP_MSWIN-x86-64.zip
p26749785_1923000DBRU_Generic.zip
p26749785_192300240416WINDBBP_MSWIN-x86-64.zip

Platform IDs for Generic Database Patches

Patch files for generic database patches, like for patch 26749785 from the example above, are registered on MOS with platform ID 2000 and in the patch inventory (inventory.xml) most often with platform ID zero. As seen in the example, you can select these by specifying actually any platform ID to option -q.

Release IDs

Release IDs are not particularly helpful as selection criteria for database patches since they tend to be lengthy and since they change every quarter for most database patches. For example, release "19.22.0.0.0DBRU" has ID 60000000010115240116212, but release "19.23.0.0.0DBRU" has ID 60000000010115240416212. (Yes, there is a pattern in these, and if you feel like using it, go ahead.) So to restrict the downloaded patch files to a particular version, it is usually easier to use command line option -r as explained above.

One exception to this "forget-about-the-release-ID" rule is probably release ID 600000000063735 for OPatch 12.2.0.1.0, since all OPatch 12.2.0.1.0 patch files get released on MOS with that release ID. And all other OPatch patch files labeled "18.*" or "19.*", at least for patching the database, are just OPatch 12.2.0.1.0 in disguise.

Language IDs

Language IDs are even less helpful as selection criteria for database patches since most database patch files are available on MOS only with language ID zero. Accordingly, specifying a non-zero language ID for these will most likely produce an empty search result.

Authentication Methods

The Oracle Patch Downloader can log on to My Oracle Support through different authentication methods:

• Basic

  Basic (but nonetheless secure) HTTPS authentication. In contrast to the following methods this one does not involve execution of JavaScript and, hence, is the fastest of all methods.

• Legacy

  "Single-paged logon", legacy authentication method. Most likely obsolete since May 2024.

• IDCS

  "Two-paged logon", authentication based on Oracle Identity Cloud Service.

By default, the downloader tries to log on through basic authentication first and, if that fails, then through IDCS-based authentication.

To override the authentication methods being attempted, specify one or more of the above methods separated by commas to command line option --authmeth. For example, to force the exclusive use of IDCS-based authentication, specify --authmeth IDCS.

Debug Mode

In debug mode the Oracle Patch Downloader dumps all HTML pages that it processes to the output folder. In addition, it logs HTTP requests and responses to a log file, also placed in the output folder. The downloader uses file names dump-<hex-timestamp>-<page-title>.xml for dumps of regular pages and file names error-<hex-timestamp>-<page-title>.xml for dumps of pages it could not process.

Note that the log file most certainly contains sensitive data, likewise for the page dumps.

The Downloader uses the APIs provided by HtmlUnit to dump HTML pages and log HTTP requests and responses. The information retrieved through these APIs differs from the actual data sent over the wire.

Installation

To build the Oracle Patch Downloader from its sources execute the following commands:

git clone [email protected]:ojehle/H35_OraclePatchDownloader.git
cd H35_OraclePatchDownloader
mvn package
java -jar target/oraclePatchDownloader-1.1.1.jar

The resulting jar is self-contained and does not require any additional dependencies during runtime.

Contributing

As experience has shown, even the smallest projects need at least some standards. If you plan to contribute more than just a couple of changed lines to the Oracle Patch Downloader, then here are some instructions on how to prepare your IDE or editor to not cause too much reformatting trouble.

Eclipse

This project comes with project specific settings named Default both for the Eclipse Java formatter and for Eclipse Java cleanup. Plus it should configure Eclipse such that it runs a format on each save of a Java source file.

Some notes on that:

• The project specific formatter and cleanup settings have been tested on Eclipse version 2024-03 (4.31.0).

• Ensure that the project specific settings are enabled by comparing Project → Properties with the following screenshots:

  

  

  

  "Enable project specific settings" should be checked on all of these, and on the former two screenshots "Unmanaged profile 'Default'" should be selected as active profile.

• Probably better avoid doing full cleanups as triggered by Source → Clean Up.... If you really feel you would need to clean up the sources in that way, do not use a custom profile, but rather the configured profile "Unmanaged profile 'Default'" from the project as shown in the following screenshot:

  

• Unfortunately, Eclipse as version 4.31.0 with the given settings seems to behave slightly inconsistently when it indents or formats Java code: A continuation line of some wrapped line might get indented differently with TAB/Ctrl+I ("Correct Indentation") and with Shift-Ctrl-F ("Format"). Mainly for that reason we have switched on source code formatting in the project's save actions.

• It is not clear to what extent the user-configured workspace settings from Window → Preferences can override or conflict with the project specific settings. If in doubt, better review your changes in the Git perspective of Eclipse before committing them.

Emacs

Install package smart-tabs-mode from MELPA and add the following to your Emacs initialization file:

(smart-tabs-insinuate 'java)

Some notes on that:

• The Emacs configuration has been tested on Emacs 28.2.

Platform and Language Codes

Code	Platform
537P	Acme Packet 1100
529P	Acme Packet 3820
540P	Acme Packet 3900
561P	Acme Packet 3950
530P	Acme Packet 4500
538P	Acme Packet 4600
560P	Acme Packet 4900
534P	Acme Packet 6100
531P	Acme Packet 6300
551P	Acme Packet 6350
527P	Acme Packet OS
512P	Apple iOS
293P	Apple Mac OS X (Intel) (32-bit)
522P	Apple Mac OS X (Intel) (64-bit)
421P	Apple Mac OS X (PowerPC)
553P	Apple macOS
559P	Apple macOS ARM (64-bit)
516P	Embedded Linux ARM
521P	Embedded Linux on cnMIPS
519P	Embedded Linux on POWER e500v2
517P	Embedded Linux on POWER Systems
294P	Embedded Linux SH4
515P	Embedded Linux x86
228P	FreeBSD - x86
361P	Fujitsu BS2000
312P	Fujitsu BS2000/OSD (SQ series)
285P	Fujitsu BS2000/OSD (SX series)
504P	Fujitsu MSP-EX
513P	Google Android
554P	Google Android 9
549P	HP NonStop (Guardian) on x86
548P	HP NonStop (OSS) on x86
304P	HP NonStop Itanium (Guardian)
308P	HP NonStop Itanium (OSS)
303P	HP NonStop S-series (Guardian)
89P	HP OpenVMS Alpha
243P	HP OpenVMS Itanium
1P	HP OpenVMS VAX
87P	HP Tru64 UNIX
197P	HP-UX Itanium
278P	HP-UX Itanium (32-bit)
999P	HP-UX PA-RISC (32-bit)
59P	HP-UX PA-RISC (64-bit)
299P	ia64
319P	IBM AIX on POWER Systems (32-bit)
212P	IBM AIX on POWER Systems (64-bit)
43P	IBM i on POWER Systems
211P	IBM S/390 Based Linux (31-bit)
30P	IBM z/OS on System z
314P	IBM z/VM on System z
227P	IBM: Linux on POWER Big Endian Systems
542P	IBM: Linux on POWER Little Endian Systems
209P	IBM: Linux on System z
297P	iTron
528P	Linux ARM 32-bit VFP HardFP ABI
523P	Linux ARM 32-bit VFP SoftFP ABI
564P	Linux ARM 64 bit (DB19 OL79)
541P	Linux ARM 64-bit
214P	Linux Itanium
556P	Linux MIPS 64-bit
525P	Linux SPARC
46P	Linux x86
226P	Linux x86-64
99999P	Metadata-Only Patch
912P	Microsoft Windows (32-bit)
558P	Microsoft Windows AArch64
208P	Microsoft Windows Itanium (64-bit)
539P	Microsoft Windows Phone
233P	Microsoft Windows x64 (64-bit)
276P	Monta Vista x86
277P	Monta Vista x86-64
536P	Net-Net 4250
533P	Net-Net 9200
547P	Netra Server X5-2 for Communications
535P	Netra X3-2 for Acme Packet
1234P	NLS Generic Platform
2234P	NLS Merged Translations
3234P	NLS Pseudo Translation
313P	OpenSolaris
309P	Oracle JRockit Virtual Edition x86 (32-bit)
311P	Oracle JRockit Virtual Edition x86-64 (64-bit)
506P	Oracle Solaris Express
453P	Oracle Solaris on SPARC (32-bit)
23P	Oracle Solaris on SPARC (64-bit)
173P	Oracle Solaris on x86 (32-bit)
267P	Oracle Solaris on x86-64 (64-bit)
511P	QNX UNIX
28P	SCO UNIX
505P	SPARC
550P	StorageTek Hardware
619P	Stratus PA-RISC VOS
26P	Symbian EPOC
555P	Talari
532P	Tekelec
316P	Unisys OS 2200
520P	VxWorks
280P	x86 32 bit
282P	x86 64 bit

Code	Language
67L	Albanian (SQ)
8L	Arabic (AR)
26L	Brazilian Portuguese (PTB)
101L	Bulgarian (BG)
3L	Canadian French (FRC)
102L	Catalan (CA)
103L	Croatian (HR)
66L	Cyrillic Kazakh (CKK)
62L	Cyrillic Serbian (CSR)
30L	Czech (CS)
5L	Danish (DK)
6L	Dutch (NL)
118L	ESTONIAN (ET)
7L	Finnish (SF)
2L	French (F)
4L	German (D)
104L	Greek (EL)
107L	Hebrew (IW)
28L	Hungarian (HU)
106L	Icelandic (IS)
46L	Indonesian (IN)
108L	Italian (I)
15L	Japanese (JA)
16L	Korean (KO)
29L	Latin American Spanish (ESA)
63L	Latin Serbian (LSR)
119L	LATVIAN (LV)
109L	Lithuanian (LT)
10L	Norwegian (N)
110L	Polish (PL)
18L	Portuguese (PT)
111L	Romanian (RO)
112L	Russian (RU)
14L	Simplified Chinese (ZHS)
113L	Slovak (SK)
114L	Slovenian (SL)
11L	Spanish (E)
13L	Swedish (S)
115L	Thai (TH)
117L	Traditional Chinese (ZHT)
116L	Turkish (TR)
37L	UK English (GB)
39L	Ukrainian (UK)
43L	Vietnamese (VN)
999L	Worldwide Spanish (ESW)