How To Update The Serial Number On Oracle X64 Platforms

7/nine

4 Troubleshooting STB and STB Components

This chapter describes common troubleshooting steps for Oracle Services Tools Packet (STB) and associated components. The following troubleshooting resolution steps are provided:

• Troubleshooting STB Installation

• Troubleshooting Explorer Collections

• Troubleshooting SNEEP

4.1 Troubleshooting STB Installation

Refer to the sections below for information on troubleshooting your Oracle Services Tools Bundle installation:

• STB Error Messages

• STB Check Sum Fault

• The Installation Log

• Possible Issues with IPS Installation

• Tracing an STB Installation

• Package Removal Error in STB Log File (Solaris 11 or Afterwards, Only)

4.1.1 STB Error Messages

It is possible that an error occurs when you are running Explorer. Near STB install error messages accept the format STB-nnnnn. You tin can get more than data on this equally follows:

explorer -X STB-nnnnn              

For case:

• Error message: STB-02004: Cannot publish the IPS repository

• Cause: The IPS repository in the installation area cannot exist published.

• Action: Check that the origins of all agile IPS repositories on this box are reachable.

iv.1.ii STB Check Sum Error

Fault message: STB-00010: Incorrect payload cheque sum

Crusade: This error message indicates that the STB payload, which was extracted from the installation script, does not have the correct cheque sum. If the cheque sums do not match, and so it is most probable that the STB bundle has become corrupted. Possible causes for such a file corruption include:

• If the STB bundle was downloaded onto a non-Solaris machine (such as a PC), and then copied to the Solaris automobile, and so the copying process from i system to the other may have encountered an mistake. Such an error may contribute to the file corruption.

• If the STB package was downloaded via FTP to a Solaris system, then it could get corrupted if y'all did non use bin (binary) and instead transferred the file every bit ASCII text.

• If the STB packet was downloaded to a PC, then the download settings of the PC may have automatically unzipped the package, which could cause the file to become corrupted.

Action: In virtually cases, you can download the STB bundle over again and retry the installation. If possible, download the STB bundle from Oracle directly to the Solaris system that you want to install STB.

1. Select a platform to download from Knowledge Article 1153444.1. You can access this commodity in My Oracle Support:

   https://support.oracle.com                  

2. From the Patchset README, note the size of the STB bundle:

   

   In this example for the SPARC 64-bit platform, the size is 44M (46960701 bytes).

3. Click the Download button to download the file.

4. In the directory where you lot downloaded the STB bundle, run the ls -l command. The output for this control would look like this:

   # ls -fifty  -rw-r--r--   one root    root   46960701 Jun 15 12:12 <patchID>_SOLARIS64.zip                  

The file size from the ls -fifty control output should match the size you note in the patchset README.

4.i.iii The Installation Log

When errors occur during the installation, you can review the installation log, which keeps track of all attempted installations. The log is located in /var/log with a file name install_stb-v<STB-VERSION>.log (for case, /var/log/install_stb-v7.0.log). This log tin can provide you lot with more than data on an issue. For every install that you practice, the installer writes log messages to this file. Brand sure that y'all only inspect the relevant (probably last) part of the log file.

Note:

It is besides possible to trace the STB installation. See Tracing an STB Installation.

4.1.4 Possible Problems with IPS Installation

Other IPS installation fault examples include:

• STB-02004: Cannot publish the IPS repository

• STB-02023: IPS installation of back up/sneep failed

• STB-02023: IPS installation of support/act failed

• STB-02013: Cannot unpublish the IPS repository

Typically y'all will observe lines like this in the log file:

Installing support/explorer ...     Errors were encountered while attempting to retrieve     bundle or file information for the requested operation.    Details follow:     http protocol error: code: 404 reason: Not Institute    URL: 'http://host1.mycompany.com/solaris11/release/    solaris/manifest/0/back up%2Fexplorer@6.10%2C5.11-12.4.17%3A20120417T105756Z'.    (happened 4 times)     STB-02023: IPS installation of support/explorer failed              

If you get such an error message while installing STB and see lines similar the above ones in the log file, the nigh likely cause is that at that place is an active IPS repository on the organization, which can no longer exist reached. To list the active IPS repositories, run:

# pkg publisher              

Output should look similar:

PUBLISHER             TYPE     Status   URI solaris               origin   online   file:///var/tmp/stb/extract/ipsrepo/              

In this case, nosotros meet a file-based IPS repository point to /var/tmp/stb/extract/ipsrepo. If someone has removed this file, the IPS deportment, triggered by the STB installation, will all neglect.

A similar case can happen when an HTTP-based repository can no longer exist reached.

The remedy in this cases is to "conciliate" the non-reachable IPS repository past:

pkg set-publisher –G <ips repository> <publisher>              

four.one.five Tracing an STB Installation

You might find the STB installation slow or might fifty-fifty think the installation is hanging.

If you matter this is the case, start by using the -verbose pick with the ./install_stb.sh script so that you go more data of what the installer is doing currently.

If yous are still convinced that the installer is hanging, please transport a full trace to Oracle. Yous can practice this past:

./install_stb.sh –verbose –trace 2>&1 | tee /tmp/trace              

Practise non forget to blazon I(install) to showtime the bodily installation.

4.1.6 Package Removal Mistake in STB Log File (Solaris eleven or Later, Simply)

When installing Oracle STB eight.02 on a Solaris 11 domain, the command line output shows:

Identified actions for Oracle Explorer Information Collector (eight.02):  - Remove -SUNWexplo  - Remove -SUNWexplu  - Remove -SUNWrda              

The installation log contains letters like:

- Failed to remove SUNWexplo pkgrm: Fault: unable to change current working directory to </var/sadm/pkg/SUNWexplo/install>              

Similar messages can exist nowadays for SUNWexplu, SUNWrda, and SUNWsneep. The message is non always present for all packages.

These messages are normal.

The rationale is that it is possible, but not advisable, to install SVR4 packages on Solaris 11 or after. When SVR4 packages were installed, the STB installer will remove the SVR4 package and replace information technology with the IPS counterpart.

Since the IPS packages have legacy entries, it is non easy to exam that the SVR4 packages are already on the system to avoid the message.

Having the legacy entries in the IPS packages, prevents the installation of the SVR4 packages on Solaris 11 or later.

4.2 Troubleshooting Explorer Collections

This section addresses known bug and workaround solutions for the Oracle Explorer Information Collector. The following Explorer-specific troubleshooting steps are provided:

• Tracing an Explorer Collection

• Finding Explorer Modules in RDA

• Explorer Mistake Messages

• ACT Truncated Output

• Running Oracle Explorer Data Collector in the Background

• Explorer Hangs on Zone pkg.ksh (Solaris Only)

iv.2.i Tracing an Explorer Collection

When an Explorer collection is running, there are three modes of operation:

• The Explorer engine is running, which sets up the collection environs, starts upwardly the Explorer modules and RDA, and does the mail service-processing (packaging the collection results and possibly forwarding them).

• One of the native Explorer modules is running (for instance, the sysconfig module).

• RDA is started, which sets upwardly the RDA collection surround, runs the RDA modules for Explorer, and converts the RDA results to Explorer format.

You lot tin can trace each of these modes of operations:

• Engine tracing: set the EXP_TRACE environment variable to a non-empty value.

• Native Explorer module tracing (for example, the sysconfig module): set the EXP_TRACE_<module name in upper case> environs variable to a non-empty value. For example:

  EXP_TRACE_SYSCONFIG=1                  

  To actuate this tracing, information technology is necessary to also actuate the Explorer engine tracing (EXP_TRACE=1).

• RDA Explorer tracing: add Set EXP_TRACE_RDA=i and export it and/or add together RDA_XPLR_TRACE="1" equally an entry to the rdainput.txt file in the Explorer configuration directory.

four.2.ii Finding Explorer Modules in RDA

To find out if the collection of an Explorer module is delegated to RDA, issue the following statement:

# <RDA location>/rda.sh -X Explorer can              

This will typically list a series of modules. Typical output would look like:

samba ib lp etcextended              

You can find out the RDA name of the module by inspecting the mrc.cfg file. The mrc.cfg file is located in the collect/EXPLORER subdirectory of the RDA base of operations directory for Oracle Explorer Data Collector vii.3 or higher. Older versions of Explorer keep it in the modules subdirectory of the RDA base directory.

Note:

For Oracle Explorer Data Collector 7.three or higher, the collection of all modules is delegated to RDA.

4.2.3 Explorer Error Letters

All Explorer error letters are at present in the following format:

EXPL-nnnnn: ....              

To become more information on an Explorer error message, run:

explorer –X EXPL-nnnnn              

For example:

explorer –10 EXPL-00032              

Output will wait like this:

EXPL-00032: The temporary directory is not a directory  Cause: The specified location to be used as temporary directory is not a directory  Action: Remove the specified location or utilise an alternate location by setting the EXP_TMP environment variable.              

four.2.iv ACT Truncated Output

Oracle Explorer Data Collector Release 6.6 may truncate the Act output on certain Solaris 10 systems that has ACT versions between 8 and eight.14 installed. This truncation is done to avoid file system overflow described in Act problems 6897128.

Work around: Download and install the latest version of Human activity into the system.

4.2.5 Running Oracle Explorer Data Collector in the Background

If y'all attempt to run Oracle Explorer Data Collector as a background process in a final without -SR, -esc, or -case options, then the tool may wait for you to input (on Explorer type) at the groundwork. In such cases, the tool will not complete its execution and volition wait on your input.

Work around: You may note a stopped message on the terminal indicating the suspended process at the background. You will need to bring the suspended process to foreground then provide the appropriate response to all the Oracle Explorer Data Collector to keep its run.

four.two.half dozen Explorer Hangs on Zone pkg.ksh (Solaris Just)

Launching the Oracle Explorer Data Collector from crontab can sometimes cause the awarding to hang. This problem is caused by an OS issue, which does not allow crontab directly or indirectly to run:

/usr/sbin/zlogin <zone> /usr/bin/pkginfo -l              

Run into bug 14674009 in My Oracle Support for more than details:

https://support.oracle.com              

The installation of the latest CPU patch and patch 119254-86 solves this consequence on Solaris 10.

4.3 Troubleshooting SNEEP

This department lists the various errors and issues encountered from using SNEEP:

• SNEEP Performance

• EEPROM nvramrc Formatting with Veritas VxVM

• SNEEP Returns Series Number unknown

• Preserving SNEEP Settings After EEPROM Restore or Supercede

• Preserving SNEEP Settings Subsequently OBP Firmware Update

• Unable to Set hostname and hostid Tags

• Overriding the Size Limit of the EEPROM nvramrc Variable

• SNEEP Reports a Bad String

4.3.i SNEEP Performance

On most machines, SNEEP responds immediately; however, with other machines it tin take longer.

Some Sunday platforms provide the product serial number in the output of the prtdiag command. The prtdiag command communicates with the Platform Information and Control Library (PICL) daemon, and picld may need to communicate with the system controller or ILOM to become electric current hardware information.

This communication may take a few seconds, with xv seconds not uncommon.

If SNEEP does not respond for a infinitesimal or more than, the likely crusade is that picld is non still ready, or perchance never started.

SNEEP may wait for picld to respond, retrying prtdiag several times earlier giving upwardly. By default, at that place will be messages in the system log /var/adm/letters indicating these retries.

On Solaris 10, you can find out if picld is enabled and functioning by checking the picl service. It should look something like this:

$ svcs picl  STATE     TIME         FMRI  online    8:10:48      svc:/system/picl:default              

Prior to Solaris ten, you tin verify that the picld process is running, but that does not necessarily hateful that information technology is functioning properly:

$ ps -ef | egrep 'PI[D]|picl[d]'  UID    PID   PPID C   STIME   TTY   Time   CMD}}  root   296   one    0   January 11  ?     ii:26   /usr/lib/picl/picld              

4.three.2 EEPROM nvramrc Formatting with Veritas VxVM

If yous use Veritas VxVM to manage your storage volumes, and subsequently making a alter to the root book, y'all may find that later using SNEEP, the EEPROM nvramrc may look strange (for example, everything is flattened onto one long line).

This is an instance of the upshot of a change made to Veritas VxVM's treatment of the EEPROM nvramrc variable. This change has been returned to normal with recent VxVM "Rolling Patches" for their Maintenance Packs (MP). Versions of SNEEP before update 1.79 were non prepared for this alter, and could lose track of the device allonym names for the root volume, which VxVM stores in the eeprom nvramrc. This could even cause a system to fail to boot without manual assist.

SNEEP update i.92 and in a higher place are not bothered by this modify, and tin can even repair the nvramrc variable if y'all either fix a value with SNEEP, or expect for the automatic repair at the next reboot.

Note:

It is very important to upgrade to at least SNEEP update i.92.

Usually, the easiest value to ready is the series number, although you could make up any tag and value to fix:

sneep                   # get the serial number sneep -south serialnumber   # set information technology over again to correct the nvramrc  or  sneep -t annihilation -south anyvalue    # brand a change to nvramrc sneep -t annihilation -south ""          # delete the made-up entry              

If sneep -t reveals whatsoever strange SNEEP tags made accidentally past SNEEP releases lower than 1.92 (tags such as devalias or cr), you can remove them by setting them to an empty value.

Upgrade to SNEEP update 1.92 or newer, then remove them by setting them to a null value.

$ sneep -t cr -s ""  $ sneep -t devalias -due south ""              

iv.3.three SNEEP Returns Serial Number unknown

Depending on your hardware platform, running SNEEP may return a value of unknown. About older Dominicus platforms take no way to know or written report their serial number until after the serial number is manually entered into EEPROM with SNEEP.

Almost newer Sun platforms natively provide series number data through some hardware-based mechanism which tin can be read past SNEEP, but there are a diversity of means in which this data is provided.

While SNEEP is able to take advantage of many of the mechanisms which provide "hardware" serial number support, at this time, SNEEP does not know how to find serial data on every platform which can provide information technology in some fashion.

As office of the ongoing development of SNEEP, back up is being added for as many of these methods and platforms as chop-chop as possible; notwithstanding, non every mechanism will be supported.

For example, some platforms require a user to interactively log in to a system controller over a public network in order to obtain the serial. Even if SNEEP was capable of this, the configuration required for this is across what most users are willing to do.

Information technology is easier for the user to log in manually one time to obtain the serial, and then utilise SNEEP to store that serial into the EEPROM where it can be retrieved easily in the hereafter.

If yous are not using the latest SNEEP update, your version may not know how to get the serial number from your machine. A newer update of SNEEP might exist able to do this.

In that location is also the possibility that your machine is ane of the platforms for which information technology is e'er necessary that the user inform SNEEP of the serial number of the motorcar using:

sneep -due south serialnumber              

This should only need to exist done i fourth dimension, and after that, SNEEP volition get in very easy for the user to become the serial number any time that information technology is needed.

4.3.four Preserving SNEEP Settings After EEPROM Restore or Supersede

Your SNEEP settings are preserved if you demand to restore your EEPROM to default settings or if y'all demand to replace the EEPROM entirely. SNEEP maintains a backup file /etc/default/SUNWsneep in which it keeps a re-create of all settings. In instance OBP defaults have been restored or if the EEPROM has been replaced without preserving the contents, SNEEP automatically restores the SNEEP EEPROM settings from the fill-in file when the system is rebooted. Alternatively, the data can be recovered but by asking for it with SNEEP, and so setting the returned value again with SNEEP.

SNEEP has options designed to get in easy to recover data with very trivial endeavour, and under nearly circumstances it will be able to automatically recover the data at the next organization boot. When this is done at arrangement startup, SNEEP volition log a message to tell you if the EEPROM is not consistent with the backup and whether or not it was recovered.

See the SNEEP human being folio for the usage of -T, -d, and -P options.

4.3.v Preserving SNEEP Settings Afterwards OBP Firmware Update

Your SNEEP settings are preserved if you need to update your OBP firmware. It is possible to lose the EEPROM settings in an OBP firmware update, but while it in one case was common, these updates have been very reliable and safe for several years.

If there is a problem, SNEEP will automatically recover the values from the backup.

4.3.6 Unable to Prepare hostname and hostid Tags

The hostname and hostid information is ever retrieved from the hostname and hostid commands. To maintain data consistency, SNEEP no longer allows users to prepare these tags, which can and did create confusing and alien situations.

If SNEEP sees that tags exist, they are erased and further requests for the values of these tags are satisfied using the advisable commands.

4.three.7 Overriding the Size Limit of the EEPROM nvramrc Variable

By default, SNEEP will foreclose you from using too much nvram, but most platforms provide more than nvram chapters than SNEEP will allow you to consume. If yous are certain that there is more available on your detail platform, you tin override SNEEP. If y'all need a few more than bytes than SNEEP allows by default, there is no danger.

Notwithstanding, if you exceed the true maximum capacity of the nvramrc variable, some SPARC systems will testify no immediate effect, while others will drop into OBP. After that, the system may not kick until the EEPROM defaults are restored:

( OBP> prepare-defaults )              

A Solaris x86 system may not accept an immediate reaction, simply may neglect to boot later on.

Caution:

Every bit these are very serious consequences, it is strongly recommended that you lot do Not override the condom limits. If you do so, you are entirely responsible for whatsoever arrangement outage or other damages. SNEEP takes care to notify you lot of this.

4.3.8 SNEEP Reports a Bad Cord

If SNEEP reports a Bad Cord and is unresponsive, then yous probably have an onetime version of SNEEP and are using a locale or language setting which involves UTF-8. The default tr programme used in SNEEP did not work in UTF-eight locales. This is no longer a problem after SNEEP 2.5_R1.75.

Source: https://docs.oracle.com/cd/E35557_01/doc.81/e35226/ch4_troubleshooting.htm

Posted by: michaelmathe1943.blogspot.com

Share this post