Oracle® Enterprise Manager Ops Center

Release Notes

12c Release 4 (12.4.0.0.0)

F17670-01

April 2019

Oracle Enterprise Manager Ops Center is Oracle's comprehensive system management solution for managing the physical and virtual operating systems, servers, and network and storage devices (or assets) in your data center. With Oracle Enterprise Manager Ops Center you can use the discovery and management features to add your data center assets to the list of managed assets. You can provision, update (patch), monitor, and manage the physical and virtually-managed assets in one or more of your data centers from a single console, regardless of where the asset is located.

Caution:

If you are upgrading an existing version of Oracle Enterprise Manager Ops Center, refer to the Oracle Enterprise Manager Ops Center Upgrade before attempting to upgrade. All instructions must be followed as stated in the Upgrade document.

Note:

If you use Ops Center to monitor Oracle SuperCluster, do not upgrade to Oracle Enterprise Manager Ops Center 12c Release 4 (12.4.0.0.0). Oracle SuperCluster is supported only by a subset of Oracle Enterprise Manager Ops Center versions. Refer the Oracle Enterprise Manager Ops Center Certified Systems Matrix to verify if Oracle SuperCluster platform is supported by Oracle Enterprise Manager Ops Center version that you use.

This document includes the following information about the Oracle Enterprise Manager Ops Center 12c Release 4 (12.4.0.0.0) software:

1.1 Where to Find Documentation

The following documents are related to installing or upgrading to Oracle Enterprise Manager Ops Center 12c Release 4 (12.4.0.0.0).

You can download the documents from the Oracle Enterprise Manager Ops Center Documentation Library at https://docs.oracle.com/cd/ops-center-12.4/index.htm

1.2 New Features Included In This Release

Oracle Enterprise Manager Ops Center 12c Release 4 (12.4.0.0.0) contains multiple new features and enhancements.

See the Oracle Enterprise Manager Ops Center What's New In This Release document for an overview of the new features and enhancements.

1.3 Acquiring the Software

You can download the product for installation from the Oracle Technology Network.

Navigate to http://www.oracle.com/technetwork/oem/ops-center/index.html for downloading the software.

1.4 Prerequisites

Before installing the Oracle Enterprise Manager Ops Center software, you should plan for the installation and how you will use the software.

See the Preparing Your Environment chapter in Oracle Enterprise Manager Ops Center Installation for Oracle Solaris Operating System or the Preparing Your Environment chapter in Oracle Enterprise Manager Ops Center Installation for Linux Operating Systems for prerequisite information.

OCDoctor is a tool utility that checks systems for installation prerequisites, troubleshoots issues, and tunes systems for the Oracle Enterprise Manager Ops Center installation. For more information, see the OCDoctor chapter in Oracle Enterprise Manager Ops Center Administration. The guide is available in the Oracle Enterprise Manager Ops Center Documentation Library at https://docs.oracle.com/cd/ops-center-12.4/index.htm.

1.5 Known Issues

This section explains the known issues that you might encounter when installing, or upgrading to, Oracle Enterprise Manager Ops Center 12c and the possible workarounds available.

For information on known issues related to installation and postinstallation configuration, see the Oracle Enterprise Manager Ops Center Release Notes.

The following are known issues that you might encounter:

1.5.1 Oracle Linux Installer Does Not Check for Required Packages

The installation does not check for the required packages such as ksh and zlib.i686 on Oracle Linux.

Workaround

Install the ksh and zlib.i686 packages before or after the Oracle Linux installation.

1.5.2 Oracle Solaris 10 Branded Zone Creation Fails: Global Zone Cannot Find OS Image

If a Global Zone is in a private network and cannot directly access the Enterprise Controller, it cannot use ISO images stored in the Enterprise Controller library.

Workaround

Copy the ISO image to the target system and use the local copy.

1.5.3 Discovery of M6 Server Can Fail if PDOM Does Not Respond Correctly

If the PDOM hangs or is slow to respond during an M6 discovery, the system can be displayed as a generic server.

1.5.4 Newly Configured Network From Outside Ops Center Does Not Appear in the User Interface

If you configure a private network on the OS level outside of Oracle Enterprise Manager Ops Center, and the OS instance is the Enterprise Controller OS which is not agent managed, the newly configured network does not appear in the user interface.

Workaround

Restart the Enterprise Controller.

1.5.5 Execute Operation Is Not Enabled For Groups of Assets if One Asset Is Offline

The "Execute Operation" action is greyed out for groups of Operating System assets if at least one of them is in an unknown/offline state.

Workaround

Select the Operational Plan first and then select the group as a target.

1.5.6 Install Fails With Permission Error While Installing SUNWuces.pkg

An installation can fail during the Install application packages step when SUNWuces.pkg failed to install. This occurs when the "users" group is an LDAP managed group, not a local group.

Workaround

The first workaround is to continue the installation if it is in progress, or restart the installation if it has been interrupted.

The second workaround is to manually create a local "users" group and manually install the SUNWuces.pkg.

1.5.7 OS Provisioning From New Remote Proxy Controller Repeatedly Fails with HTTP 500 Error

When attempting to OS Provision LDOM guests from a newly installed Proxy Controller, the AI downloads keep failing with an http 500 error showing different URL locations that are noted as the failure points.

For example:

23:23:19 Download: 25862/71703 items 251.0/723.4MB 34% complete (7.4M/s)
23:23:24 Download: 27804/71703 items 338.3/723.4MB 46% complete (14.9M/s)
23:23:29 Download: 30360/71703 items 354.3/723.4MB 48% complete (10.2M/s)
23:23:34 Download: 32781/71703 items 376.3/723.4MB 52% complete (3.8M/s)
23:23:39 Download: 33078/71703 items 403.1/723.4MB 55% complete (4.9M/s)
23:23:45 Download: 33080/71703 items 408.2/723.4MB 56% complete (3.3M/s)
23:23:48 Error occurred during execution of 'generated-transfer-1197-1' checkpoint.
23:23:48 Failed Checkpoints:23:23:4823:23:48 generated-transfer-1197-123:23:4823:23:48 Checkpoint execution error:23:23:4823:23:48 http protocol error: code: 500 reason: Internal Server Error23:23:49 URL: 'http://<IP Address>:8003/IPS/solaris/file/1/bf7046c17838279aaf2639f92317deaf14349673'23:23:4923:23:49 Automated Installation Failed. See install log at /system/volatile//install_logAutomated Installation failedPlease refer to the /system/volatile/install_log file for details

Workaround

Edit the Apache KeepAliveTimeout and ProxyTimeout variables to increase them. Doubling the values should resolve the issue, but both can be further increased if necessary.

  • The KeepAliveTimeout variable is in the /var/opt/sun/xvm/uce/etc.opt/server/uce_server/httpd.conf file.

  • The ProxyTimeout variable is in the /var/opt/sun/xvm/uce/etc.opt/server/uce_server/ips-proxy.conf file.

1.5.8 Oracle Solaris 11 Kernel Zones With the Same Hostid Are Correlated

Ops Center relies on a kernel zone hostid to identify it uniquely from the global zone or the kernel zone, and to persist its configuration among different global zones.

If 2 kernel zones have the same hostid, they will correlate together, which means that there will be only 1 kernel zone listed in the UI, and management of a kernel zone may target the other kernel zone with the same hostid.

Having the same hostid is possible if the user manually sets the same hostid for 2 kernel zones, or if the same hostid is generated for 2 KZs configured in the same second.

Workaround

Manually change the hostid for both of the kernel zones, then refresh the global zone.

1.5.9 Logical Domains Can Be Created Without Sufficient Memory To Run

The Logical Domain creation process allows the creation of LDoms with 1 GB of memory, but Oracle Solaris 11.2 requires a minimum of 2 GB. The creation job fails and the LDom cannot be used.

Workaround

When creating a Logical Domain with Oracle Solaris 11.2, do not assign it less than 2 GB of memory.

1.5.10 After Restoring, a Duplicate Enterprise Controller Appears in the User Interface

After backing up the Enterprise Controller and restoring from that backup on a new system with the same hostname, a duplicate Enterprise Controller system is displayed in the UI in an unknown state.

Workaround

Delete the missing assets from the user interface. If you cannot delete the missing assets from the user interface, delete the missing assets from CLI.

  1. Launch the CLI on the system hosting the Enterprise Controller.

    /opt/SUNWoccli/bin/oc
    
  2. Type connect to connect to the Enterprise Controller.

  3. Type gear to enter the gear mode.

  4. Type list to list the assets to get the UUID of the asset you want to delete.

  5. Delete the assets by using the following command:

    delete -U <UUID> -f
    

1.5.11 Zstat Produces Excessive Amounts of Data in /var/adm/exacct/zstat-process

On Oracle Solaris 10 systems, zstat on occasions produces excessive amounts of data in /var/adm/exacct/zstat-process resulting in filesystem fillup in production environment.

Workarounds

A temporary workaround is to run the following command:

svcadm restart svc:/application/xvm/zstat:default

However, this only clears the file temporarily, it will continue to grow.

A permanent workaround tells zstat not to record detailed tracking of system processes. Run the following commands on the system:

% mkdir -p /etc/opt/sun/oc
% echo "zstat_exacct_allowed=false" > /etc/opt/sun/oc/zstat.conf
% chmod 755 /etc/opt/sun /etc/opt/sun/oc
% chmod 644 /etc/opt/sun/oc/zstat.conf
% chown -Rh root:sys /etc/opt/sun/oc
% /opt/SUNWxvmoc/bin/agentadm stop
% /opt/SUNWxvmoc/bin/agentadm start

1.5.12 Cannot Correlate Kernel Zone Instances if No Hostid in Kernel Zone Operating System

If a hostid generated in zonecfg is not within the range 0x0 to 0x7fffffff, then there is no hostid in the kernel zone operating system.

As the hostid is the correlation entry in Oracle Enterprise Manager Ops Center, following are the consequences:

  • If the kernel zone is running, the kernel zone asset is discovered or is removed from the Oracle Enterprise Manager Ops Center as you get the hostid from the OS.

  • If the kernel zone is not running, the kernel zone asset is recreated with the hostid set.

  • If the OS of the kernel zone is discovered, an OS is created but it does not correlate with the kernel zone asset as there is no hostid.

Workaround 1

Update the kernel zone to a SRU when the hostid problem is fixed. Remove the operating system of the kernel zone from the Oracle Enterprise Manager Ops Center and rediscover the asset with or without agent.

Workaround 2

  1. Change the hostid in the kernel zone by providing a hostid that it supports using the command:

    zonecfg -z <kz> set hostid=...
    
  2. Remove the image UUID by using the command:

    zonecfg -z <kz> remove attr name=image-uuid
    
  3. Refresh the global zone. This destroys the current kernel zone model and recreates a new one with a new hostid.

  4. If the operating system is running an agent, unconfigure and reconfigure the agent. If the OS is discovered agentlessly, delete the OS asset in Oracle Enterprise Manager Ops Center and rediscover the OS.

1.5.13 USB VNIC Interfaces Are Not Filtered and Causes Various Issues

After discovering a system with a USB VNIC, an extra ethernet fabric appears in the domain model.

During the system discovery, the Oracle Enterprise Manager Ops Center must skip the USB interfaces. But, as the system is discovered with a USB VNIC with an IP plumbed on it, all the USB interfaces are considered as regular interfaces and causes an extra ethernet fabric to appear in the domain model. This leads to incorrect domain model and affects the network functionality.

1.5.14 Do Not Create an ETH Fabric for USB Interfaces

Oracle Enterprise Manager Ops Center does not support USB network creation or representation.

As the USB network creation or representation is not supported, creating fabric for a USB data link on a server is not useful.

1.5.15 Non-Global Zone Persistent Configuration Should Not Be Updated if Dynamic Reconfiguration Fails

When performing dynamic zone management operations, current Oracle Enterprise Manager Ops Center behavior is to update the persisted zone configuration prior to the live zone configuration.

If a job fails to update the live zone configuration, the persisted zone configuration may be updated successfully. To synchronize the live zone configuration with the configuration displayed in Oracle Enterprise Manager Ops Center, restart the non-global zone.

1.5.16 Database Migration From Local 12c to Remote 12c with Container Times Out

Migrating local 12c database to remote 12c database with container times out after several hours as it was unable to lock an object during migration.

Ops Center migration from local database to remote database fails with the following error:

ORA-04021: timeout occurred while waiting to lock object

The ORA-04021 timeout error is caused by another session locking or using a package that you are attempting to execute. During migration, the Enterprise Controller drops and recreates the remote DB procedure copy_oc_schema, if it exists from previous migration attempt. If the previous migration was aborted, this may cause a lock hold in the remote schema procedure copy_oc_schema and the current migration waits until the lock is released or timeout occurs.

Workaround

Apply the solution found in MOS Note 2079362.1: Ops Center migration from local database to remote database fails with error ORA-04021: timeout occurred while waiting to lock object.

1.5.17 Update Management and Delete Asset Should Clear the Session

When performing Update Management and Delete Assets actions on a Proxy Controller, the session times out due to inactivity.

The session times out due to the presence of an active old session. This leads to performance issues.

Workaround

Restart the Proxy Controller.

1.5.18 Reverting to Earlier Version Without Reverting Customer-Managed Database Can Cause Issues

Once you have upgraded to the latest version of OpsCenter, if you are using a customer-managed database, you cannot revert to an earlier version by reverting only the Enterprise Controller system to a saved boot environment. You must revert both the Enterprise Controller system and database system to the same pre-upgrade state.

If you revert to an earlier version, all monitoring data gathered since the reversion point is lost.

1.5.19 LUNs for Zone Storage Does Not Get Locked

If a deployment plan is initiated to create multiple zones in one job, using LUNs from a SAN library as zone storage, the deployment wizard uses a LUN for each zone from the SAN library as storage. The LUNs used are not locked until the matching zone gets created.

If a second deployment plan is initiated using the same SAN library, when the first deployment plan is still in progress, the deployment wizard will use the same available LUNs for zone storage which were already used earlier.

Therefore, both jobs will partially fail with only a few zones created. Do not submit these kind of jobs in parallel.

1.5.20 Users and Groups Created by Ops Center Installation Not Specified

The accounts that are created as part of the Oracle Enterprise Manager Ops Center 12c installation are not listed in the documentation.

The following users and groups are created in Oracle Enterprise Manager Ops Center 12c:

  • Enterprise Controller Users - svctag, allstart, uce-sds, xvm, webservd

  • Enterprise Controller Groups - jet, uce-sds, webservd

  • Proxy Controller Users - svctag, allstart, uce-sds

  • Proxy Controller Groups - jet, uce-sds

The svctag and xvm users are expected to have the default UIDs of 95 and 60, respectively. The UIDs for the other users depend on the OS settings and are not expected to be a specific value.

1.5.21 Zones Using iSCSI Storage Are at Risk to Get Broken During Global Zone Reboot

On an Oracle Solaris platform, when you reboot the global zone hosting the kernel zone based on iSCSI storage, an error is displayed.

Workaround

  1. After the global zone has booted successfully, manually attach zones on the global zone again by using the following command:

    zoneadm -z <ngz> attach
    
  2. The zones disappear from the user interface. Hence refresh the Global zone to add them again.

1.5.22 Create Kernel Zone Job Fails When Overriding CPU Model

During the creation of zone with a shared CPU, if you choose the allow override any profile values option, the shared CPU value is changed to dedicated CPU value and hence the creation of kernel zone fails.

1.5.23 Ability to Create Kernel Zone Using Infiniband

The discovery of Kernel Zones using Infiniband is supported, but creating a Kernel Zone with Infiniband from Oracle Enterprise Manager Ops Center user interface is not supported.

1.5.24 Dynamic Editing of a S10 Branded Zone Should Not Be Allowed in the User Interface

The dynamic network edit of a zone is based on the Global Zone Solaris version. For Oracle S10 branded zones, the dynamic network edition should not be allowed.

Workaround

You can start a job when the zone is running. This job will not dynamically edit the zone configuration. Restart the S10 branded zone at the end of the job for the zone configuration to be updated.

1.5.25 Improve Error Message on Failed Hardware Discovery

During the discovery of Oracle SPARC S7-2 and Oracle SPARC S7-2L systems, the discovery fails with an error message due to the servicestate property which is disabled under the /SP/services/ipmi/ file on the service processor.

The following error message is displayed:

Discovery driver reported: Fail to retrieve the ethernet ports IPMI for host <IP> exit with errors: 
Error: Unable to establish IPMI v2 / RMCP+ session Get Channel Info command failed

This error message does not describe the cause for the failure and the fix for the issue.

Workaround

Log in to the service processor and enable the servicestate property value by using the following commands:

cd /SP/services/ipmi/ 
set servicestate=enabled 
show /SP/services/ipmi/

/SP/services/ipmi 
 Targets:

 Properties:
                servicestate = enabled
                v1_5_sessions = disabled

1.5.26 Kernel Zone Creation - IP Missing if Many Connections on the Same Global Zone Lower-link

When creating a Kernel Zone from Oracle Enterprise Manager Ops Center user interface, if the user requires 2 interfaces, in order to connect 2 networks with an IP address in both interface in the Kernel Zone and if both networks are connected on the same interface in the Global Zone that will host the Kernel Zone, then both anets are properly created but one of the IPs may be missing in the Kernel Zone OS (the datalink is not configured in the OS).

Workaround

Create the Kernel Zone without a maximum of 1 network interface from the same Global Zone lower-link. Once the Kernel Zone is created, additional network connections may be added from the same Global Zone lower-link using the Attach Network action.

1.5.27 Kernel Zone Dynamic Add Network Does Not Set Network IP in Kernel Zone Operating System

When adding a new network interface with an IP address to a running kernel zone from its server asset, the new anet will be added only in the persistent zone configuration but does not add in the live zone configuration.

As a consequence, the new IP address will not be set in the kernel zone operating system.

Workaround

  1. Halt the Kernel zone, Add the network while the Kernel zone is halted, Boot the Kernel zone and set the new IP address in the Kernel zone OS.
  2. Add the network to the running Kernel zone (dynamic configuration will not be updated), reboot the Kernel zone (this will update the dynamic configuration) and set the new IP address in the Kernel zone OS.
  3. On the global zone that hosts the Kernel zone, add the following new property in /opt/sun/n1gc/lib/XVM_SERVER.properties file.

    zone.kz.force.dynamic.reconfig=true

  4. Restart the global zone VC agent:
    /opt/SUNWxvmoc/bin/agentadm stop  
    /opt/SUNWxvmoc/bin/agentadm start 
    

    Perform add network operation dynamically.

1.5.28 Could Not Perform Operation After Automatic Recovery on LDOM Guests

After an automatic move of the logical domain guest from a source server to a destination server, launching some operations on the recovered guest (reboot, delete, migrate, connect network, move metadata, automatic recovery) fails.

Error message displayed is as follows:

Could not perform operation because host ID collision has been found. You should apply the repair procedure described in Ops Center documentation and try again (61021).

Workaround

Some operations fail when the host ID collision detection finds two images for the recovered guest, on source server and destination server. This issue will be automatically fixed when the source server is reachable again, since the image gets cleaned once the server is available.

You can also set collision.actions.check.hostid to false. Navigate to Administration > Configuration > Virtualization panel to set the configuration variable. This will prevent checking host ID collisions before launching these operations. After the source server is available again, set the variable back to true.

1.5.29 Mismatch in Vlan Tagging Information Shown on Network Dashboard and Details Tab

There is mismatch between vlan tagging information shown on network Dashboard and Network Details tab.

For any selected network from Network accordion, Network Dashboard always shows vlan Tagging attribute value as enabled, whereas Network Details tab shows vlan Tagging attribute value as enabled only if vlan id is greater than VLAN_MIN value which is currently set to 2, otherwise it shows it as disabled for untagged networks.

Workaround

Follow the Network Details tab to get the correct vlan tagging information about the network.

1.5.30 Proxy Controller OFFLINE after Rollback and Subsequent Upgrade

When upgrade fails at a customer site for some reason, rolling back to previous version, the subsequent successful upgrade leaves Proxy Controller in OFFLINE state, until it is manually restarted by the customer. Without restart, the Proxy Controller remains unreachable (OFFLINE), affecting the upgrade of associated Agent Controllers.

Workaround

If Proxy Controller remains unreachable (OFFLINE) after upgrade, use proxyadm stop -w && proxyadm start -w to restart it.

1.5.31 Host ids for LDOMs should be 8 Characters Long

oocli is not able to process correctly with shorter host ids. For example, reporting host ids collisions.

Workaround

Host ids must be 8 characters long with pre-pending zeros if necessary.

/opt/SUNWoccli/bin/oc -e 'connect -u root -p xxxxxxx ; collisions ;  check -t  hostid -v 0x00999996' 

1.5.32 Remote Proxy Controller Unable to Configure Non-OC Publishers

If a user adds non OpsCenter content to the Solaris 11 Library using the EC BUI, when installing / configuring a remote Proxy Controller on Solaris 11, it is not able to configure the non-OC IPS publishers.

Workaround

On the remote proxy controller, execute the following commands:
# rm -rf /var/opt/sun/xvm/uce/opt/server/proxy/* 
#/var/opt/sun/xvm/utils/install_ips_ac.sh -P 127.0.0.1 

1.5.33 Without Upgrading to the Latest Version, Ops Center Agent and Proxy Controllers will fail with Java Version >= 1.8.0_181

JDK 1.8.0_181 removes bundled Java derby. This breaks OpsCenter agent controllers and proxy controllers that rely on Java derby for monitored data. With Solaris 11.3 SRU 34 and Solaris 11.4, Java version greater than or equal to 1.8.0_181 is delivered

Workaround

Do not upgrade Java/Solaris to the versions mentioned above without upgrading Ops Center agent and proxy controllers to the latest version (12.4.0).

1.5.34 Latest fallback Boot Images Missing for T7

The latest boot image in the EC library for T7 is sol-11_3_17_5_0-fallback_boot-sparc.pkg, but it should be sol-11_3_31_6_0-fallback_boot-sparc.pkg.

Workaround

Contact Oracle support.

1.5.35 Occasionally Scheduled Jobs Go Missing

It is occasionally observed that on certain environments scheduled jobs go missing.

1.5.36 Create Oracle Solaris 11.4 Zone Fails when any Language other than English is Specified in the Wizard

When a language other than "English" is specified in the Create Oracle Solaris Zone wizard for Solaris 11.4 zones, the job fails at the Start Zone task. The SMF service svc:/milestone/config:default on the non-global zone goes into maintenance. This is also observed outside of Ops Center when the zone is created through through command line as well.

1.5.37 Actions on ILOM fail if Port 162 is Occupied on the Proxy Controller

Actions like discover, delete, and reset SP on the ILOM will fail if port 162 is occupied by any process other than the java process belonging to the Ops Center Proxy Controller.

Workaround

Port 162 must be available for the Proxy Controller. No other application on the Proxy Controller should use this port.

1.5.38 MOS SR Create Operation Fails

On the Ops Center BUI, under Message Center, choosing the action Open SR on an incident fails with the error pop-up, Error creating SR.

1.5.39 User uce-sds Lacks sys Group Membership after Upgrade to 11.4

Workaround

After upgrading Solaris to Solaris 11.4 on the Enterprise Controller and Proxy Controller, the user uce-sds loses the sys group membership. This is an underlying Solaris bug. This impacts the UCE SMF services (uce-scheduler and uce-server). This affects the corresponding operations like patching Solaris 10 servers, downloading OS image, and so on.

  1. Manually add the user uce-sds to the /etc/group directory of the sys group.
    Example: (After adding uce-sds)
    # grep sys: /etc/group 
    sys::3:bin,adm,root,uce-sds
    
  2. Restart the Enterprise Controller services.

1.5.40 DST Change Impacts Scheduled Jobs

Scheduled jobs run an hour prior on a Daylight Savings Time end date.

1.5.41 Two Expired Root CA Certificates

Glassfish (version 3.1.2.19) is bundled with OpsCenter 12.4.0 and delivered in the following:
/usr/lib/cacao/lib/tools/template/domain/cacerts.jks

Note:

Glassfish has several entries that have either expired or set to expire in 2019. These will be addressed by OCDoctor.

1.5.42 Rebooting Solaris 11.4 with NFS Mounts in /etc/vfstab kills smf:nfs/client and Prevents HA EC and Agent Startup

The existing NFS Library mountpoints in /etc/vfstab prevent nfs/client SMF service starts after an OS reboot. This prevents Agent Controller startup on Global zone with associated NAS libraries or Enterprise Controller startup on an EC-HA node configured with a NAS mounted voting disk. This affects EC or an Agent Controller only when they are running on a Solaris 11.4 Operating System. Solaris 11.2 and 11.3 are not affected.

Wordaround

  1. Comment out the NFS mount entries in /etc/vfstab (Using # at the beginning of the line).
  2. Run the following command:
    $ svcadm clear nfs/client
    
  3. Uncomment the entries from /etc/vfstab (Remove # from the beginning of the entries commented in the first step).
  4. Start the Agent controller or Enterprise Controller services.

1.5.43 With EC and PC Running on Non-global Zones on a Solaris 11.3 Global Zone, Upgrading that Global Zone to Solaris 11.4 Fails

As an example, a Solaris 11.3 Global zone has 2 non-global zones – one running the Enterprise Controller and the other running a remote Proxy Controller. The following upgrade path is performed on these non-global zones: OpsCenter 12.3.3 JIDR1 (On Solaris 11.3) -> OpsCenter 12.4.0 (On Solaris 11.3). There is no agent running on the global zone (not supported). Manually upgrading the global zone at this point to Solaris 11.4 fails with:
Planning linked: 0/2 done; 1 working: zone:eczone 
Linked progress: \pkg: update failed (linked image exception(s)): 

A 'sync-linked' operation failed for child 'zone:eczone' with an unexpected 
return value of 1 and generated the following output: 
pkg: 0/1 catalogs successfully updated: 
   
Framework error: code: 6 reason: Could not resolve host: 
oracle-oem-oc-mgmt-eczone 
URL: 'https://oracle-oem-oc-mgmt-eczone:8002/IPS/versions/0/
The global zone publisher looks like the following:
(GZ)# pkg publisher 
PUBLISHER                   TYPE     STATUS P LOCATION 
solaris                     origin   online F http://ipsrepo.oracle.com/solaris11/support/ 
GZ#
The non-global zone EC publisher looks like the following:
(NGZ) # pkg publisher
PUBLISHER            TYPE     STATUS     P       LOCATION
solaris  (syspub)    origin   online    T       <system-repository>
solaris  (syspub)    origin   online    F       https://oracle-oem-oc-mgmt-eczone:8002/IPS/
cacao                origin   online    F       https://oracle-oem-oc-mgmt-eczone:8002/IPS/
mp-re   (non-sticky) origin   online    F       https://oracle-oem-oc-mgmt-eczone:8002/IPS/
opscenter            origin   online    F       https://oracle-oem-oc-mgmt-eczone:8002/IPS/

Workaround

  1. Ascertain that the OpsCenter version of the EC/PCs on all the underlying non-global zones is the same.
  2. Unset all publishers pointing to OpsCenter’s Solaris 11 IPS repository on the Global zone (if any) and all the non-global zones. Disabling does not work.

    Example on the NGZ:

    (NGZ)# pkg set-publisher -G https://oracle-oem-oc-mgmt-eczone:8002/IPS/ solaris
    (NGZ)# pkg set-publisher -G https://oracle-oem-oc-mgmt-eczone:8002/IPS/ cacao
    (NGZ)# pkg set-publisher -G https://oracle-oem-oc-mgmt-eczone:8002/IPS/ mp-re
    (NGZ)# pkg set-publisher -G https://oracle-oem-oc-mgmt-eczone:8002/IPS/ opscenter
    

    Note:

    There might be other publishers pointing to OpsCenter’s IPS repository. For example: ha-cluster and so on.
  3. The publisher opscenter needs to be added to the global zone, so that the non-global zones can find the required OpsCenter packages during the OS upgrade. To do this, copy the file oc-archive.p5p from the OpsCenter EC/PC installation on one of the non-global zones.
    For example:
    (GZ)# cp <EC-zonepath>/root/var/opt/sun/xvm/images/IPS/oc-archive.p5p /var/tmp/
    
  4. Configure the OpsCenter publisher on the global zone:
    For example:
    (GZ)# pkg set-publisher -G ‘*’ -g /var/tmp/oc-archive.p5p opscenter
    
  5. Try the OS upgrade to Solaris 11.4. If this works – proceed to Step 6. If not, then it might be because of a Solaris bug. The symptom of this bug is the following error during the upgrade:
    pkg sync-linked: Package 'network/ipfilter' must be uninstalled or upgraded 
    if the requested operation is to be performed. 
      Reject:  pkg://solaris/network/ipfilter@0.5.11-0.175.3.34.0.3.0 
      Reason:  Package network/ipfilter is not installed in global zone. 
      Reason:  Excluded by synced parent incorporation 
    'consolidation/osnet/osnet-incorporation'
    

    The above solaris bug is related to the two packages network/ipfilter and system/locale/setterm, if they are installed on any of the non-global zones, but not on the global zone. The workaround is to install these packages on the global zones as well. The OS upgrade to 11.4 should now be successful. The above error may look different if the package network/ipfilter is already installed on the global zone or not installed on any of the non-global zones.

  6. Restart the global zone to Solaris 11.4.
  7. Fix the OpsCenter publishers on the global zone and the non-global zones:
    1. Unset the temporary OpsCenter publisher on the global zone:
      (GZ)# pkg unset-publisher opscenter
      
    2. For non-global zones that had OpsCenter publishers configured, login to the non-global zone running the EC and run the following command first, followed by each non-global zone running PCs.
      (GZ)# /var/opt/sun/xvm/utils/install_ips_ac.sh -P 127.0.0.1
      

      This will reconfigure the OpsCenter publishers on the non-global zones.

1.5.44 Installation on a Freshly Installed Solaris 11.4 solaris-small-server Group Fails

On a freshly installed Solaris 11.4 small-server group, installation of OpsCenter 12.4.0 might fail with the following error:
Executing job: jobs/04installPrereqs.pl --install 
Can't exec "pkginfo": No such file or directory at /var/tmp/xvmoc_full_bundle/./SunOS_sparc/Product/installer/Modules/util.pm line 1123. 
Use of uninitialized value in split at /var/tmp/xvmoc_full_bundle/./SunOS_sparc/Product/installer/Modules/util.pm line 1123. 
scripts/should_install_pkg[280]: pkginfo: not found [No such file or directory] 
Error in running pkginfo against
/var/tmp/xvmoc_full_bundle/./SunOS_sparc/Product/components/packages/SFWrpmlib.pkg. 
Error determining whether  /var/tmp/xvmoc_full_bundle/./SunOS_sparc/Product/components/packages/SFWrpmlib.pkg is newer than installed.

This is because the svr4 package is no longer automatically installed with the solaris-small-server group installation in Solaris 11.4.

Workaround

Install the svr4 package before installing OpsCenter:
$ pkg install package/svr4

1.5.45 Cannot Rediscover a Greenfield Non-global Zone without Metadata. Disks are not Modeled

For zones created through OpsCenter, rediscovering the global zone without previous metadata, would lose the disk modeling of these greenfield non-global zones.

Workaround

  1. Before rediscovering a global zone with greenfield non global zones, ensure the existing metadata is backed up. The new agent will re-read the good metadata.
  2. (This will require zone downtime) Change the mountpoint of the zone and delete the image-uuid attribute, to discover the zone as if manually created:
    • Halt the zone:
      $ zoneadm -z <zoneName> halt
      
    • Get the zone path:
      $ zonecfg -z <zoneName> info zonepath
      
    • Get the corresponding zpool:
      $ zfs list -o name <zonepath>
      
    • Change the mountpoint of the zone datasets:
      $ zfs set -r mountpoint=<newMountpoint> <zpoolName>
      
    • Change the zonepath to this new mountpoint:
      $ zonecfg -z <zoneName> set zonepath=<newMountpoint>
      
    • Delete the image-uuid attribute:
      $ zonecfg -z <zoneName> remove attr name=image-uuid
      
    • Reboot the zone:
      $ zoneadm -z <zoneName> boot
      
  3. If both step 1 and step 2 are not possible, file a ticket with Oracle support.

Oracle Enterprise Manager Ops Center Release Notes, 12c Release 4 (12.4.0.0.0)

F17670-01

Copyright © 2007, 2019, Oracle and/or its affiliates. All rights reserved.

This software and related documentation are provided under a license agreement containing restrictions on use and disclosure and are protected by intellectual property laws. Except as expressly permitted in your license agreement or allowed by law, you may not use, copy, reproduce, translate, broadcast, modify, license, transmit, distribute, exhibit, perform, publish, or display any part, in any form, or by any means. Reverse engineering, disassembly, or decompilation of this software, unless required by law for interoperability, is prohibited.

The information contained herein is subject to change without notice and is not warranted to be error-free. If you find any errors, please report them to us in writing.

If this is software or related documentation that is delivered to the U.S. Government or anyone licensing it on behalf of the U.S. Government, then the following notice is applicable:

U.S. GOVERNMENT END USERS: Oracle programs, including any operating system, integrated software, any programs installed on the hardware, and/or documentation, delivered to U.S. Government end users are "commercial computer software" pursuant to the applicable Federal Acquisition Regulation and agency-specific supplemental regulations. As such, use, duplication, disclosure, modification, and adaptation of the programs, including any operating system, integrated software, any programs installed on the hardware, and/or documentation, shall be subject to license terms and license restrictions applicable to the programs. No other rights are granted to the U.S. Government.

This software or hardware is developed for general use in a variety of information management applications. It is not developed or intended for use in any inherently dangerous applications, including applications that may create a risk of personal injury. If you use this software or hardware in dangerous applications, then you shall be responsible to take all appropriate fail-safe, backup, redundancy, and other measures to ensure its safe use. Oracle Corporation and its affiliates disclaim any liability for any damages caused by use of this software or hardware in dangerous applications.

Oracle and Java are registered trademarks of Oracle and/or its affiliates. Other names may be trademarks of their respective owners.

Intel and Intel Xeon are trademarks or registered trademarks of Intel Corporation. All SPARC trademarks are used under license and are trademarks or registered trademarks of SPARC International, Inc. AMD, Opteron, the AMD logo, and the AMD Opteron logo are trademarks or registered trademarks of Advanced Micro Devices. UNIX is a registered trademark of The Open Group.

This software or hardware and documentation may provide access to or information about content, products, and services from third parties. Oracle Corporation and its affiliates are not responsible for and expressly disclaim all warranties of any kind with respect to third-party content, products, and services unless otherwise set forth in an applicable agreement between you and Oracle. Oracle Corporation and its affiliates will not be responsible for any loss, costs, or damages incurred due to your access to or use of third-party content, products, or services, except as set forth in an applicable agreement between you and Oracle.