Go to main content

Oracle® Solaris 11.3 Release Notes

Exit Print View

Updated: March 2018
 
 

Issues When Updating to Oracle Solaris 11.3

The following problems might occur when you update to Oracle Solaris 11.3.

Bad Policy Token Error When Updating From Releases Prior to Oracle Solaris 11.1 SRU 9.2 (16773078)

The following error message might be displayed when updating from releases prior to Oracle Solaris 11.1 SRU 9.2:

driver (<driver>) upgrade (addition of policy 'tpd_member=true') failed with
return code 1
command run was: /usr/sbin/update_drv -b /mnt -a -p tpd_member=true <driver>
command output was:

------------------------------------------------------------
Bad policy token: ``tpd_member''.
------------------------------------------------------------

If this error message is displayed, the first boot after the update might take longer.

Workaround: No action is required. Ignore the error message.

L3 VRRP Might Change Existing IP Configuration After an Update from Oracle Solaris 11.1 to Oracle Solaris 11.3 (16720867, 16727299, 16720923)

The introduction of the Layer 3 Virtual Router Redundancy Protocol (L3 VRRP) feature might change the existing IP configuration for a few systems after an update from Oracle Solaris 11.1 to Oracle Solaris 11.3. Certain private IP protocol properties which were previously private are now made public due to the introduction of L3 VRRP.

The following table lists the IP protocol properties that now have public names.

Table 3  IP Protocol Properties With Public Names
IP Protocol Property
Public Name
_arp_publish_count
arp_publish_count
_arp_publish_interval
arp_publish_interval
_ndp_unsolicit_count
ndp_unsolicit_count
_ndp_unsolicit_interval
ndp_unsolicit_interval
_send_redirects
send_redirects

All static IP addresses that are configured over any VRRP virtual network interface cards (VNICs) are also converted to the VRRP type.

For more information, see About the Layer 3 VRRP Feature in Configuring an Oracle Solaris 11.3 System as a Router or a Load Balancer.

Workaround: No action is required.

MPxIO Support for Toshiba Internal SAS Hard Disk Drives Prevents Rollback to Any Older Version of Oracle Solaris (15824482)

Starting with Oracle Solaris 11.2, Solaris multipathing supports Toshiba hard disk drives with the following product IDs:

  • AL13SEB600

  • MBF2600RC

  • MK1001GRZB

  • MK2001GRZB

After updating to Oracle Solaris 11.2 and rebooting to the new boot environment, you cannot roll back to your older version of Oracle Solaris. The following error message is displayed:

root@smem10a:~# beadm activate S11U1SRU20B04-z_stress-2.10-1
Error while accessing "/dev/rdsk/c2t500003942823F352d0s0": 
No such file or directory
Unable to activate S11U1SRU20B04-z_stress-2.10-1.
Error installing boot files.

Workaround: Choose one of the following workarounds:

  • Roll back to the older version of Oracle Solaris by using Open Boot PROM (OBP) or GRand Unified Bootloader (GRUB).

  • Disable MPxIO configuration for the specific HBA port that is connected with the TOSHIBA disk. For more information about disabling MPxIO, see the stmsboot(1M) man page.

ZFS Pool Information Becomes Stale After Running the stmsboot Command With –e Option (15791271)

After running the stmsboot command with the –e option to enable the MPXIO (multipathing) feature on the next boot, the ZFS pool device path information becomes stale for a brief period. As a result, zpool.cache is not updated properly. The following error message might be displayed when updating from some earlier Oracle Solaris releases, earlier than Oracle Solaris 11.2 SRU 7, to Oracle Solaris 11.3 when the pkg update or beadm activate command is executed:

Error while accessing /dev/rdsk/c2d1s0&#8243;: No such file or directory

Workaround: Run the zpool status command and then run the beadm command.

Upgrading S11.2 to S11.3 Results in Oracle Grid Infrastructure 12.1.0.1.0 Startup Hang (21511528)

Oracle RDBMS 12.1.0.1 configured with automatic system global area (SGA) memory management fails to start due to an unexpected mprotect() failure against an optimized shared memory (OSM) segment. The following error message is displayed:

ORA-27122: unable to protect memory

Workaround: Use at least Oracle Grid Infrastructure 12.1.0.2 on Oracle Solaris 11.3.

Ops Center: Communication Between the Agent Controller and the Corresponding Proxy Controller of Ops Center Breaks (21464720)

Oracle Solaris 11.3 includes the latest Java versions, which could cause a break in the communication between the agent controller and its corresponding proxy controller of Ops Center. For more information and a solution for the issue, see Document 2036973.1 on My Oracle Support: “Upgrading Java on an Oracle Enterprise Manager Ops Center Agent Controller to versions Java 6u101, 7u85, 8u51, or later breaks the communication between the Agent Controller and its corresponding Proxy Controller. (Doc ID 2036973.1).”