Skip to main content

Problems Resolved in XCP 2340


Problems Resolved in XCP 2340
The following table lists the problems resolved in XCP 2340.
Table 3-19  Problems Resolved in XCP 2340
RTI No. RTIF2-201215-002
Model SPARC M10-4S
Description If the threshold of the system caps for power consumption is exceeded during dynamic reconfiguration of a physical partition (PPAR DR), Oracle Solaris may inadvertently be shut down.
Workaround Before performing PPAR DR, execute the setpowercapping command to disable the power capping function.
RTI No. RTIF2-201117-003
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description While the OS console is in use, the execution of the console command to connect to the OS console may cause an unexpected message to appear before the login prompt.
Workaround There is no effective workaround.
RTI No. RTIF2-201116-001
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description In a system with a connected PCI expansion unit, an XSCF reboot may cause the "SCF process down detected" error in some very rare cases.
Workaround There is no effective workaround.
RTI No. RTIF2-201116-005
Model SPARC M10-4S
Description In a system with a building block configuration, "XB-XB Interface Correctable Error" may inadvertently be registered in the error log in some very rare cases.
Workaround There is no effective workaround.
RTI No. RTIF2-201116-006
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description On a system with XCP 2330 and XCP 2332 applied, no connection over HTTPS can be made to XSCF Web when elliptic curve cryptography is used.
Workaround There is no effective workaround.
RTI No. RTIF2-201116-010
Model SPARC M10-1
Description If a failure occurs on the I/O board in a PCI expansion unit, "UNSPECIFIED" instead of the faulty component may be registered for the suspected location in the error log.
[Example]
XSCF> showlogs error -rvM
Date: Mar 21 13:15:30 JST 2017
    Code: 20000000-00c7010000ff0000ff-030000020000000000000000
    Status: Notice                 Occurred: Mar 21 13:15:30.364 JST 2017
    FRU: UNSPECIFIED
    Msg: Boot process failed
Workaround There is no effective workaround.
RTI No. RTIF2-201116-011
Model SPARC M10-4
Description If a PCIe card failure occurs in the PCI expansion unit connected to PCI slot #9 in the SPARC M10-4, the registered suspected location in the error log does not show the number of the PCI expansion unit.
[Example]
 FRU: /BB#1/PCI#9/PCIBOX#/PCI#1
Workaround There is no effective workaround.
RTI No. RTIF2-201110-004
Model SPARC M10-4S
Description In a system with a building block configuration, the "XSCF hang-up is detected" error may be detected in some rare cases.
Workaround There is no effective workaround.
RTI No. RTIF2-201109-009
Model SPARC M10-4S
Description Hypervisor abort might occur if live migration of a guest domain and PPAR DR are performed sequentially.
Workaround There is no effective workaround.
RTI No. RTIF2-171013-001
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description Security fixes.
(CVE-2016-6304, CVE-2016-2182, CVE-2016-2183, CVE-2016-6302, CVE-2016-6306,
CVE-2016-6515, CVE-2017-3731, CVE-2016-7431, CVE-2016-7429, CVE-2016-7433)
For details, see the Oracle website relating to the following:
Critical Patch Updates
Workaround Update the XCP firmware to XCP 2340 or later.
RTI No. RTIF2-170421-001
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description Booting from an external DVD drive may fail.
[Example of output message]
{0} ok boot /pci@8000/pci@4/pci@0/pci@2/pci@0/usb@4,1/cdrom@2/disk
 Boot device: /pci@8000/pci@4/pci@0/pci@2/pci@0/usb@4,1/cdrom@2/disk File and args:
  Can't open boot device
Workaround
1. Stop OpenBoot PROM at the ok prompt.

2. Perform the following steps to change the OpenBoot PROM environment variable nvramrc.

a. Execute the nvedit command, and write nvramrc as shown below.

    {0} ok nvedit
    0: : patch-zero 0 to my-adr0 0 ;
    1: ' patch-zero ' 0 ' new-instance d# 28 + token@ (patch
b. Press the [Ctrl] + [C] keys to exit nvedit and return to the ok prompt.

c. Execute nvstore to save the edited contents.

    {0} ok nvstore
d. Check whether nvramrc was written correctly.

    {0} ok printenv nvramrc
    nvramrc = : patch-zero 0 to my-adr0 0 ;
                ' patch-zero ' 0 ' new-instance d# 28 + token@ (patch
    {0} ok
3. Set the OpenBoot PROM environment variable use-nvramrc? to true.

    {0} ok setenv use-nvramrc? true
    use-nvramrc? = true
4. Execute the reset-all command to restart OpenBoot PROM.

    {0} ok reset-all
 
After executing the reset-all command and the system stops at the ok prompt, boot from a DVD drive.
Example: For booting from a DVD drive connected to a front USB port
 {0} ok boot /pci@8000/pci@4/pci@0/pci@2/pci@0/usb@4,1/cdrom@2/disk
 
To restore the original settings from the above settings in [Workaround], perform the following steps.
1. Stop OpenBoot PROM at the ok prompt.

2. Delete the set values from the OpenBoot PROM environment variable nvramrc.

a. Execute nvedit to delete ": patch-zero 0 to my-adr0 0 ;" and
"' patch-zero ' 0 ' new-instance d# 28 + token@ (patch" from nvramrc.
You can delete lines one by one by pressing the [Ctrl] + [K] keys twice at the beginning of a line.

    {0} ok nvedit
    0:
b. Press the [Ctrl] + [C] keys to exit nvedit and return to the ok prompt.

c. Execute nvstore to save the edited contents.

  {0} ok nvstore

d. Check whether the nvramrc setting was deleted correctly.

    {0} ok printenv nvramrc
    nvramrc =
    {0} ok
3. Set the OpenBoot PROM environment variable use-nvramrc? to false.
If it was set to true before this workaround was performed, this step is unnecessary.

    {0} ok setenv use-nvramrc? false
    use-nvramrc? = false
    {0} ok
4. Execute the reset-all command to restart OpenBoot PROM.

    {0} ok reset-all
RTI No. RTIF2-170407-001
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description The ttydm process may fail and the following logs may be registered if you reboot the XSCF by executing the rebootxscf(8), switchscf(8), or flashupdate(8) command while console messages are continuously output from the control domain.
[Example]
XSCF> showlogs error -vr
Date: Mar 02 22:23:00 JST 2015
    Code: 20000000-00fcff000e020000ff-010400010000000000000000
    Status: Notice                 Occurred: Mar 02 22:22:54.396 JST 2015
    FRU: /FIRMWARE,/MBU
    Msg: SCF process down detected
    Diagnostic Code:
        00000000 00000000 0000
        00000000 00000000 0000
        00000000 00000000 0000
        74747964 6d2e3237 36342e62 7a320000
        00000000 00000000 0000
You can check the ttydm process by confirming that the first four bytes on the fourth line of "Diagnostic Code" have the value "747479646d".
Workaround There is no effective workaround.
Logs are collected. The XSCF is rebooted.
RTI No. RTIF2-170406-001
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description From XCP 2230 onwards, a process may produce a core dump if Dynamic Resource Management (DRM) of Oracle VM Server for SPARC is used or a virtual CPU is deleted by executing the ldm remove-vcpu command.
Workaround There is no effective workaround.
RTI No. RTIF2-170406-002
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description When you select the [XSCF] - [Logs] - [Console Log] menu from XSCF Web and display log information, the following error appears in a pop-up window and the screen returns to the login window.
Description Server Internal error or the page you visit does not exist or has been deleted! Sorry for any inconvenience!
Workaround There is no effective workaround.
Execute the showlogs command in the XSCF shell to check the console log.
RTI No. RTIF2-170405-004
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description If a security key is used for booting Oracle Solaris on the control domain and you save the domain configuration information by using the following procedure, the next boot fails.
1. Execute the set-security-key command or the set-ascii-security-key command at the ok prompt of OpenBoot PROM on the control domain to set a security key.

2. Save the domain configuration information to the XSCF by executing the ldm add-spconfig command of Oracle VM Server for SPARC.

3. Boot Oracle Solaris on the control domain.

This problem does not occur on a guest domain.
[Example] Booting Oracle Solaris fails after a security key is set
SPARC M10-4S, No Keyboard
Copyright (c) 1998, 2016, Oracle and/or its affiliates. All rights reserved.
OpenBoot 4.38.5, 126.0000 GB memory available, Serial #*********.
[ 2.21.0 ]
Ethernet address **:**:**:**:**:**, Host ID: ********.
-- Omitted --

Boot device: /pci@8000/pci@4/pci@0/pci@9/network@0 File and args:
100 Mbps full duplex Link up
100 Mbps full duplex Link up
100 Mbps full duplex Link up
SunOS Release 5.11 Version 11.3 64-bit
Copyright (c) 1983, 2015, Oracle and/or its affiliates. All rights reserved.
NOTICE: Configuring iSCSI boot session...
Hostname: solaris
 
solaris console login: root
Password:
Last login: Wed Jan ** **:**:** 2017 on console
Jan ** **:**:** solaris login: ROOT LOGIN /dev/console^M
Oracle Corporation SunOS 5.11 11.3 September 2015
You have mail.
root@solaris:~# ldm ls-spconfig
factory-default [current]
root@solaris:~# ldm add-spconfig ldm-cfg1
root@solaris:~# ldm ls-spconfig
factory-default
ldm-cfg1 [current]
root@solaris:~# init 6
-- Omitted --

SPARC M10-4S, No Keyboard
Copyright (c) 1998, 2016, Oracle and/or its affiliates. All rights reserved.
OpenBoot 4.38.5, 126.0000 GB memory available, Serial #*********.
[ 2.21.0 ]
Ethernet address **:**:**:**:**:**, Host ID: ********.
 
100 Mbps full duplex Link up
Boot device: net File and args:
100 Mbps full duplex Link up
ERROR: /iscsi-hba: iSCSI login failed
 
ERROR: /packages/obp-tftp: Could not open /iscsi-hba/disk
 
Boot load failed
 
Evaluating:
 
{0} ok
Workaround There is no effective workaround.
[How to restore]
After executing the ldm add-spconfig command from Oracle VM Server for SPARC to store the domain configuration information in the XSCF, set the security key again by executing the set-ascii-security-key command or set-security-key command at the ok prompt of OpenBoot PROM.
RTI No. RTIF2-170405-006
Model SPARC M10-4S
Description In a system with a building block configuration, the following error log may be registered when the input power is turned on (AC ON) or the rebootxscf -a command is used.
 XSCF data synchronization failed
At this time, the physical partition cannot be powered on.
Workaround There is no effective workaround.
Reboot all the XSCFs by using the rebootxscf -a command.
RTI No. RTIF2-170405-011
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description The showinterimpermit command cannot be executed with the fieldeng privilege.
Workaround Execute the command with any of the following user privileges.
platadm, platop, pparadm, pparmgr, pparop
RTI No. RTIF2-170331-001
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description On a SPARC M10 system with SPARC64 X processors, Hypervisor may abort after performing a live migration on a guest domain.
Workaround There is no effective workaround.
RTI No. RTIF2-161215-001
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description If a live migration of a guest domain fails, logical domains on the physical partition as the migration source may panic or Hypervisor may abort.
Workaround There is no effective workaround.
If a live migration fails, reboot the guest domain as soon as possible.
RTI No. RTIF2-161117-001
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description If you boot/reboot logical domains after the elapse of a certain period of time from physical partition (PPAR) power-on, the boot/reboot may fail.
This problem occurs in a boot/reboot from an internal LAN.
This problem happens during the time period described below.
- The first period is the 15-minute period after the elapse of "24 days, 20 hours, 16 minutes, and 23.648 seconds" from physical partition (PPAR) power-on. Each subsequent period is the 15-minute period after the elapse of a multiple of time of "49 days, 17 hours, 02 minutes, and 47.296 seconds."

 (15-minute period on the 25th day and then the 75th day, 125th day, 175th day, and so on at an interval of approximately 50 days)

[Example]
Period of 15 minutes from 24th day, 20 hours, 16 minutes, and 23.648 seconds to (same day and hour) 31 minutes and 23.648 seconds
Period of 15 minutes from 74th day, 13 hours, 19 minutes, and 10.944 seconds to (same day and hour) 34 minutes and 10.944 seconds

[Example of failure] Oracle Solaris is booted from internal LAN in control domain on SPARC M10-4
SPARC M10-4S, No Keyboard
Copyright (c) 1998, 2016, Oracle and/or its affiliates. All rights reserved.
OpenBoot 4.38.5, xx.xxxx GB memory available, Serial #xxxxxxxx.
[ 2.19.0 ]
Ethernet address xx:xx:xx:xx:xx:xx, Host ID: xxxxxxxx.
1000 Mbps full duplex Link up
Boot device: net:dhcp File and args:
1000 Mbps full duplex Link up
Timed out waiting for BOOTP/DHCP reply
Timed out waiting for BOOTP/DHCP reply
Timed out waiting for BOOTP/DHCP reply
Timed out waiting for BOOTP/DHCP reply
Timed out waiting for BOOTP/DHCP reply
Timed out waiting for BOOTP/DHCP reply
:
:
Workaround There is no effective workaround.
[How to restore]
Execute the sendbreak command from the XSCF. After the system stops at the ok prompt and 15 minutes elapse, execute the boot command.
RTI No. RTIF2-150226-002
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description CHECK LED of the old master XSCF blinks during the master/standby switchover of XSCFs.
Workaround There is no effective workaround.
Ignore the blinking CHECK LED of the old master XSCF during the master/standby switchover of XSCFs.
RTI No. RTIF2-141008-001
Model SPARC M10-1, SPARC M10-4, SPARC M10-4S
Description If the setsnmp command tries to set enabled immediately after disabled was set, the "Agent enable failed" message may appear, and the command may fail.
This may cause the following phenomena.
1. If it is left as is, "SCF panic detected" is registered in the error log, and the XSCF is rebooted.

2. The setting is not reflected, so if the showsnmp command is executed, it displays "Enabled" for "Agent Status" and "None" for "Enabled MIB Modules."

3. If the showsnmp command is executed, it displays "Enabled" for "Agent Status" and "SP MIB" for "Enabled MIB Modules." However, the setting has not actually been reflected, so rebooting the XSCF causes "None" to be displayed for "Enabled MIB Modules."

4. The setting has not been reflected, so when the showsnmp command displays "None" for "Enabled MIB Modules," the system using the cluster software detects the following error: "FJSVcluster: Error: DEV: 7240: Cannot connect to the XSCF."

[Example] After the setsnmp command fails, the setting is not reflected.
XSCF> setsnmp disable
XSCF> setsnmp enable
setsnmp: Agent enable failed
XSCF> showsnmp
 
Agent Status: Enabled
Agent Port: 161
System Location: System-Location
System Contact: System-Contact

Status: Enabled
Community String: public
 
Enabled MIB Modules: None
XSCF>
Similarly, this problem also occurs when the setsnmpusm command or setsnmpvacm command is executed to change the setting while the SNMP setting is enabled.
Workaround To change the setting to enabled after it was changed to disabled, wait at least 30 seconds, the minimum interval.
To change the setting with the setsnmpusm command or setsnmpvacm command, temporarily change the SNMP setting to disable, and then wait at least 30 seconds, the minimum interval.
[How to restore]
Temporarily set disable with the setsnmp command, and then wait at least 30 seconds, the minimum interval, to set enable.
Next, execute the showsnmp, setsnmpusm, or showsnmpvacm command to confirm that the setting has been reflected as expected.