Fujitsu M10 Product Notes page 126

Table 3-12 Problems resolved in XCP 2210 (continued)
SPARC M10-
RTI No. 1 4 4S
RTIF2- x x x
140212-025
RTIF2- x x x
140221-002
RTIF2- x x x
140303-002
Fujitsu M10/SPARC M10 Systems Product Notes for XCP Version 2240 ･ March 2015
112
Description
In case of SPARC M10-4/M10-4S, if all
I/O cease to be used on the CPU memory
unit lower (CMUL) of the chassis which
construct the physical partition (PPAR),
or in case of SPARC M10-1, if all I/O
cease to be used on the motherboard unit
(MBU), if the PPAR is powered on
without replacing the failed components,
the starting of hypervisor is aborted.
Or, after Oracle Solaris is started, the
state of Oracle VM Server for SPARC
becomes "suspended". If all I/O cease to
be used, any of the following must have
occurred.
- I/O power supply has failed.
- All PCI Express root complex have
failed.
- All PCI switches have failed.
The "LDC Protocol info from PPAR
(PPAR ID X : Domain Service Data Send
Failed)" message is registered while
logical domains are running and the
"WARNING: ds@x: ds_handle_recv:
invalid message length, received xxx
bytes, expected xxx" message is
displayed on the control domain console.
In some very rare cases, XSCF does not
start after performing AC OFF/ON.
Workaround
There is no effective workaround.
[How to Restore]
In case of SPARC M10-4S/M10-4, replace
the CMUL on which the I/O has failed.
In case of SPARC M10-1, replace the
MBU on which the I/O has failed.
There is no effective workaround.
Reset XSCF with the "rebootxscf -a"
command. After that, execute the "svcs"
command on the control domain to
restart the "picl", "fmd" and the "ldmd"
services.
There are no effective workaround.
Contact the maintenance staff if this
symptom occurs.