Fault is not propagated to the SP
Update with additional fmstat information. Seeing this problem with uncorrectable error injections only. Correctable errors appear fine.
Before EI:
# fmstat -m etm -z
NAME VALUE DESCRIPTION
etm_rd_body_response 1 ETM response msg bodies rcvd from xport
etm_rd_hdr_response 1 ETM response msg headers rcvd from xport
etm_resp_q_max_len 512 max enqueable response msgs to xport
etm_wr_body_control 1 ETM control msg bodies sent to xport
etm_wr_hdr_control 1 ETM control msg headers sent to xport
# mtst -b cpuid=44 kdfrfus
# fmstat -m etm -z
NAME VALUE DESCRIPTION
etm_os_nvlist_pack_fail 1 nvlist_pack failures
etm_rd_body_fmaevent 1 ETM fmaevent msg bodies rcvd from xport
etm_rd_body_response 1 ETM response msg bodies rcvd from xport
etm_rd_fmd_bytes 1704 bytes of FMA events rcvd from FMD
etm_rd_fmd_fmaevent 1 FMA events rcvd from FMD
etm_rd_hdr_fmaevent 1 ETM fmaevent msg headers rcvd from xport
etm_rd_hdr_response 1 ETM response msg headers rcvd from xport
etm_rd_max_ev_per_msg 1 max FMA events per ETM msg from xport
etm_rd_xport_bytes 680 bytes of FMA events rcvd from xport
etm_resp_q_max_len 512 max enqueable response msgs to xport
etm_wr_body_control 1 ETM control msg bodies sent to xport
etm_wr_body_response 1 ETM response msg bodies sent to xport
etm_wr_drop_fmaevent 1 dropped FMA events to xport
etm_wr_fmd_bytes 680 bytes of FMA events sent to FMD
etm_wr_fmd_fmaevent 1 FMA events sent to FMD
etm_wr_hdr_control 1 ETM control msg headers sent to xport
etm_wr_hdr_response 1 ETM response msg headers sent to xport
Log File = /net/hah-ha2-nfs.east/export/ds02/d510/fire-sinclair2/data3/ssqa/projects/maramba/logs/fma/cpu-mem/LDOM/float_reg_UE-frfu-kdfrfus.log
Work Around
Not an elegant workaround, but....
# svcadm disable fmd
# rm -rf /var/fm/fmd/ckpt
# rm -rf /var/fm/fmd/rsrc
# svcadm enable fmd
has been reported to return the system to a working state when this bug is hit.