Problem Statement:

No errors reported after a Fireplane Interconnect injected fault.  Solaris continues to run without any message in /var/adm/messages and no 'service' LEDs are lit.

The following error messages on the domain shell may be posted:

Oct 31 10:50:45 sc-fortress Domain-A.SC: Domain A has a SYSTEM ERROR

Oct 31 10:50:45 sc-fortress Domain-A.SC: /N0/C4 encountered the first error

Oct 31 10:50:45 sc-fortress Domain-A.SC: ArAsic reported first error on /N0/C4

Oct 31 10:50:45 sc-fortress Domain-A.SC:

/partition0/domain0/C4/ar0:

>>> L2CheckError[0x6150] : 0x00808080

AccIncSyncErr [24:21] : 0x4 accumulated incoming mismatch

FE [15:15] : 0x1

INCSyncErr [08:05] : 0x4 Ports [9:6] incoming mismatched against

internal expected incoming

Oct 31 10:50:45 sc-fortress Domain-A.SC: This domain is still running because

error pause is not enabled for this domain

Cause:

Hardware fault on the L2 repeater board (Safari Bus).

The cause for concern lies in the fact that a fault was injected, reported, and remained on the system for the duration of time that SunVTS was run, with no additional errors encountered.

Assumptions:

None

Solution:

The error would have immediately paused the domain if enable-error-pause had been set to true.

The only recovery from this situation is to "keyswitch off/keyswitch on" the domain.

Work around:

None

Recommendations (Best Practices):

enable-error-pause should be set to true.

See Also:

bugid 4363706

Keywords:
SCApps : Safari Bus : FRU : LED