oracle数据库rac节点故障,两节点rac，故障时不能驱逐故障节点

有点类似于这个bug

Bug 18971331 : DB NODE#1 IS CRASHED WITH HUGE CONSUMPTION OF CPU RESOURCE TO OCSSD.BIN

Hdr: 18971331 11.2.0.4 PCW 11.2.0.4 CSS PRODID-5 PORTID-226

Abstract: DB NODE#1 IS CRASHED WITH HUGE CONSUMPTION  OF CPU RESOURCE TO OCSSD.BIN

*** 06/12/14 07:19 pm ***

PROBLEM:

--------

The problem of the customer is crashed a DB node#1 of Exadata rack at two

times between 2 months. At this point, we focused on crashing DB node#1 at

05:23 May 25.

teams so that while initializing a problem, %CPU of an ocssd.bin is using

2287.2% which is an analysis result of Linux Team.

DIAGNOSTIC ANALYSIS:

--------------------

<<< ocssd.log of DB Node#1 >>>

...

2014-05-25 05:25:46.049: [    CSSD][764115264]clssgmShutDown: Received

abortive shutdown request from client.

2014-05-25 05:25:46.049: [

CSSD][764115264]###################################

2014-05-25 05:25:46.049: [    CSSD][764115264]clssscExit: CSSD aborting from

thread GMClientListener

2014-05-25 05:25:46.049: [

CSSD][764115264]###################################

2014-05-25 05:25:46.049: [    CSSD][764115264](:CSSSC00012

clssscExit: A

fatal error occurred and the CSS daemon is terminating abnormally

bcexadb01, number 1, has experienced a failure in thread number 15 and is

shutting down

2014-05-25 05:25:46.050: [    CSSD][764115264]clssgmThreadRecovery:recovering

clntlsnr mutex

2014-05-25 05:25:46.050: [    CSSD][764115264]clssnmQueueClientEvent:

Sending Event(6), type 6, incarn 288609881

2014-05-25 05:25:46.050: [    CSSD][764115264]clssnmQueueClientEvent: Node[1]

state = 3, birth = 288609879, unique = 1396727807

2014-05-25 05:25:46.050: [    CSSD][764115264]clssnmQueueClientEvent: Node[2]

state = 3, birth = 288609877, unique = 1393111836

2014-05-25 05:25:46.050: [    CSSD][764115264]clssscExit: Starting CRSD

cleanup

...

<<< ocssd.log of DB Node#2 >>>

...

2014-05-25 05:26:45.950: [    CSSD][3088668992]clssnmWaitOnEviction: Node

kill could not beperformed. Admin or connection validation failed

2014-05-25 05:26:45.950: [    CSSD][3102861632]clssnmvDiskEvict: Kill block

write, file o/192.168.10.7;192.168.10.8/DBFS_DG_CD_02_bcexaceladm02 flags

0x00010004, kill block unique 1396727807, stamp 3554575728/3554575728

2014-05-25 05:26:45.950: [    CSSD][3098130752]clssnmvDiskEvict: Kill block

write, file o/192.168.10.9;192.168.10.10/DBFS_DG_CD_02_bcexaceladm03 flags

0x00010004, kill block unique 1396727807, stamp 3554575728/3554575728

2014-05-25 05:26:45.950: [    CSSD][3107592512]clssnmvDiskEvict: Kill block

write, file o/192.168.10.5;192.168.10.6/DBFS_DG_CD_03_bcexaceladm01 flags

0x00010004, kill block unique 1396727807, stamp 3554575728/3554575728

2014-05-25 05:26:45.950: [    CSSD][3088668992]clssnmWaitOnEvictions: node 1,

undead 1, EXADATA fence handle 13 kill reqest id 0, last DHB (1400963145,

4235069254, 4232085), seedhbimpd TRUE

2014-05-25 05:26:45.950: [    CSSD][3088668992]clssnmrCheckEDFenceStatus:

Node bcexadb01, number 1, EXADATA fence handle 13 has status 1, status

request return code ORA-0

2014-05-25 05:26:45.950: [    CSSD][3088668992]clssnmCheckKillStatus: Node 1,

bcexadb01, down, due to timeout of DHB; last NHB TOD invariant clock

4235069294, TOD invariant clock 4235069254

2014-05-25 05:26:45.950: [    CSSD][3088668992]clssnmWaitOnEviction: node(1)

exceeded graceful shutdown period, IPMI-kill allowed if configured

2014-05-25 05:26:45.950: [    CSSD][3219560160]clssgmQueueGrockEvent:

groupName(DGBCDB0) count(2) master(1) event(2), incarn 4, mbrc 2, to member

1, events 0x0, state 0x0

2014-05-25 05:26:45.950: [    CSSD][3088668992]clssnmWaitOnEviction: Node

kill could not beperformed. Admin or connection validation failed

...

WORKAROUND:

-----------

NONE.

RELATED BUGS:

-------------