============
Envirionment
============
-------------------------------------
OS: LINUX 5.6
DB: 10.2.0.5 RAC
Storage:ASM
Testing ASM Disk: DS1301
-------------------------------------

=====================================================================
Case 1: ASM WITHOUT ScanOrder config (etc/sysconfig/oracleasm-_dev_oracleasm)
=====================================================================

ROOT:
--------------------------------------------------------------------
1 Confirmed ORACLEASM_SCANORDER is blank
--------------------------------------------------------------------
grep ORACLEASM_SCANORDER /etc/sysconfig/oracleasm

---------------------------------------------------------
2 Get ASM disk major and miner number
---------------------------------------------------------
--Check if ASM disk major num is 253, if not,DM scanorder is not configure.

for i in `/etc/init.d/oracleasm listdisks`;do devnum=`/etc/init.d/oracleasm querydisk -d $i` &&echo $devnum;done

Disk "DS1001" is a valid ASM disk on device [8, 97]
Disk "DS1101" is a valid ASM disk on device [8, 113]
Disk "DS1102" is a valid ASM disk on device [253, 66]
Disk "DS1105" is a valid ASM disk on device [253, 59]
Disk "DS1201" is a valid ASM disk on device [8, 129]

Disk "DS1301" is a valid ASM disk on device [8, 145]

Disk "DS1401" is a valid ASM disk on device [8, 161]

-------------------------------------------------------------------------------
3 Get the real sd disk(HBA based) ASMlib choosed --sdj
-------------------------------------------------------------------------------
cat /proc/partitions |grep 145 |grep '\<8\>'
8 145 262108507 sdj1

--confirmed
ls -lth /dev/sd*|grep '8, 145'
brw-r----- 1 root disk 128, 145 Oct 18 21:29 /dev/sdeh1
brw-r----- 1 root disk 68, 145 Oct 18 21:29 /dev/sdbv1
brw-r----- 1 root disk 8, 145 Oct 18 21:29 /dev/sdj1

multipath -ll|grep sdj
\_ 2:0:0:6 sdj 8:144 [active][ready]

multipathd -k
>show paths
2:0:0:6 sdj 8:144 1 [active][ready] XXX....... 41/120

----------------------------------------------------------
4 Stop DB and ASM for data security safe
----------------------------------------------------------
Oracle:

su - oracle
srvctl stop instance -d DBNAME -i Instname
srvctl stop asm -n Nodehostname

---------------------------------------------------------
5 Hot Delete the sdj physical storage link
---------------------------------------------------------

ROOT:

--Confirm link

ll /sys/bus/scsi/drivers/sd/2\:\0:\0:\6/

lrwxrwxrwx 1 root root 0 Oct 18 21:28 block:sdj -> ../../../../../../../../../../../../block/sdj

lrwxrwxrwx 1 root root    0 Oct 18 21:28 bus -> ../../../../../../../../../../../../bus/scsi
--w------- 1 root root 4096 Oct 18 21:28 delete
-r--r--r-- 1 root root 4096 Oct 18 21:28 device_blocked
-rw-r--r-- 1 root root 4096 Oct 18 21:28 dh_state
lrwxrwxrwx 1 root root    0 Oct 18 21:28 driver -> ../../../../../../../../../../../../bus/scsi/drivers/sd
lrwxrwxrwx 1 root root    0 Oct 18 21:28 generic -> ../../../../../../../../../../../../class/scsi_generic/sg10
-r--r--r-- 1 root root 4096 Oct 18 21:28 iocounterbits

--Hot delete link

echo 1 > /sys/bus/scsi/drivers/sd/2\:\0:\0:\6/delete

--Confirmedd if sdj deleted

ls -lth /dev/sdj

----------------------------
6 Try if DB can start
----------------------------
Oracle:

su - oracle
srvctl stop asm -n Nodehostname
sqlplus / as sysdba
startup --failed

-------------------------------------------------------
7 Reboot OS to recovery device link sdj
-------------------------------------------------------
Root:

reboot

ls -lth /dev/sdj --recover
$CRS_HOME/bincrs_stat -t -v

=================================================================
Case 2: ASM WITH ScanOrder config(/etc/sysconfig/oracleasm-_dev_oracleasm)
=================================================================

ROOT:
---------------------------------------------------------------
1 Edit ASM config file to enable dm scan first
---------------------------------------------------------------

vi /etc/sysconfig/oracleasm
Change:
ORACLEASM_SCANORDER=""
to
ORACLEASM_SCANORDER="dm"

--------------------------------
2 Restart ASM service
--------------------------------
Oracle:
--Stop ASM and DB
sqlplus / as sysdba
shutdown immedaite
srvctl stop asm -n Nodehostname

Root:
/etc/init.d/oracleasm restart

----------------------------------------------
3 Check if ASM multipath enable
----------------------------------------------
Root:

for i in `/etc/init.d/oracleasm listdisks`;do devnum=`/etc/init.d/oracleasm querydisk -d $i` &&echo $devnum;done --device mayjor num should be 253

-----------------------------------------------------------
4 Hot Delete the physical storage path sdj
-----------------------------------------------------------

--Confirmed the sdj scsi id
multipath -ll|grep sdj
\_ 2:0:0:6 sdj 8:144 [active][ready]

--Hot delete sdj

echo 1 > /sys/bus/scsi/drivers/sd/2\:\0:\0:\6/delete

----------------------------
5 Try if DB can start
----------------------------
Oracle:

su - oracle
srvctl start asm -n Nodehostname
sqlplus / as sysdba
startup --success

--------------------------------------
6 Repeat 3-5 for other path
--------------------------------------

DB can be up normally for each path deleted with one path available

ITPub博客

Lab-Linux Enable ASM using multipath configuration Testing

最新文章