How to configure ACFS File System in RAC using GI 19c

Posted on by

ACFS i.e Oracle ASM Cluster File System is cluster file system service used for High availability services.
Example – For achieving high availability in goldenate for Oracle RAC, we can use ACFS for goldengate relate files

ENVIRONMENT DETAILS:
ORACLE GRID VERSION – 19.7.0
NODES               – node1 , node2
OS                  – OEL7

First we need to make sure the GI Home is patched with the latest version

Patching Node1 and Node2

These can be downloaded from MOS, we need GI RU and OPatch patch.

#GI 19c Patch Number 31305339
p31305339_190000_Linux-x86-64.zip
#OPatch version
p6880880_190000_Linux-x86-64.zip

Replace the GRID_HOME/Opatch with the latest one and set up the env bars.

export ORACLE_HOME=/u01/app/grid
export PATH=$ORACLE_HOME/bin:$PATH

unzip the patch 31305339 and check that everything is fine

unzip -q p31305339_190000_Linux-x86-64.zip
//the above created /home/oracle/31305339

Node1

Then execute to check that everything is ok

cd $ORACLE_HOME/OPatch
sudo ./opatchauto apply /home/oracle/31305339 -analyze -oh /u01/app/grid

Then

cd $ORACLE_HOME/OPatch
sudo ./opatchauto apply /home/oracle/31305339 -oh /u01/app/grid

Node2

This will fix the problem described here

/u01/app/grid/oui/bin/runInstaller -silent -ignoreSysPrereqs -attachHome ORACLE_HOME="/u01/app/grid" ORACLE_HOME_NAME="OraGI19Home1" CRS=true

Then just patch it

export ORACLE_HOME=/u01/app/grid
sudo ./opatchauto apply /home/oracle/31305339 -oh /u01/app/grid

Additional steps

Then we need to workaround a bug, explanation here

Node1

sudo su -
export ORACLE_HOME=/u01/app/grid
export PATH=$ORACLE_HOME/bin:$PATH
crsctl set resource use 1
#then
crsctl disable has
crsctl stop has
crsctl enable has
crsctl start has

Node2

sudo su -
export ORACLE_HOME=/u01/app/grid
export PATH=$ORACLE_HOME/bin:$PATH
crsctl set resource use 1
#then
crsctl disable has
crsctl stop has
crsctl enable has
crsctl start has

ACFS configuration

Configuring the ACFS filesystem

Node1

sudo mkdir -p /u01/app/fs
sudo chown oracle:oinstall /u01/app/fs

Node2

sudo mkdir -p /u01/app/fs
sudo chown oracle:oinstall /u01/app/fs

Create an ASM DISKGROUP

Set env vars and list disc candidates

asmcmd lsdsk --candidate -p
#on node1
export ORACLE_HOME=/u01/app/grid
export PATH=$ORACLE_HOME/bin:$PATH
export ORACLE_SID=+ASM1

create ~/hatake.xml with the below content

<dg name="HATAKE" redundancy="external">
   <dsk string="/dev/oracleasm/asm-disk4"/>
   <dsk string="/dev/oracleasm/asm-disk5"/>
   <dsk string="/dev/oracleasm/asm-disk6"/>
   <a name="compatible.asm" value="19.0.0.0.0" />
   <a name="compatible.rdbms" value="19.0.0.0.0" />
</dg>

if advm is in use it cal also be added, not needed for this example

<a name="compatible.advm" value="19.0.0.0.0" />

then within asmcmd execute

asmcmd
>mkdg hatake.xml
>lsdg

 Mount the diskgroup from other nodes

srvctl start diskgroup -diskgroup HATAKE -n node1,node2
Create a new volume KAKASHIVOL1 in the diskgroup HATAKE
#asmcmd
volcreate -G HATAKE -s 57G KAKASHIVOL1
volinfo --all
[oracle@node1 ~]$ asmcmd
ASMCMD> volinfo --all
no volumes found
ASMCMD> volcreate -G HATAKE -s 57G KAKASHIVOL1
ASMCMD> volinfo --all
Diskgroup Name: HATAKE
	 Volume Name: KAKASHIVOL1
	 Volume Device: /dev/asm/kakashivol1-412   <---- This is the volume device
	 State: ENABLED
	 Size (MB): 58368
	 Resize Unit (MB): 64
	 Redundancy: UNPROT
	 Stripe Columns: 8
	 Stripe Width (K): 1024
	 Usage: 
	 Mountpath: 
export VOLUME_NAME=`asmcmd volinfo --all | grep "Volume Device:" | awk -F': ' '{print $2}'`
echo $VOLUME_NAME

Check that the mount point was successfully created

crsctl stat res ora.HATAKE.KAKASHIVOL1.advm -t
[oracle@node1 ~]$ crsctl stat res ora.HATAKE.KAKASHIVOL1.advm -t
--------------------------------------------------------------------------------
Name           Target  State        Server                   State details       
--------------------------------------------------------------------------------
Local Resources
--------------------------------------------------------------------------------
ora.HATAKE.KAKASHIVOL1.advm
               ONLINE  ONLINE       node1                    STABLE
               ONLINE  ONLINE       node2                    STABLE
--------------------------------------------------------------------------------
[oracle@node1 ~]$
Create ACFS FILESYSTEM In the KAKASHIVOL1 VOLUME
 sudo /sbin/mkfs -t acfs $VOLUME_NAME
[oracle@node1 ~]$ sudo /sbin/mkfs -t acfs $VOLUME_NAME
mkfs.acfs: version                   = 19.0.0.0.0
mkfs.acfs: on-disk version           = 39.0
mkfs.acfs: volume                    = /dev/asm/kakashivol1-412
mkfs.acfs: volume size               = 61203283968  (  57.00 GB )
mkfs.acfs: Format complete.
[oracle@node1 ~]$

Register the ACFS file system with CRS

sudo $ORACLE_HOME/bin/srvctl add filesystem -d $VOLUME_NAME -m /u01/app/fs -u oracle -fstype ACFS -autostart ALWAYS

Check the resource status

crsctl stat res ora.hatake.kakashivol1.acfs -t
[oracle@node1 ~]$ crsctl stat res ora.hatake.kakashivol1.acfs -t
--------------------------------------------------------------------------------
Name           Target  State        Server                   State details       
--------------------------------------------------------------------------------
Local Resources
--------------------------------------------------------------------------------
ora.hatake.kakashivol1.acfs
               OFFLINE OFFLINE      node1                    STABLE
               OFFLINE OFFLINE      node2                    STABLE
--------------------------------------------------------------------------------
[oracle@node1 ~]$

Start the ACFS file system resource

sudo $ORACLE_HOME/bin/srvctl start filesystem -d $VOLUME_NAME
[oracle@node1 ~]$ crsctl stat res ora.hatake.kakashivol1.acfs -t
--------------------------------------------------------------------------------
Name           Target  State        Server                   State details       
--------------------------------------------------------------------------------
Local Resources
--------------------------------------------------------------------------------
ora.hatake.kakashivol1.acfs
               ONLINE  ONLINE       node1                    mounted on /acfspoc,
                                                             STABLE
               ONLINE  ONLINE       node2                    mounted on /acfspoc,
                                                             STABLE
--------------------------------------------------------------------------------
[oracle@node1 ~]$

Now validate the ACFS mount points

srvctl config filesystem
[oracle@node1 ~]$ srvctl config filesystem
Volume device: /dev/asm/kakashivol1-412
Diskgroup name: hatake
Volume name: kakashivol1
Canonical volume device: /dev/asm/kakashivol1-412
Accelerator volume devices: 
Mountpoint path: /u01/app/fs
Mount point owner: oracle
Mount point group: oinstall
Mount permissions: owner:oracle:rwx,pgrp:oinstall:r-x,other::r-x
Mount users: 
Type: ACFS
Mount options: 
Description: 
ACFS file system is enabled
ACFS file system is individually enabled on nodes: 
ACFS file system is individually disabled on nodes:

verify everything worked fine

#node1
[oracle@node1 ~]$ df -kh /u01/app/fs
Filesystem                Size  Used Avail Use% Mounted on
/dev/asm/kakashivol1-412   57G  193M   57G   1% /acfspoc
#node2
[oracle@node2 OPatch]$ df -kh /u01/app/fs
Filesystem                Size  Used Avail Use% Mounted on
/dev/asm/kakashivol1-412   57G  193M   57G   1% /acfspoc

#node1
[oracle@node1 ~]$ cd /u01/app/fs
[oracle@node1 acfspoc]$ touch fred.is_cool
#node2
[oracle@node2 OPatch]$ ls -l /u01/app/fs
total 64
-rw-r--r-- 1 oracle oinstall     0 Nov  9 18:33 fred.is_cool
drwx------ 2 root   root     65536 Nov  9 18:30 lost+found

Leave a Reply

Your email address will not be published.

Privacy Policy