Stop Cluster, unmount and remount oracle shared storage, Start Cluster .

===========

General

===========

In short:
Stop Cluster

unmount  Oracle shared storage

Do DBA work
remount Oracle shared storage

Start Cluster.

How to know if this is a Veritas or Pacemaker?

getaclu

VCS - > Veritas

Starhome Technote

http://10.135.10.64/portal/projects/howto/mount-cl-fs-locally.html

=====================

Pacemaker

=====================

=============

Stop the cluster 

==============
Stop the cluster
#> pcs cluster stop --all

Tag and enable 

#> for i in $(vgscan |grep Ora | cut -d '"' -f 2) ;do vgchange --addtag sometag $i ; vgchange -ay --config 'activation{volume_list=["@sometag"]}' $i ; done

Mount locally

#> mount -t xfs -f -b size=4096 -m crc=0 /dev/OraVg1/db1 /oracle_db/db1
#> mount -t xfs -f -b size=4096 -m crc=0  /dev/OraVg2/Ora_Exp /backup/ora_exp
#> mount -t xfs -f -b size=4096 -m crc=0  /dev/OraVg2/Ora_Online /backup/ora_online
#> mount -t xfs -f -b size=4096 -m crc=0  /dev/OraVg3/db2 /oracle_db/db2

==============
Start the cluster 
==============

Unmount it all locally:

#> umount -f /mnt/oratmp /oracle_db/db1 /backup/ora_exp /backup/ora_online /oracle_db/db2

Delete tag and deactivate 
#> for i in $(vgscan |grep Ora | cut -d '"' -f 2) ;do vgchange -an $i ; vgchange --deltag sometag $i ; done

Restart the cluster 

#> pcs cluster start --all

=====================

VCS

=====================

Take screenshot of current status

df -hP | grep ora

/dev/vx/dsk/OraDg1/db1        79G   26G   54G  33% /oracle_db/db1

/dev/vx/dsk/OraDg2/Ora_Online 159G   16G  143G 10% /backup/ora_online

/dev/vx/dsk/OraDg2/Ora_Exp    100G  7.0G   93G  8% /backup/ora_exp

/dev/vx/dsk/OraDg3/db2        199G  662M  197G  1% /oracle_db/db2

Stop Cluster

hastop -all 

or

hastop -all -force

Perioticly check cluster status until it is unavailable - like the following message:

hastatus -summary

VCS ERROR V-16-1-10600 Cannot connect to VCS engine

VCS WARNING V-16-1-11046 Local system not available

or freeze just the oracle group

hagrp -freez ora_igt_sg

Mount locally the oracle mount points

vxdg import OraDg1

vxdg import OraDg2

vxdg import OraDg3

vxvol -g OraDg1 startall

vxvol -g OraDg2 startall

vxvol -g OraDg3 startall

mount -t vxfs /dev/vx/dsk/OraDg1/db1        /oracle_db/db1

mount -t vxfs /dev/vx/dsk/OraDg2/Ora_Online /backup/ora_online

mount -t vxfs /dev/vx/dsk/OraDg2/Ora_Exp    /backup/ora_exp

mount -t vxfs /dev/vx/dsk/OraDg3/db2        /oracle_db/db2

df -hP | grep ora

Do the Oracle stuff

Dismount locally the oracle mount points

umount /oracle_db/db1 

umount /backup/ora_online

umount /backup/ora_exp

umount /oracle_db/db2

vxvol -g OraDg1 stopall

vxvol -g OraDg2 stopall

vxvol -g OraDg3 stopall

vxdg deport OraDg1

vxdg deport OraDg2

vxdg deport OraDg3

Start the service

Run the following command on ALL cluster nodes:

node a

hastart

node b

hastart

===========

Additional Commands

===========

move service to another node

as root

pcs status

pcs resource cleanup ora_igt_rg

pcs resource disable ora_igt_rg

pcs resource enable ora_igt_rg

pcs resouce show ora_igt_rg

To create a service

as root Check on oracle service definition

/etc/sysconfig/env.oracledb 

for oracle service - check this file:

/etc/systemd/system/dbora.service

This defines the oracle service, ORACLE_HOME, ORACLE_SID

/etc/sysconfig/evn.oracledb

run this:

systemctl status dbora

systemctl status dbora.service

systemctl start dbora.service

systemctl deamon

pcs resource create dbora_igt_ap systemd:dbora op stop interval=0 timeout=120s on-fail="block" monitor interval=30s timeout=600s start interval=0 timeout=120s --group ora_igt_rg

pacemaker cluster commands info

pcs status - this will show cluster status

pcs resource show - this will show current node status

pcs resource show dbora_igt_ap - this will show oracle service info

 Resource: dbora_igt_ap (class=systemd type=dbora)

  Operations: monitor interval=30s timeout=600s (dbora_igt_ap-monitor-interval-30s)

              start interval=0 timeout=120s (dbora_igt_ap-start-interval-0)

              stop interval=0 on-fail=block timeout=120s (dbora_igt_ap-stop-interval-0)

pacemaker cluster commands oracle

pcs resource move ora_igt_rg PIPNVHED901G

pcs resource cleanup ora_igt_rg 

pcs resource disable ora_igt_rg

pcs resource enable ora_igt_rg

pacemaker cluster commands restart oracle service

pcs status

pcs resource cleanup ora_igt_rg

pcs resource disable ora_igt_rg

pcs resource enable ora_igt_rg

pcs resource restart ora_igt_rg PIPNVHED901G

pcs status

pacemaker cluster commands other

--pcs resource disable oracle

--pcs resource enable oracle/19/dbhome_1/rdbms/log/startup

Check corosync in short

root>% vi /etc/corosync/corosync.conf 

root>% pcs cluster sync

server901G: Succeeded

server902G: Succeeded

root>% pcs cluster reload corosync

Corosync reloaded

root>% corosync-cmapctl | grep totem.token

runtime.config.totem.token (u32) = 5000

runtime.config.totem.token_retransmit (u32) = 1190

runtime.config.totem.token_retransmits_before_loss_const (u32) = 4

totem.token (u32) = 5000

Change corosync timeout

1. Edit /etc/corosync/corosync.conf on one of the cluster nodes

Add the required line if does not exist or update the value if the line does exist.

For 5 seconds, set the value to 5000 msec.

totem {

   version: 2

   secauth: off

   cluster_name: rhel7-cluster

   transport: udpu

   rrp_mode: passive

   token: 15000      <--- If this line is missing, add it, otherwise update the value.-->

}

2. Propagate the updated corosync.conf to the rest of the nodes as follows:

pcs cluster sync

3. Reload corosync.
    This command can be run from one node to reload corosync on all nodes and does not require a downtime.

pcs cluster reload corosync

4. Confirm changes

corosync-cmapctl | grep totem.token

For example:

corosync-cmapctl | grep totem.token

runtime.config.totem.token (u32) = 5000

runtime.config.totem.token_retransmit (u32) = 1190

runtime.config.totem.token_retransmits_before_loss_const (u32) = 4