In this blog i will present step to configure non-Cluster DB2 9.7 HADR configuration. Following topics are covered in this blog
a) Pre-Configuration Details
b) Setting up non-cluster DB2 9.7 HADR
c) Manual Starting/Stopping HADR steps
d) Manul Failover/Failback Steps
e) Force Failover/Failback Steps
Part 0 : DB2 9.7 HADR with TSA Part 00 - Introduction
a) Pre-Configuration Details
b) Setting up non-cluster DB2 9.7 HADR
c) Manual Starting/Stopping HADR steps
d) Manul Failover/Failback Steps
e) Force Failover/Failback Steps
Part 0 : DB2 9.7 HADR with TSA Part 00 - Introduction
A) Pre-Configuation Details
| Item # | Description | Command | |
|---|---|---|---|
| System On which command will be executed => | Primary | Standby | |
| 1 | IP Address and Hostname | a) Public Network IP => 192.168.5.22 Hostname => Mumbai b) Private Network (Optional) IP => 192.168.5.23 Hostname => priv01 | a) Public Network IP => 192.168.5.33 Hostname => London b) Private Network (Optional) IP => 192.168.5.34 Hostname => priv02 |
| 2 | Virtual IP (Configured by TSA) | IP => 192.168.5.55 Virtual Hostname => newyork | |
| 3 | Create New Mount point (Repet step a to Step i for each new hard disk added to the virtual machine) | a) Add two hard disk to virtual machine and restart the virtual machine b) Log on to the OS to see the hard disk is added ls /dev/sd*-- All the hard disk which show /dev/sd[alphabet] are newly added hard disk. -- All the hard disk which shows /dev/sd[albhabet][number] are the existing hard disk on the server. c) Partition the hard disk using FDISK. The sequence of command to FDISK are "n", "p", "1", "Return(Enter Button on Keyboard)", "Return(Enter Button on Keyboard)","p" and "w" d) New partition get added to the list ls /dev/sd* e) Create the physical Volumne pvcreate /dev/sd[alphabet][number] where : alphabet = Character of new added disk Number = Number of the newly formated disk For Example, pvcreate /dev/sdc1 f) Create the file system on the volumne mkfs -t ext3 -m 1 -v /dev/sd[alphabet][number] For example:- mkfs -t ext3 -m 1 -v /dev/sdc1 g) Check the new FS created tune2fs -l /dev/sd[alphabet][number] For example;- tune2fs -l /dev/sdc1 h) Create the mount directory mkdir /db2fs01 i) Modify the /etc/fstab file to include following line /dev/sdc1 /db2fs01 ext3 defaults 1 1 | |
| 4 | Check OS details | a) OS Versionoslevel -a b) Patch details c) Kernal Level | |
| 5 | Check the Network Details | a) Public IP Address b) Private IP Address c) Virtual IP Address | |
| 6 | Check DB2 details | a) Check DB2 version db2level b) Mount point for DATA c) Mount point for Logs d) Mount point for archive logs | |
B) Setting up non-Cluster HADR
| Item # | Description | Command | |
|---|---|---|---|
| System On which command will be executed => | Primary | Standby | |
| 1 | Public Network and hostname | Network device : eth0 Hostname : Mumbai IP => 192.168.5.22 | Network device : eth0 Hostname : London IP => 192.168.5.33 |
| 2 | Virtual IP | Hostname => newyork Virtual IP => 192.168.5.55 | |
| 3 | Add the public/Virtual IP details to /etc/hosts file | 192.168.5.22 mumbai 192.168.5.33 london 192.168.5.55 newyork | 192.168.5.22 mumbai 192.168.5.33 london 192.168.5.55 newyork |
| 4 | Check the connectivity between nodes | ping mumbai ping london | ping mumbai ping london |
| 5 | Create DB2 Instance Owner and DB2 Fence Owner Group GID and GROUP name must be same on both node | groupadd -g 1003 db2iadm1 groupadd -g 102 db2fadm1 | groupadd -g 1003 db2iadm1 groupadd -g 102 db2fadm1 |
| 6 | Create DB2 Instance Owner user and DB2 fence owner user | useradd -u 21000 -g 1003 -d /home/db2inst1 -m -c "DB2 Instance Owner user" db2inst1 useradd -u 108 -g 102 -d /home/db2fenc1 -m -c "DB2 Fence Owner user" db2fenc1 | useradd -u 21000 -g 1003 -d /home/db2inst1 -m -c "DB2 Instance Owner user" db2inst1 useradd -u 108 -g 102 -d /home/db2fenc1 -m -c "DB2 Fence Owner user" db2fenc1 |
| 7 | Check DB2 ports in /etc/services files.Must be same on each of the nodes | a)DB2 TCPIP Communication port 6000 b) DB2 HADR port DB2_HADR_MUM=50011/tcp DB2_HADR_LON=50012/tcp | a)DB2 TCPIP Communication port 6000 b) DB2 HADR port DB2_HADR_MUM=50011/tcp DB2_HADR_LON=50012/tcp |
| 8 | Check ~/sqllib/db2nodes.cfg under DB2 instance owner user | 0 mumbai 0 | 0 london 0 |
| 9 | Create directory a) Backup b) Archive log | chmod 777 /db2fs mkdir /db2fs/db2bkp mkdir /db2fs/db2arch mkdir /db2fs/db2log | chmod 777 /db2fs mkdir /db2fs/db2bkp mkdir /db2fs/db2arch mkdir /db2fs/db2log |
| 10 | Update DB2 parameters | a) DB2 TCPIP communication parameters db2set DB2COMM=tcpip db2 "update dbm cfg using svcename 60000" db2stop force;db2start b) Enable DB2 to use SSH password less communication db2set -i db2inst1 DB2RSHCMD=/usr/bin/ssh db2stop force;db2start c) Update DB2 HADR database parameters Note Parameter values are differenct on each node update db cfg for sample using logarchmeth1 'DISK:/db2fs/db2arch' update database configuration for sample using LOGINDEXBUILD ON update db cfg for Sample using hadr_local_host mumbai update db cfg for Sample using hadr_remote_host london update db cfg for Sample using hadr_local_svc DB2_HADR_MUM update db cfg for Sample using hadr_remote_svc DB2_HADR_LON update db cfg for Sample using hadr_remote_inst db2inst1 update db cfg for Sample using hadr_timeout 120 update db cfg for Sample using hadr_peer_window 120 update db cfg for Sample using hadr_syncmode nearsync | a) DB2 TCPIP communication parameters db2set DB2COMM=tcpip db2 "update dbm cfg using svcename 60000" db2stop force;db2start b) Enable DB2 to use SSH password less communication db2set -i db2inst1 DB2RSHCMD=/usr/bin/ssh db2stop force;db2start c) Update DB2 HADR database parameters Note Parameter values are differenct on each node. This values are set after RESTORING the database on STANDYBY update db cfg for Sample using hadr_local_host london update db cfg for Sample using hadr_remote_host mumbai update db cfg for Sample using hadr_local_svc DB2_HADR_LON update db cfg for Sample using hadr_remote_svc DB2_HADR_MUM update db cfg for Sample using hadr_remote_inst db2inst1 update db cfg for Sample using hadr_timeout 120 update db cfg for Sample using hadr_peer_window 120 update db cfg for Sample using hadr_syncmode nearsync |
| 11 | Backup the database | backup database sample online to /db2fs01/db2bkp | No Operation |
| 12 | Copy the backup image to Standby machine | scp SAMPLE.0.db2inst1.NODE0000.CATN0000.20101021100242.001 db2inst1@hadrtest2:/home/db2inst1/bkp | No Operation |
| 13 | Create new database on standby from the offline backup | No Operation | restore database sample replace history file This will set the database in rollforward pedning state as we created the database from online full backup of primary database after enabling primary database for archival logging. It will be in ROLLFORWARD PENDING STATE. No changes needed. DO NOT TAKE THE DATABASE OUT OF ROLL FORWARD PENDING STATE |
| 14 | Start HADR on Standby | No Operation | start hadr on database sample as standby |
| 15 | Start HADR on Primary | start hadr on database sample as primary | No Operation |
| 16 | Check the HADR Status (must be same on both nodes) | db2pd -hadr -db sample HADR State = Peer | db2pd -hadr -db sample HADR State = Peer |
| 17 | Check if HADR is working | 1) Archive current log archive log for db sample 2) Check the first log sequence get db cfg for sample | grep -i first 3) Check if the archive log file is shifted to Standby. (May take some time) db2pd -hadr -db sample | 1) Archive current log archive log for db sample 2) Check the first log sequence get db cfg for sample | grep -i first 3) Check if the archive log file is shifted to Standby. (May take some time) db2pd -hadr -db sample |
C) Manual Starting/Stopping HADR Steps
| Item # | Description | Command | |
|---|---|---|---|
| Starting HADR steps | Primary | Standby | |
| 1 | Start Instance on Primary/Standby | db2start | db2start |
| 2 | Start HADR on Standby machine | No Operation | start hadr on database sample as standby |
| 3 | Start HADR on Primary Machine | start hadr on database sample as primary | No Operation |
| 4 | Check if DB2 HADR is running | db2pd -hadr -db sample | db2pd -hadr -db sample |
| Stopping HADR steps | Primary | Standby | |
| 1 | Stop HADR on database | stop hadr on database sample | No Operation |
| 2 | Deactivate Database | deactivate database sample | No Operation |
| 3 | Stop instance | db2stop force | No Operation |
| 4 | Deactivate the standby database | No Operation | deactivate database sample |
| 5 | Stop HADR on database | No Operation | stop hadr on database sample |
| 6 | Stop Instance | No Operation | db2stop force |
D) Manul Failover/Failback Steps
| Item # | Description | Command | |
|---|---|---|---|
| Failover Operation | Primary | Standby | |
| 1 | Identify the Standby node | db2pd -db sample -hadr | db2pd -db sample -hadr |
| 2 | Execute takeover command on standby | No Operation | takeover hadr on database sample |
| Failback Operation | Primary | Standby | |
| 1 | Identify the new Standby node | db2pd -db sample -hadr | db2pd -db sample -hadr |
| 2 | Execute takeover command on old Primary which is now new Standby after failover | takeover hadr on database sample | No Operation |
E) Force Failover
| Item # | Description | Command | |
|---|---|---|---|
| Force Failover (Test Scenario) | Primary | Standby | |
| 1 | Kill the DB2 instance procee to simulate Test Scenario | ps -eaf | grep -i db2sysc kill -9 | No Operation |
| 2 | Check the status of the standby | No Operation | db2pd -hadr -db sample HADR State will be in Disconnected state |
| 3 | Execute the force failover on Standby | No Operation | takeover hadr on database sample by force |
| 4 | Bring the primary Server online | db2start start hadr on database sample as standby | No Operation |
| 5 | Failback to Primary machine | takeover hadr on database sample | No Operation |
| 6 | Check HADR status | db2pd -hadr -db sample | db2pd -hadr -db sample |
