Environment
Situation
Resolution
To Upgrade an inactive cluster node see: Upgrading Collectors in a Cluster
NOTE: The three servers explained in this article are: Reload Server, NODE1 and NODE2
Reload Server
1. Make sure that Reload 5.05 is first installed on the Reload Server. See: Upgrading Reload
2. Make sure that on the Reload Server a profile exists that represents the profile that will be backing up the GroupWise post office
NOTE: If a profile already exists, and this profile is being converted to a Collector/Server profile then do the following
a. In Reload Web Administration, choose the profile, and go to the Configure tab for the profile
b. Select the Post Office and POA Settings panel
c. On this panel, look at the line that reads: NOTE: Disk Space Required For Collector
d. Make sure that on the volume that houses the post office there is sufficient space to support the collector.
NOTE: The calculation mentioned in step d does not take into account Document Managment Libraries. If DMS is involved, then calculate the size of the GWMDS directory structure X 2. That is the additional disk space you will also need to support a Reload Collector.
e. Set Allow Paired Collector to Enabled to proceed with creating a Reload Collector
NODE1
1. Take these actions on the cluster node that is currently active. We will call this NODE1
2. Install the Reload Collector Software on NODE1
a. Obtain the Reload Collector software from the Reload Server at the following path:
/opt/beginfinite/reload/setup/collector/reload_collector.rpm
b. Once you have copied the Reload Collector RPM file to NODE1 install it with the command:
rpm -ivh reload_collector.rpm
3. Exchange SSH keys with the Reload Server
a. On the server NODE1 type in the following command:
ssh-copy-id -i /root/.ssh/id_rsa.pub<address of the Reload Server>
b. If you have any issues with this command working properly, make sure to work them out before proceed. Ultimately the goal is for you to be able to ssh to the Reload server from NODE1 without being prompted for a password. This functionality is needed to support Rsync.
4. Create a Reload Collector Profile (we will explain how to do a post office, a domain is similar)
a. On NODE1 run Reload administration: reload
b. Choose Create profile.
c. Select Post Office
d. Name the profile with the same name as the profile that already exists on the Reload Server.
NOTE: The profile must first exist on the Reload Server before you can proceed!
e. When prompted as to where to put the Reload profile, put is at <Post Office Path>/collect
f. Then choose Collector as the profile type
g. You will be prompted for the Address of the Reload Server, make sure to the use the same address you used when the ssh-copy-id command.
h. Next indicate the SSH port to the Reload Server, 22 is the default and generally the port you should keep.
i. Next, indicate the Path to the Reload Profile on the Reload Server
j. Next indicate the Path to the Live Post Office which is currently running on NODE1
k. Proceed to Test the profile and resolve any issues as needed.
5. Go to Reload Web Administration for the Reload Collector on NODE1
NOTE: Reload Web Administration is on NODE1 at port 5555 just like the Reload Web Administration on the Reload Server
6. Configure the Post Office profile on NODE1 to be configured for Clustering.
a. Choose the Post Office profile
b. On the Configure tab, select Reload Collector Settings panel
c. Set This Collector is on a Cluster Node to Enabled
7. Perform a backup, and make sure the backup completes normally.
NODE2
1. Take these actions on the cluster node that is currently active. We will call this NODE2
2. Install the Reload Collector Software on NODE2
a. Obtain the Reload Collector software from the Reload Server at the following path:
/opt/beginfinite/reload/setup/collector/reload_collector.rpm
b. Once you have copied the Reload Collector RPM file to NODE2 install it with the command:
rpm -ivh reload_collector.rpm
3. Exchange SSH keys with the Reload Server
a. On the server NODE2 type in the following command:
ssh-copy-id -i /root/.ssh/id_rsa.pub<address of the Reload Server>
b. If you have any issues with this command working properly, make sure to work them out before proceed. Ultimately the goal is for you to be able to ssh to the Reload server from NODE2 without being prompted for a password. This functionality is needed to support Rsync.
4. Obtain the following files from NODE1 and put these files in the same directories on NODE2
a. The contents of the /opt/beginfinite/reload/conf directory (including the ../conf/cron directory)
b. The reload.pem file in /opt/beginfinite/reload/license
5. Now that the license is in place, the Reload Daemon needs to be restarted on NODE2 so that is is aware of the license file. On NODE2 type in the command to restart the Reload Daemon:
reloadd restart
6. Go to Reload Web Administration for the Reload Collector on NODE2
a. Reload Web Administration is on NODE2 at port 5555 just like the Reload Web Administration on the Reload Server
b. Perform a backup, and make sure the backup actually fails.
You should see text similar to this in the Event Log when it fails:
DATE: Fri_Dec_11 TIME: 14:57:25 PROFILE: COLLECT1 - STARTING JOB FOR PROFILE: COLLECT1 DATE: Fri_Dec_11 TIME: 14:57:25 PROFILE: COLLECT1 - DISK SPACE IN USE AT JOB BEGINNING: 46% DATE: Fri_Dec_11 TIME: 14:57:25 PROFILE: COLLECT1 - JOB TYPE: COLLECTOR (INCREMENTAL) BACKUP DATE: Fri_Dec_11 TIME: 14:57:25 PROFILE: COLLECT1 - [GRE_DBS_AGENTS] Start Database Backup DATE: Fri_Dec_11 TIME: 14:57:34 PROFILE: COLLECT1 - [GRE_DBS_OFUSER] Start OFUSER Backup DATE: Fri_Dec_11 TIME: 14:57:38 PROFILE: COLLECT1 - [GRE_DBS] Terminating Processing, Post Office Volume for Profile Not Mounted. DATE: Fri_Dec_11 TIME: 14:57:38 PROFILE: COLLECT1 - [GRE_DBS] This must not be the active node for the Post Office
NODE1 & NODE2
Disable Backup Age Monitoring on both nodes. Backup Age Monitoring is not necessary at the Collector, it is only necessary on the Reload Server. To do this, in Reload Web Administration, edit the Collector profile and on the Configure tab select the Preferences panel. Under Backup Age Monitoring set Monitor How Old the Most Current Backup Is: to Disabled
CONCLUSION
Because you copied the configuration files over from NODE1 to NODE2, they are configured with exactly the same paths which is the essential piece for clustering. Setting changes to NODE1 do not come over to NODE2, however generally that is not an issue because you do not need to change the settings.
Also, since the Reload Profile is located on the same volume as the post office, when the post office volume fails over to NODE2, everything will be intact.