Configure TAS Datastore servers for high availability

Configure the GlusterFS service on each of the TAS Datastore servers in high-availability environments. Typically, you can have from three up to five Datastore servers.

Before you begin

Prepare TAS Datastore server (TAS 8)
Verify that the firewall ports are open

Procedure

On one of the TAS Datastore servers, do the following:
1. If the directories for the Configuration Service and the Data Export Service do not exist, create them, and assign permissions:
  Copy
```
1
2
3
mkdir /<data_folder>/configservice_data
chmod 770 /<data_folder>/configservice_data
chown -R tas_inst:tas <data_folder>/configservice_data
```
  Copy
```
1
2
3
mkdir /<data_folder>/reports
chmod 770 /<data_folder>/reports
chown -R tas_inst:tas <data_folder>/reports
```
  where:
  - <data_folder> is the folder defined for data storage, and is by default opt/app/data.
  - configservice_data is the name of the Configuration Service directory and must not be changed.
  - reports is the name of the Data Export Service directory and must not be changed
2. Verify that GlusterFS is installed:
  Copy
```
gluster --version
```
3. Verify that GlusterFS is enabled and started:
  Copy
```
systemctl status glusterd
```
4. Create and assign permission to the bricks directories within the data directory:
  Copy
```
1
2
3
4
mkdir -p /<data_folder>/bricks/brick1/g1
mkdir -p /<data_folder>/bricks/brick2/g2
chown -R tas_inst:tas <data_folder>/bricks/
chmod 770 -R <data_folder>/bricks/
```
  where:
  - <data_folder> is the folder defined for data storage, and is by default opt/app/data.
5. Repeat step a through step d for each of the TAS Datastore servers in your deployment.
On one of the TAS Datastore servers, do the following:
1. Probe the other TAS Datastore server peers:
  Copy
```
gluster peer probe <fqdn_tas_datastore1>
...
gluster peer probe <fqdn_tas_datastore5>
```
  where:
  - <fqdn_tas_datastore1> ... <fqdn_tas_datastore5> are the FQDNs of the other TAS Datastore servers in your deployment, and can range from three to five servers.
  - For example, if you have three Datastore servers, and you are running the command from TAS Datastore server 2, you would probe the TAS Datastore servers 1 and 3.
2. Verify the status of the peer servers:
  Copy
```
gluster peer status
```
  Example of expected output for three Datastore servers:
  
  Number of Peers: 2
  
  Hostname: hostname1.labqm.com
  
  Uuid: 973b360c-4d7d-457e-ab6a-045a2d2b706a
  
  State: Peer in Cluster (Connected)
  
  Hostname: hostname2.labqm.com
  
  Uuid: c49b4146-9747-49a7-8f09-254cde66ddd9
  
  State: Peer in Cluster (Connected)
3. Create the following GlusterFS volumes:
  Copy
```
gluster volume create config_service_data replica <num_datastore_servers> \
<fqdn_tas_datastore1>:<data_folder>/bricks/brick1/g1 \
...
<fqdn_tas_datastore5>:<data_folder>/bricks/brick1/g1 force
```
  Copy
```
gluster volume create dataexport_data replica <num_datastore_servers> \
<fqdn_tas_datastore1>:<data_folder>/bricks/brick2/g2 \
...
<fqdn_tas_datastore5>:<data_folder>/bricks/brick2/g2 force
```
  where:
  - config_service_data is the volume name, and must not be changed.
  - dataexport_data is the volume name of the report folder, and must not be changed.
  - <fqdn_tas_datastore1> ... <fqdn_tas_datastore5> is the FQDN of every TAS Datastore server in your deployment, corresponding to <num_datastore_servers>. Each Datastore server must have a brick with its FQDN.
  - <data_folder> is the folder defined for data storage, and is by default /opt/app/data.
4. Fine-tune the GlusterFS volumes:
  Copy
```
gluster volume set config_service_data nfs.disable ON \
cluster.quorum-type auto cluster.favorite-child-policy majority
```
  Copy
```
gluster volume set dataexport_data nfs.disable ON \
cluster.quorum-type auto cluster.favorite-child-policy majority
```
5. Start the GlusterFS volumes:
  Copy
```
gluster volume start config_service_data 
```
  Copy
```
gluster volume start dataexport_data 
```
6. Verify that the GlusterFS volumeswere set up successfully:
  Copy
```
gluster volume info 
```
  Example of expected output:
  
  Volume Name: configservice_data
  
  Type: Replicate
  
  Volume ID: 4d7dafa2-da67-45dc-aa40-5e81d4103e38
  
  Status: Started
  
  Snapshot Count: 0
  
  Number of Bricks: 1 x 3 = 3
  
  Transport-type: tcp
  
  Bricks:
  
  Brick1: hostname1.labqm.com:/opt/app/data/bricks/brick1/g1
  
  Brick2: hostname1.labqm.com:/opt/app/data/bricks/brick1/g1
  
  Brick3: hostname1.labqm.com:/opt/app/data/bricks/brick1/g1
  
  Options Reconfigured:
  
  performance.client-io-threads: off
  
  nfs.disable: On
  
  transport.address-family: inet
  
  cluster.quorum-type: auto
On each of the TAS Datastore servers, mount the volumes for the Configuration Service and Data Export Service:
1. On the first TAS Datastore server, run:
  Copy
```
echo "<fqdn_tas_datastore>:/config_service_data \
/<data_folder>/configservice_data glusterfs defaults,_netdev" \
>> /etc/fstab
```
  Copy
```
echo "<fqdn_tas_datastore1>:/dataexport_data \
/<data_folder>/reports glusterfs defaults,_netdev" \
>> /etc/fstab
```
  and then run:
  Copy
```
mount -a
```
  where:
  - <fqdn_tas_datastore> is the FQDN of the TAS Datastore server for which you are mounting the volume.
  - configservice_data is the name of the Configuration Service directory and must not be changed.
  - dataexport_data is the name of the Data Export Service volume and must not be changed.
  - reports is the name of the Data Export Service directory and must not be changed.
  - <data_folder> is the folder defined for data storage, and is by default opt/app/data.
2. Repeat step a for each Datastore server in your deployment.

What to do next

Configure TLS for TAS servers