Converting from Multiple Mount Points to a Single Mount Point

In versions of SolOS prior to 9.12, storage-elements were typically externalized to separate storage volumes and mounted to the container using separate mount points. In version 9.12 and later, these storage-elements are now collected under a single object called a storage-group, which uses a single mount point.

Before you use these procedures, ensure that you calculate the storage requirements for your container. You must make sure your external storage is sufficient accommodate:

the minimum required storage, PLUS
additional storage for higher values of system scaling parameters, PLUS
space for spooled messages

For details, see:

If you have upgraded from SolOS version 9.11 or earlier to SolOS 9.12 or later, you can use the following procedures to convert your software event broker(s) from using multiple mount points to using a single mount point:

Standalone
High Availability

Standalone

To update a standalone container image to use a single mount point instead of multiple mount points, do the following:

These steps assume you are using Docker Engine. Similar commands are available for other container runtimes.

Make sure all the required data are in the volumes. To double check what volumes you have in the environment, run the following command:
```
docker volume ls
```
The response should look like the following:
```
DRIVER    VOLUME NAME
local     adb
local     adbBackup
local     diagnostics
local     spool
local     jail
local     var
```
Stop the Solace container:
```
docker stop --time 1200 solace
```
Remove the container:
```
docker rm solace
```
On the host operating system, create new directories in the location where you want the new storage-group to reside.
You can use any name you want for the top-level directory, but the storage-element directories must be as shown in the example (spool, spool-cache, var, spool-cache-backup, jail, and diagnostics)
The following example uses /mnt/solace as the top-level directory:
```
mkdir /mnt/solace/spool
mkdir /mnt/solace/spool-cache
mkdir /mnt/solace/var
mkdir /mnt/solace/spool-cache-backup  
mkdir /mnt/solace/jail   
mkdir /mnt/solace/diagnostics
```
Set the owner and group of the new directories to the container owner and container group by running the following command, replacing <container-user> and <container-group> with the actual values:
```
chown -R <container-user>:<container-group> /mnt/solace/
```
For example, if you use the default values for the container owner (1000001) and group (0), run the following command:
```
chown -R 1000001:0 /mnt/solace/
```

To determine what the mount point was for the old volumes, run the following command:

docker volume inspect adb

The response should look like the following. The "Mountpoint" field shows where the adb volume was mounted:

[
    {
        "CreatedAt": "0001-01-01T00:00:00Z",
        "Driver": "local",
        "Labels": {},
        "Mountpoint": "/var/lib/docker/volumes/adb/_data",
        "Name": "adb",
        "Options": {},
        "Scope": "local"
    }
]

Copy the folder contents corresponding to each volume to the new location (note the differences between the old and new storage-element names):

shopt -s dotglob
mv /var/lib/docker/volumes/adb/_data/* /mnt/solace/spool-cache
mv /var/lib/docker/volumes/internalSpool/_data/* /mnt/solace/spool
mv /var/lib/docker/volumes/diagnostics/_data/* /mnt/solace/diagnostics
mv /var/lib/docker/volumes/var/_data/* /mnt/solace/var
mv /var/lib/docker/volumes/adbBackup/_data/* /mnt/solace/spool-cache-backup
mv /var/lib/docker/volumes/jail/_data/* /mnt/solace/jail

Run the following command to create a new container that mounts the new storage-group directory (/mnt/solace) as a bind mount:

 docker create --network=host --uts=host --shm-size=1g --ulimit core=-1 \
--ulimit memlock=-1 --ulimit nofile=2448:42192 --env 'username_admin_globalaccesslevel=admin' \
--env 'username_admin_password=admin' --name=solace --mount type=bind,source=/mnt/solace,destination=/var/lib/solace,ro=false \
solace-pubsub-enterprise:<version>

If you used the Docker volume API to create the old named volumes, you can now remove them by running the following command (they are no longer used):
```
docker volume rm adb adbBackup diagnostics internalSpool jail var
```

High Availability

To update the software event brokers in a high-availability group to use a single mount point instead of multiple mount points, do the following:

In the steps that follow, the primary messaging node is referred to as solace-primary and the backup messaging node is referred to as solace-backup.

Verify that the redundancy status is correct by running the show redundancy command on each messaging node:
- On both messaging nodes, ensure Redundancy Configuration Status is Enabled and Redundancy Status is Up
- On the primary node, the Message Spool Status should be AD-Active
- On the backup node, the Message Spool Status should be AD-Standby
On the backup node, perform the steps listed above for a standalone image.

On the backup node, run the show redundancy command to ensure that the Redundancy Status is Up:

solace-backup> show redundancy
Configuration Status     : Enabled
Redundancy Status        : Up

On the backup node, if config-sync was active, ensure it is active again:

solace-backup> show config-sync
Admin Status            : Enabled
Oper Status             : Up

On the primary node, release activity to the backup node:

solace-primary> enable
solace-primary> configure
solace-primary> redundancy release-activity

On the backup node, run the show redundancy command to ensure that the Message Spool Status is now AD-Active.
On the primary node, perform the steps listed above for a standalone image.

On the primary node, re-claim activity:

solace-primary> enable
solace-primary> configure
solace-primary> no redundancy release-activity

On the primary node, run the show redundancy command to ensure that the Redundancy Status is Up:
```
solace-primary> show redundancy
Configuration Status     : Enabled
Redundancy Status        : Up
```
At this point, the backup node is still handling messaging traffic. If required, you can manually force the backup node to release activity to the primary node.

On the primary node, if config-sync was active, ensure it is active again:

solace-primary> show config-sync
Admin Status            : Enabled
Oper Status             : Up

On the monitoring node, perform the steps listed above for a standalone image.

Provide feedback