1. Backing up
In this section you will find the process for backing up the MSSQL DB that is running inside a docker container for the Lakehouse Monitor.
1. Stop the containers
First and foremost you’ll need to stop the app and the database:
$ docker compose down
You should see a message similar to
[+] Running 3/3 ⠿ Container lakehouse Removed 1.1s ⠿ Container bplm Removed 0.4s ⠿ Network centos_bplm_net Removed
This tells you that the Lakehouse monitor and the DB are down.
2. Locate the MSSQL data
Now you will need to identify the path on the host system where your MSSQL data resides. Do this by using the following approach:
$ sudo -i
# docker volume ls
DRIVER VOLUME NAME
local centos_mssql_data
local centos_spring_session
# docker volume inspect centos_mssql_data
[
{
"CreatedAt": "2023-03-15T07:27:39Z",
"Driver": "local",
"Labels": {
"com.docker.compose.project": "centos",
"com.docker.compose.version": "2.16.0",
"com.docker.compose.volume": "mssql_data"
},
"Mountpoint": "/var/lib/docker/volumes/centos_mssql_data/_data",
"Name": "centos_mssql_data",
"Options": null,
"Scope": "local"
}
]
# cd /var/lib/docker/volumes/centos_mssql_data/_data
# ls -la
total 8
drwxrwx---. 7 root root 73 Apr 5 12:36 .
drwx-----x. 3 root root 19 Mar 15 07:27 ..
drwxr-xr-x. 5 10001 root 74 Mar 15 07:27 .system
drwxr-xr-x. 2 10001 root 24 Apr 5 12:36 backup
drwxr-xr-x. 2 10001 root 4096 Mar 15 07:27 data
drwxr-xr-x. 2 10001 root 4096 Apr 6 10:10 log
drwxr-xr-x. 2 10001 root 25 Mar 15 07:27 secrets
3. Archive the data
Once in this step, you can create a backup archive of this entire directory using the tar command
# tar zcvf ~centos/mssql-backup.tgz . ./ ./.system/ ./.system/system/ ./.system/system/Temp/ ./.system/system/lsasetup.log ....
Note: in my script above I have saved the tgz file in the centos home folder, as such: ~centos/mssql-backup.tgz
Now you need to change the owner of the file so that your user, in my case centos can access the file.
# cd ~centos # chown centos mssql-backup.tgz
4. Start the containers
You can now leave the privileged session using the exit command and restart the Lakehouse containers
# exit $ docker compose up -d # make sure you're in the same directory where your docker-compose.yml file is located
The output should look similar to this:
[+] Running 3/3 ⠿ Network centos_bplm_net Created 0.1s ⠿ Container bplm Started 0.9s ⠿ Container lakehouse Started
You can check that the containers are up and running with the following command:
$ docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES 240a0312e897 blueprint.azurecr.io/bpcs/lakehouse-optimizer:20230406.1 "./start-app.sh" 4 seconds ago Up 3 seconds 2222/tcp, 0.0.0.0:4040->4040/tcp, :::4040->4040/tcp, 0.0.0.0:80->8080/tcp, :::80->8080/tcp bplm 0287182a2d92 mcr.microsoft.com/mssql/server:2019-latest "/opt/mssql/bin/perm…" 4 seconds ago Up 3 seconds 0.0.0.0:1433->1433/tcp, :::1433->1433/tcp lakehouse
5. Use scp to download the data
You can use scp to copy the tgz backup to your local machine and from there upload it to S3 or any other cloud storage solution you use.
### On your local computer $ scp <your username>@<your server host>:<path to mssql-backup.tgz> ./
For example, in my case:
$ scp centos@1.2.3.4:/home/centos/mssql-backup.tgz ./
This will copy the archive to my local computer from which point I can upload it to S3.
2. Restore
In this section you will find the process for using the backup created in the first section to restore the MSSQL DB that’s running inside a docker container for the Lakehouse Monitor.
Assumptions
I’m assuming that you have the mssql-backup.tgz file on your server under the centos user home folder, if not, you need to use scp to copy it from your local machine, so on your local machine you run this:
$ scp <path to tar>/mssql-backup.tgz <your username>@<your server host>:
1. Stop the containers
First and foremost you’ll need to stop the app and the database:
$ docker compose down
You should see a message similar to
[+] Running 3/3 ⠿ Container lakehouse Removed 1.1s ⠿ Container bplm Removed 0.4s ⠿ Network centos_bplm_net Removed
This tells you that the Lakehouse monitor and the DB are down.
2. Locate the MSSQL data
Now you will need to identify the path on the host system where your MSSQL data resides. Do this by using the following approach:
$ sudo -i
# docker volume ls
DRIVER VOLUME NAME
local centos_mssql_data
local centos_spring_session
# docker volume inspect centos_mssql_data
[
{
"CreatedAt": "2023-03-15T07:27:39Z",
"Driver": "local",
"Labels": {
"com.docker.compose.project": "centos",
"com.docker.compose.version": "2.16.0",
"com.docker.compose.volume": "mssql_data"
},
"Mountpoint": "/var/lib/docker/volumes/centos_mssql_data/_data",
"Name": "centos_mssql_data",
"Options": null,
"Scope": "local"
}
]
# cd /var/lib/docker/volumes/centos_mssql_data/_data
# ls -la
total 8
drwxrwx---. 7 root root 73 Apr 5 12:36 .
drwx-----x. 3 root root 19 Mar 15 07:27 ..
drwxr-xr-x. 5 10001 root 74 Mar 15 07:27 .system
drwxr-xr-x. 2 10001 root 24 Apr 5 12:36 backup
drwxr-xr-x. 2 10001 root 4096 Mar 15 07:27 data
drwxr-xr-x. 2 10001 root 4096 Apr 6 10:10 log
drwxr-xr-x. 2 10001 root 25 Mar 15 07:27 secrets
3. Clean the _data dir
You can either remove all of the data or, for extra-precations move it into another directory of your choosing.
3.1 Deleting the data
# rm -rfv * .system
3.2 Moving the data
# mkdir bak # mv * .system bak mv: cannot move 'bak' to a subdirectory of itself, 'bak/bak' # ls -la total 0 drwxrwx---. 3 root root 17 Apr 6 10:19 . drwx-----x. 3 root root 19 Mar 15 07:27 .. drwxr-xr-x. 7 root root 73 Apr 6 10:19 bak
4. Restoring the data
With the destination directory empty, it is now time to bring in the backup data.
# cd ~centos # tar zxvf mssql-backup.tgz -C /var/lib/docker/volumes/centos_mssql_data/_data/ ./ ./.system/ ./.system/system/ ./.system/system/Temp/ ./.system/system/lsasetup.log ./.system/system/debug/ ./.system/system/debug/PASSWD.LOG ./.system/system/system32/ ......
As you can see I used the tar command to extract the contents of mssql-backup.tgz into the _data directory (using the -C flag).
Now you can check that the data was correctly restored using the following approach:
# cd - /var/lib/docker/volumes/centos_mssql_data/_data # ls -la total 8 drwxrwx---. 8 root root 84 Apr 5 12:36 . drwx-----x. 3 root root 19 Mar 15 07:27 .. drwxr-xr-x. 5 10001 root 74 Mar 15 07:27 .system drwxr-xr-x. 2 10001 root 24 Apr 5 12:36 backup drwxr-xr-x. 7 root root 73 Apr 6 10:19 bak drwxr-xr-x. 2 10001 root 4096 Mar 15 07:27 data drwxr-xr-x. 2 10001 root 4096 Apr 6 10:10 log drwxr-xr-x. 2 10001 root 25 Mar 15 07:27 secrets
5. Starting the containers
You can now exit the privileged bash session and start the containers.
# exit $ docker compose up -d [+] Running 3/3 ⠿ Network centos_bplm_net Created 0.1s ⠿ Container lakehouse Started 0.8s ⠿ Container bplm Started
Give it 2 minutes during which time you can check to see that no restarts happen using docker ps
After 2 minutes without restarts you can check the app logs to see that the app started
$ docker logs bplm | grep Started ..... 2023-04-06 10:21:08.024 INFO 1 --- [ main] scala.App : Started App in 34.412 seconds (JVM running for 36.438)
6. Clean up of the bak folder
If in step 3 you chose to create the bak folder, it is now safe to delete it.
$ sudo -i # cd /var/lib/docker/volumes/centos_mssql_data/_data/ # ls -la total 8 drwxrwx---. 8 root root 84 Apr 5 12:36 . drwx-----x. 3 root root 19 Mar 15 07:27 .. drwxr-xr-x. 5 10001 root 74 Mar 15 07:27 .system drwxr-xr-x. 2 10001 root 24 Apr 5 12:36 backup drwxr-xr-x. 7 root root 73 Apr 6 10:19 bak drwxr-xr-x. 2 10001 root 4096 Mar 15 07:27 data drwxr-xr-x. 2 10001 root 4096 Apr 6 10:10 log drwxr-xr-x. 2 10001 root 25 Mar 15 07:27 secrets # rm -rfv bak
Add Comment