Backup and Restore procedure

migmeneses · 21 August 2025 07:42

Hi All,

On a maas server I’ve tried following the backup and restore procedures from:

How to back up MAAS

However, it does not work correctly:

for backing up and restoring the PostgreSQL DB. I used:
sudo -u postgres pg_dump maas_db -c > “backup/$(date +%s)_maas_dump.sql”
and
sudo -u postgres psql maas_db < /home/ubuntu/backup/$(unix date)_maas_dump.sql
That one actually worked for me, because previously I was having permission issues with the procedure presented in the docs.
but for restoring the snap maas, it is where I am having issues. True we need to reinstall maas before running restore, like mentioned here: Issue restoring from backup
Still, when restarts maas it does not remain active

$ sudo snap services maas
Service          Startup  Current   Notes
maas.supervisor  enabled  inactive  -

$ sudo snap logs maas
2025-08-20T06:44:57Z systemd[1]: Stopped Service for snap application maas.supervisor.
2025-08-20T06:44:57Z systemd[1]: Started Service for snap application maas.supervisor.
2025-08-20T06:44:57Z maas.supervisor[86775]: cannot create user data directory: /root/snap/maas/39713: No such file or directory
2025-08-20T06:44:57Z systemd[1]: snap.maas.supervisor.service: Main process exited, code=exited, status=1/FAILURE
2025-08-20T06:44:57Z systemd[1]: snap.maas.supervisor.service: Failed with result 'exit-code'.
2025-08-20T06:44:57Z systemd[1]: snap.maas.supervisor.service: Scheduled restart job, restart counter is at 5.
2025-08-20T06:44:57Z systemd[1]: Stopped Service for snap application maas.supervisor.
2025-08-20T06:44:57Z systemd[1]: snap.maas.supervisor.service: Start request repeated too quickly.
2025-08-20T06:44:57Z systemd[1]: snap.maas.supervisor.service: Failed with result 'exit-code'.
2025-08-20T06:44:57Z systemd[1]: Failed to start Service for snap application maas.supervisor.

Does anyone have the correct procedure for backing up and restoring MAAS? ( that is not the one presented on the official documentation )

Best Regards

javier-fs · 21 August 2025 08:17

Hi @migmeneses

It looks like snap is looking for a MAAS version that is not found in the system.

39713 is MAAS 3.4. If you ran sudo snap install maas in a clean environment, it will install the latest.

Could you try to install 3.4 and try again?
sudo snap install maas --channel=3.4/stable?

migmeneses · 22 August 2025 08:35

Hello Javier,

These are the steps followed:

Back up MAAS

Create a backup directory:

mkdir -p backup

Dump the database:

sudo -u postgres pg_dump maas_db -c > "backup/$(date +%s)_maas_dump.sql"

Stop MAAS:

sudo snap stop maas

Stop PostgreSQL:

sudo systemctl stop postgresql.service

Save a snap backup:

sudo snap save maas

Export the snapshot:

sudo snap export-snapshot <snapshot-id> backup/$(date +%s)_snapshot

Restart services:

sudo systemctl start postgresql.service
sudo snap restart maas

Restore MAAS

Stop and remove the MAAS snap:

sudo snap stop maas && sudo snap remove --purge maas

Restore the database:

sudo -u postgres psql maas_db < /home/ubuntu/backup/1755604162_maas_dump.sql

(Re)Install MAAS and stop:

sudo snap install --channel=3.4/stable maas
sudo snap services maas ( check if the snap is "active" )
sudo snap stop maas

Import and restore the snapshot:

sudo snap import-snapshot backup/<snapshot>
sudo snap restore <snapshot-id>

Restart services:

sudo systemctl restart postgresql.service
sudo snap restart maas

On sudo snap install --channel=3.4/stable maas I noticed MAAS is not configured if I run maas status and after sudo snap restore it is when i cannot (re)start maas again.

Regards

migmeneses · 22 August 2025 08:38

The idea here is to have a smooth, reliable and confident backup and restore procedure.

skatsaounis · 22 August 2025 10:01

What if you run sudo snap install --revision <revision-number-from-backup> instead, during the restoration? This will bring you from snapstore the same revision as the one present in the snapshot. --channel 3.4/stable will bring the latest revision available in 3.4/stable channel, which at the moment is 39713 for amd64.

migmeneses · 22 August 2025 12:24

These are the last entries:

sudo -u postgres psql maas_db < /home/ubuntu/backup/1755863836_maas_dump.sql
sudo snap import-snapshot backup/1755863946_snapshot
sudo snap install --revision 39713 maas
sudo snap services maas

Service was active:

ubuntu@poc-1:~$ sudo snap services maas                                                                                                                                                
Service          Startup  Current  Notes                                                                                                                                               
maas.supervisor  enabled  active   -

but,

$ sudo maas status
cannot create user data directory: /root/snap/maas/39713: No such file or directory

( be aware of this:

sudo ls -l /root/snap/maas/
total 8
drwxr-xr-x 2 root root 4096 Aug 22 11:53 39713
drwxr-xr-x 2 root root 4096 Aug 22 11:53 common
lrwxrwxrwx 1 root root    5 Aug 22 12:10 current -> 39713

so the directory exist )

Restore:

sudo snap restore <snapshot-id>

Restart services:

sudo systemctl restart postgresql.service
sudo snap restart maas

and from there, maas is not active anymore.
I think the problem sits just before running snap restore

migmeneses · 22 August 2025 12:25

logs:

ubuntu@poc-1:~$ sudo snap logs maas
2025-08-22T12:14:02Z systemd[1]: Stopped Service for snap application maas.supervisor.
2025-08-22T12:14:02Z systemd[1]: Started Service for snap application maas.supervisor.
2025-08-22T12:14:02Z maas.supervisor[7326]: cannot create user data directory: /root/snap/maas/39713: No such file or directory
2025-08-22T12:14:02Z systemd[1]: snap.maas.supervisor.service: Main process exited, code=exited, status=1/FAILURE
2025-08-22T12:14:02Z systemd[1]: snap.maas.supervisor.service: Failed with result 'exit-code'.
2025-08-22T12:14:02Z systemd[1]: snap.maas.supervisor.service: Scheduled restart job, restart counter is at 5.
2025-08-22T12:14:02Z systemd[1]: Stopped Service for snap application maas.supervisor.
2025-08-22T12:14:02Z systemd[1]: snap.maas.supervisor.service: Start request repeated too quickly.
2025-08-22T12:14:02Z systemd[1]: snap.maas.supervisor.service: Failed with result 'exit-code'.
2025-08-22T12:14:02Z systemd[1]: Failed to start Service for snap application maas.supervisor.

javier-fs · 22 August 2025 16:22

I went through the backup process using the PostgreSQL 16 and MAAS 3.6 and I reached the same outcome.

Could you open a bug with this information?

migmeneses · 26 August 2025 12:15

Filed bug:

zmance · 26 August 2025 18:37

For my own curiosity, were you able to roll back without issues?

migmeneses · 27 August 2025 07:05

No,
After running maas snap restore, it does not restore the maas snap correctly
The log says it cannot find directory /root/snap/maas/
when it actually exists

migmeneses · 27 August 2025 07:09

Weird,
Today, I just started maas

sudo snap start maas

Checked:

sudo snap service maas

and is running

lloydwaltersj · 27 August 2025 09:09

Interesting that re-starting MAAS after some downtime seems to have fixed it. I wonder if there’s some long running process that didn’t complete in time for your first MAAS restart?

migmeneses · 27 August 2025 11:42

Could be,
Be aware that my actions Create Backup and Restore are happening continuously.

Maybe, there should be a cooldown period before running restore.

migmeneses · 28 August 2025 08:48

Yesterday, 2025-08-27
Created a new backup, trying to restore it today, and back to square 1

migmeneses · 28 August 2025 08:52

syslog:

Aug 28 08:45:40 poc-1 systemd[1]: Stopped PostgreSQL RDBMS.
Aug 28 08:45:40 poc-1 systemd[1]: Stopping PostgreSQL RDBMS...
Aug 28 08:45:40 poc-1 systemd[1]: Stopping PostgreSQL Cluster 17-main...
Aug 28 08:45:45 poc-1 systemd[1]: postgresql@17-main.service: Deactivated successfully.
Aug 28 08:45:45 poc-1 systemd[1]: Stopped PostgreSQL Cluster 17-main.
Aug 28 08:45:45 poc-1 systemd[1]: postgresql@17-main.service: Consumed 7min 19.342s CPU time.
Aug 28 08:45:45 poc-1 systemd[1]: Starting PostgreSQL Cluster 17-main...
Aug 28 08:45:47 poc-1 systemd[1]: Started PostgreSQL Cluster 17-main.
Aug 28 08:45:47 poc-1 systemd[1]: Starting PostgreSQL RDBMS...
Aug 28 08:45:47 poc-1 systemd[1]: Finished PostgreSQL RDBMS.
Aug 28 08:45:47 poc-1 systemd[1]: Started Service for snap application maas.supervisor.
Aug 28 08:45:48 poc-1 maas.supervisor[203920]: cannot create user data directory: /root/snap/maas/39713: No such file or directory
Aug 28 08:45:48 poc-1 systemd[1]: snap.maas.supervisor.service: Main process exited, code=exited, status=1/FAILURE
Aug 28 08:45:48 poc-1 systemd[1]: snap.maas.supervisor.service: Failed with result 'exit-code'.
Aug 28 08:45:48 poc-1 systemd[1]: snap.maas.supervisor.service: Scheduled restart job, restart counter is at 1.

migmeneses · 1 September 2025 10:00

Today 2025-09-01,
I was available to start MAAS services without any major issues.
sudo snap start maas

So, it looks like:

The regular procedure shown for backup and restore in the official docs is not reliable.
There is a cool-down period before (re)starting Maas services successfully!

I am going to try this on another environment and see the results!

migmeneses · 1 September 2025 12:59

Tried on DigitalOcean droplets, and same result.