PostgreSQL17 does not start after Barman PITR recovery

[sudha]

Hello all,

I’m testing Point-in-Time Recovery (PITR) with Barman and I’ve run into an issue I’d like some clarity on.

• I have below backup 
  
  where the end time says 2025-09-02 22:00:35.116829+00:00, but since it is continuously receiving the WAL files , I attempted a PITR using

barman recover --get-wal --remote-ssh-command "ssh postgres@sudha-dev-vm" --target-time="2025-09-03 10:38:07.471841+00:00" bitbucket-pgdb1-u22 20250902T220001 /var/lib/postgresql/17/main

The barman log says "Your PostgreSQL server has been successfully prepared for recovery!"
However I am unable to start Postgresql service , getting below error.

Can someone please help with my understanding of PITR?  My questions are:

1. Since Barman is continuously receiving WAL files, should I be able to perform PITR to a timestamp after the base backup’s end_time (e.g., the next day)?

2. If yes, what might cause this checkpoint error? Missing WALs, misconfigured restore_command, or something else?
   

   Thanks in advance for any guidance!

[Jan Birk - gmail]

Den 2025-09-03 kl. 17.26 skrev sudha:

> The barman log says "Your PostgreSQL server has been successfully
> prepared for recovery!"
> However I am unable to start Postgresql service , getting below error

Is there anything in the log before the screendump with the tail ?

Ie:

ERROR: Connection problem with ssh
ERROR: Remote 'barman get-wal' command has failed!

or other ERRORs ?

/Jan

[sudha]

Just to clarify — are you asking about the PostgreSQL logs or the Barman logs?

On the PostgreSQL side, I don’t see any additional errors apart from the ones I shared in the screenshots.

On the Barman side, I also don’t see any errors in the logs — the PITR process reports success. However I have attached the screenshot of PITR command that reported success. 

I’ve retried the PITR again today, and I’m still facing the same issue when starting PostgreSQL.

[sudha]

Yes, you are right, I see below errors. what is this about? 

[Luca Ferrari]

On Thu, Sep 4, 2025 at 4:23 PM sudha <sudhaaru...@gmail.com> wrote:
>
> Yes, you are right, I see below errors. what is this about?

Could it be that your operating system user (postgres) is not able to
ssh-in the barman repository?

Could you please avoid sending images and copy and paste log text? It
makes it simpler to read.

Luca

[sudha]

Why would the target server’s postgres user need to be able to SSH into the Barman server?

Since the restore is initiated from the Barman server to the target server, I would expect the connection to go the other way — i.e. the Barman server should be able to SSH into the target server’s postgres user to copy the restored files.

[sudha]

Hi,
I am able to SSH both ways between my Barman server and the target PostgreSQL server using the postgres and barman users without any issues. However, when I run a restore, I still get the error:

ERROR: Connection problem with ssh