Knowledge Base

Plesk Obsidian 18.0.41 Backup process can be stuck indefinitely when uploading to remote FTP storage

Symptoms

On a server that runs Plesk Obsidian version 18.0.41, the backup process upload to remote FTP can be stuck indefinitely (since Plesk update to 18.0.41), while it keeps spawning a lot of backupmng processes
Messages that are similar to the following can be found in the backup logs at /var/log/plesk/PMM/backup-YYYY-MM-DD-hh-mm-ss-ms:

INFO: TransportError Transport error: unable to check directory existence: Curl error: (7) Couldn't connect to server: Last FTP request: Last FTP response: [common/plesk-utils/PMM/repository-transport/transport.cpp:TransportError]
virtual bool plesk::tRepositoryFtp::IsDirExist(const string&)
INFO: pmm-ras finished. Exit code: 125
INFO: The utility succesfully executed.

INFO: Curl error: (28) Timeout was reached: Last FTP request: STOR backup_2202161157.tar: Last FTP response: 150 Opening BINARY mode data connection for backup_2202161157.tar

INFO: Unable to resume upload. Have got 550 error code from FTP server
INFO: Cancel export
INFO: Data sending interrupted, because of lost connection to FTP
INFO: Error occured while writing archive to FTP Curl error: Unable to resume an interrupted upload: (25) Upload failed (at start/before it took off): Last FTP request: STOR backup_2202200441.tar: Last FTP response: 550 backup_2202200441.tar: Disc quota exceeded: Connection to the FTP server has lost

Cause

This behavior was caused by a product issue:

#PPPM-13411 "Scheduled backup processes no longer get stuck indefinitely when backing up to FTP storage if the configured FTP server is unavailable or if the provided credentials do not match."
Fixed in:
- Plesk Obsidian 18.0.42 - 28 February 2022

Resolution

Since the bug has been resolved permanently when Plesk Obisdian 18.0.42 came out, you should simply update your Plesk version to the latest one, as explained in the following article:

How to update Plesk Obsidian to the latest build?

Workaround

If updating Plesk to the latest build is not possible for some reason, you may execute the following workaround:

Connect to the Plesk server via SSH

Stop the cron service

On Debian/Ubuntu:

# systemctl stop cron

On CentOS/RHEL/CloudLinux:

# systemctl stop crond
Kill the stuck backup processes by using their PID:

# kill -9 `ps auxfw | grep -i pmm | grep -v grep | awk {'print $2'}`

# kill -9 `ps auxfw | grep -i backup | grep -v grep | awk {'print $2'}`
Start the cron service

On Debian/Ubuntu:

# systemctl stop cron

On CentOS/RHEL/CloudLinux:

# systemctl stop crond
Either wait until the next scheduled backup is launched or launch the process manually through the following command:

# /opt/psa/admin/sbin/backupmng

Warning: If the tasks are launched manually through the command above, all the pending scheduled backups will be launched at the same time. Consider this for resource usage.