Incremental backup not working

[LaurentR2D2]

Hello,

I have a scheduled backup operation for all the server. It should make incremental backups every day and a full backup once a week, but, instead, it makes a full backup every day. The incremental backup option is checked. The backup is scheduled to start at midnight every day. The backup is saved locally. I haven't yet configured a remote location.

My server :
CPU Intel(R) Xeon(R) CPU X3450 @ 2.67GHz (8 core(s))
Version Plesk Obsidian v18.0.21_build1800191128.17 os_CentOS 7
OS CentOS Linux 7.7.1908 (Core)

Thank you

 

[Bitpalast]

Interesting, I have observed the same issue on our Obsidian test machine, but thought that this is due to the very small size of the backup there. I thought that maybe the system calculates whether it is worthwile to do a delta of the files or wether the volume is so small that is is better to make a full backup instead. But maybe it is a bug.

 

[LaurentR2D2]

It must be a bug. I've just migrated my Plesk install on this server, so the backup size is the same, and I never had this problem with the previous server which was running Plesk Obsidian too.
Also, the previous server was less powerful and had less RAM (4Gb with 1 Gb swap partition), and I never had any Swap usage monitoring alarm as I have with this server which is shipped with 8 gb RAM and a 1 Gb swap partition. The backup process occupies nearly all the swap partition.

 

[IgorG]

It must be a bug.

Important - PLEASE FOLLOW THIS INSTRUCTION IF YOU BELIEVE THAT FOUND A BUG!

 

[Bitpalast]

Reported by @LaurentR2D2 in Forwarded to devs - Backup problem : incremental backup doesn't work and backup use all the swap

 

[zarza]

Hello,
Why this has been mark as solved if it is not solution on post? or in bug report?

Currently we have more of 50+ servers that after upgrade to Obsidian are not doing incrementals just full backup.

Can you upload a patch or a file to upload until you release a new update with its fix?

 

[Bitpalast]

I apologize for the confusion. The issue is considered "resolved" here, because it was reported to Plesk and is now being processed. There won't be any feedback in this thread once a micro update is published that truly resolves the issue in Plesk, but it is sure being taken care of. The "resolved" status here allows the Certified Plesk Professionals and Plesk Gurus, who frequently visit the forum and check the threads, to avoid useless effort in re-reading threads that have already been taken to a solution level.

So at the moment, there is nothing we can do here. Plesk will most likely find the bug and solve it in a future update. Such updates are published frequently.

 

[Pascal_Netenvie]

Yes same problem here.
All backups failed with these messages :

Warning:
Unable to find base backup for creation of incremental one, the full backup will be created. Error: Failed to exec pmm-ras: Transport error: unable to list directory: Curl error: (28) Timeout was reached: Last FTP request: LIST Last FTP response: 200 Type set to A
Error:
Unable to create remote backup. Error: Unable to export backup: Transport error: Unable to create the directory 'ftps://xxx.xxx.xxx.57/backup_flexcjd1_cda/': Curl error: (21) Quote command returned error: Last FTP request: QUIT Last FTP response: 221 Goodbye

 

[Bitpalast]

@Pascal_Netenvie: This is probably a different issue.
Please check this article and try the solution described there: Cannot configure FTP repository in Plesk Backup Manager: Transport error: unable to list directory: Curl error: Timeout was reached

 

[Pascal_Netenvie]

Thanks Peter but this KB seems to be for Centos and we use Debian.
And i don't see any backup or extra version for ftp_psa in /etc/xinetd.d/

 

[Bitpalast]

I suggest that you first try to create an FTP connection manually from your server's console to your backup space and check whether - once you are logged in - you can descend into the backup_flexcjd1_cda directory. Then check whether you can create and delete directories in your FTP space and whether you can transfer files into it. You should first test that, because the issue could simply be an issue with the FTP storage space.

 

[Pascal_Netenvie]

I successfully connect from source server to backup server with lftp and i successfully downloaded file, created and deleted a folder.
BUT i had to use command "set ssl:verify-certificate no" for this to work.
Whatever everything worked fine before Plesk 18 upgrade on source server ...

 

[Bitpalast]

According to the ftps-address in your log, you are not connecting to a domain name, but to an ip address. Normally, an SSL certificate needs a domain name, not an ip address. A certificate will never be valid when tested against an ip address, because it is made out to a domain name. Could the problem be that you rather need to connect to a server name instead of the IP off your FTP backup space? You could also try to switch to ftp instead of ftps (at least for testing purposes) in the GUI settings.

 

[Pascal_Netenvie]

Yes all backups (from many servers) connect to backup server with IP address.
But only one server have this problem and it is since Plesk upgrade to v18 ...
I will try with domain ... I'll keep you informed.

 

[Pascal_Netenvie]

It seems that it don't come from connection by IP address cause sometimes it works and sometimes no ...
See image : 12.17.2019-16.25.32.

 

[Pascal_Netenvie]

I just tested with domain name (secured with wildcard certificate) instead of IP address (to connect to backup server) and i still have problems.
5 tries : It worked 2 times and failed 3 times ...

Fail 1 & 3 > Full backup error :

 Unable to create remote backup. Error: Unable to export backup: Transport error: Unable to create the directory 'ftps://xxxxxx.com/xxxxxx/': Curl error: (21) Quote command returned error: Last FTP request: QUIT Last FTP response: 221 Goodbye

Fail 2 > Incremental error :

Unable to create remote backup. Error: Unable to export backup: Curl error: Unable to resume an interrupted upload: (28) Timeout was reached: Last FTP request: STOR xxxxx.com_1912181935.tar: Last FTP response: 200 Type set to I: Connection to the FTP server has lost

 

[Bitpalast]

"... an interrupted upload: (28) Timeout was reached: ..."
Clearly a network of FTP server issue. This is not an issue caused by the Plesk installation or your local server.

 

[Pascal_Netenvie]

Ok il will continue to search but problems started with Plesk upgrade from v17.5 to v18.0.
So it must have changed something ...