[SOLVED] Push jobs take much longer now

[leesteken]

I just noticed that my daily push jobs of encrypted backups from one PBS (4.2.4) to others take hours instead of minutes. One of which is at Tuxis (4.2.0) and not under my control. None of the PBS have the encryption keys and they don't need them as the backups are client-side encrypted.
Was there a recent change that all data is sent instead of early detection of duplicate data? One of the backups is a host (using proxmxox-backup-client) of a 100GB of which 98% was reusable before. Am I the only one experiencing this with Push jobs? Apologies if I missed an obvious thread about this.
I did not recognize this change in the release notes but I might have missed them, and I don't know how to read them except just before applying updates.
Is there any way I can get the optimized/reduced network traffic back again?

 

[fiona]

Hi,
thanks for the report! The regression is known and fixed in current master already: https://bugzilla.proxmox.com/show_bug.cgi?id=7878

 

[Onslow]

I did not recognize this change in the release notes but I might have missed them, and I don't know how to read them except just before applying updates.

Before installing updated packages I do:

cd /var/cache/apt/archives
apt-get --download-only dist-upgrade
find . -name "*.deb" -ctime -20 -print0 | xargs -r0 apt-listchanges

(This -20 is with a big room for error, but it doesn't make an information noise, as next executions of this command don't display old changes of already installed packages).

Then after having a look at the changes I apply downloaded packages in a convenient moment - with usual
apt-get dist-upgrade

 

[leesteken]

thanks for the report! The regression is known and fixed in current master already: https://bugzilla.proxmox.com/show_bug.cgi?id=7878

And the fix reach no-subscription already and resolved the issue completely.