At this moment, you can set only 1 rule to make backups (daily, weekly, or monthly). My idea is to make multiple backup rules (ex. which daily backups should be deleted) or simply make more backup options.

Ex:
I want to make daily backup and delete every older than 2 days, except one day of the week and one day of the month.

ex interface:
Make [daily/weekly/monthly] backup and keep only:
[n] daily backups (enabled only for daily backups)
[n] weekly backups (enabled for daily and weekly backups)
[n] monthly backups (enabled for daily, weekly and monthly backups)

So, if someone picks daily backups from 2 days, but also wants a 1 monthly backup, the script makes daily backups and deletes every backup older than 2 days except 1 monthly backup. This way it doesn't make double work - backup for daily backups and backup for monthly backups.

This may also be a security issue.
Currently there are no special protection of user content in the backup. Plesk protects only its sensitive data, web content archived without any encryption.

If some gains ADMIN access to a wordpress/drupal/other web applications install parts of the Plesk container file system, cron jobs, etc will be compromised and would undoubtably increase exponentially the chances of the system to be exploited with virus, rootkits, malware “indirectly” affecting Plesk. (as an analogy no point to build a metal wall if the door is still glass)

So, in other words ... if a wordpress installation is compromised the installations of malicious (or not) plugins can lead to the installations of rootkit, malware and virus at the container. I cannot see for a second how this would be good for any Virtuozzo container running Plesk and how you could consider Plesk not being affect if for instance the container is used to conduct a DDOS to a random host.

Initial Thread
http://forum.odin.com/threads/plesk-12-x-x-backup-manager-encryption-security-issue.334691/#post-786917

As soon as any account is created in Plesk, a mail account is setup in Smartermail, used for managing the mail domain. This is one of the few minor issues for those using Plesk in a load balanced solution (which is actually remarkably possible!). It is also tricky for those running a limited smartermail license (e.g. 250 accounts, many of which are now unwanted Plesk management accounts).

It should not be so difficult to setup the mail account only when mail is enabled on a domain, and have Plesk just leave the mail server alone for domains that do not have mail enabled.

I wrote some more here:

http://forum.parallels.com/pda/index.php/t-265082.html

Hi,

Please consider implementing Mod Security v3.x.x (aka libmodsecurity) for NGINX and Apache in the next Plesk update.

At the moment any Plesk user if he wants to use Mod Security (official version supported by Plesk) is forced to use it as a web server:

• Apache

or

• Apache + NGINX

Any Plesk user who wants to use only NGINX as a web server and without using Apache at the moment cannot use Mod Security because Plesk does not currently support it for NGINX exclusively web servers.

Here are some of the advantages of Mod Security v3.x.x (aka libmodsecurity and these advantages apply to any type of webserver, apache, NGINX, IIS, etc., etc.,) compared to the now old and obsolete Mod Security v2.x.x as reported on the official GitHub page relating to SpyderLabs - Mod Security available at this link https://github.com/SpiderLabs/ModSecurity for anyone who wants to deepen the subject.

"Libmodsecurity is one component of the ModSecurity v3 project. The library codebase serves as an interface to ModSecurity Connectors taking in web traffic and applying traditional ModSecurity processing. In general, it provides the capability to load/interpret rules written in the ModSecurity SecRules format and apply them to HTTP content provided by your application via Connectors.

If you are looking for ModSecurity for Apache (aka ModSecurity v2.x), it is still under maintenence and available: here.

What is the difference between this project and the old ModSecurity (v2.x.x) ?

• All Apache dependences have been removed


• Higher performance


• New features


• New architecture

Libmodsecurity is a complete rewrite of the ModSecurity platform. When it was first devised the ModSecurity project started as just an Apache module. Over time the project has been extended, due to popular demand, to support other platforms including (but not limited to) Nginx and IIS. In order to provide for the growing demand for additional platform support, it has became necessary to remove the Apache dependencies underlying this project, making it more platform independent.

As a result of this goal we have rearchitected Libmodsecurity such that it is no longer dependent on the Apache web server (both at compilation and during runtime). One side effect of this is that across all platforms users can expect increased performance. Additionally, we have taken this opprotunity to lay the groundwork for some new features that users have been long seeking. For example we are looking to nativly support auditlogs in the JSON format, along with a host of other functionality in future versions.

It is no longer just a module.
The 'ModSecurity' branch no longer contains the traditional module logic (for Nginx, Apache, and IIS) that has traditionally been packaged all together. Instead, this branch only contains the library portion (libmodsecurity) for this project. This library is consumed by what we have termed 'Connectors' these connectors will interface with your webserver and provide the library with a common format that it undersands. Each of these connectors is maintained as a seperate GitHub project. For instance, the Nginx connector is supplied by the ModSecurity-nginx project (https://github.com/SpiderLabs/ModSecurity-nginx).

Keeping these connectors seperated allows each project to be have different release cycles, issues and development trees. Addtionally, it means that when you install ModSecurity v3 you only get exactly what you need, no extras you won't be using."

Thanks in advance for the support.

The settings for mail servers have a help tip in the email addresses management page, and in the users' Plesk Account pages.

If a Plesk server is set up strictly as a mail server, and has DNS handled remotely, the settings in the Help Tips can be incorrect.

Currently the tips only show the server as the root domain even if the mail server's actual DNS points (as an example) mail.domain.tld to the server, requiring the settings to include the subdomain.

If I add the domain as "subdomain.domain.tld" in settings, then the server tips are correct but the email addresses all contain the subdomain as well, which is not the desired behavior at all.

We need a way to add our own subdomain to the server's help tips listing, or have the server query external DNS and add it automatically to it's own setup help tips.