rrServer and internal changes

rrServer

Local jobs.db and log file
The rrServer saves its app log file and jobs.db to the local drive.
A second background app syncs the files to the fileserver.
Improves responsibility of the rrServer in case of network hickups.
Job files as local memory cache file
Instead of saving the jobs every 30 minutes, you can enable a memory cache file.
The rrServer saves the jobs uncompressed to the local drive once they are received.
Fast SSD recommended.
Cache/Synced fileserver support
A new settings to tell the rrServer to wait for files after submission and after a client has finished its frame segment.
No RR Client licenses for unused machines.
Clients that are offline for a few days will be internally removed from the RR license assignment.

TCP data encoding improved
New password hashing "Argon2"
Hash-Only login (Ability to save a password hash in your local user prefs or a file to use in python scripts)
Better encoding of preview jpegs in rrJobData (if enabled)
Optional Job password set at submission to change job later (for automated submission)
Job data send from rrServer to rrClient is encoded

These new RR9 features are marked in green:

Royal Render Priority Level
Priority Level	1	2	3	4+5	6	7
	Client Assignment (set by user, group, installed apps or hardware)	Exclusive	Client privileged for	Client Averaging/Distribution modes (Average= all get the same amount of clients)	Render Order	Job Setting "Priority" (+ rrConfig modification )
Available values	Assigned Deassigned	Any of these types or a combination: Project Render app Renderer User Division and Client Count and Expire date	Any of these types: Project Render app Renderer User Division and "Abort other jobs"	Average projects Average apps Average users	Global list to order priority based on Project Render app Renderer User Division	1 2 3 4 ... 98 99 First-In First-Out Average (same prio => same client count)
Examples	Redshift jobs are assigned to clients with GPU only. Houdini jobs to machines with high RAM. Nuke jobs to machines which	In this example "average clients" is enabled. All projects get same ammount of clients. But projectXY has a deadline soon, so you assign a minimum of 50 clients to that project until end of the month. And Nuke is important, so 10 Clients for Nuke all the time.	If a Houdini sim job is send, it should render on the machines with a lot of RAM.	All projects in the company get the same part of the farm to end "priority wars".	App "Houdini" has a higher priority than division "Turntable". Next are "Maya" jobs. At last jobs of user "RobotAutomation"