Keyword search: 

2010 Releases

4/12/2014 5:49 PM
You can subscribe to this wiki article using an RSS feed reader.


FileReplicationPro 2010 Releases
  • Download the latest version for trial or upgrade - Here

  • First time install Quick Start Guide - Here 

  • Upgrade instructions - Here 




Release : 6.4.5   Date: Dec. 31, 2010

Operating System Affected : All    

Installation: Major Upgrade (New Installation) Instructions  Here



Version 6.4.5 fixes a few bugs



  • Fixed a bug in replication which was OSX specific - having to do with OSX forksl which caused a runtime error

  • A fix for long file names and acl problems on Windows platformFixes a problem with replicating whole Widows partition from the root folder.

  •  Fixed a bug in 2-way mirror replication when there was no access to folder. It could cause removal of all data from remote host. FRP  in this case thought that all files were removed on the source and was then deleting them on the remote server.

  • Fixed a condition in which live status events could cause out of memory error in the Mangement Server when there were too many events sent at one time from remote servers.  The transmission of these reports has been broken down so they are smaller.







Release : 6.4.1   Date: Dec. 1, 2010

Operating System Affected : All    

Installation: Major Upgrade (New Installation) Instructions  Here 




Major New Feature - Real time replication to and from all Linux OSes


Minor New Features


  • Now using FSEvents API in OSX (Leopard and Snow Leopard) - ensures all file system events are sent to FRP
  • A new s2s.rename_on_delete=true or false, when set to true delete operations in mirror type jobs are changed to rename, there is a suffix added to file or directory name of .$FRPdeleted; these files are ignored in further replications but can be restored at any time by renaming them without the suffix.
  • A new s2s.show_hidden_folders_in_browser=true  or false (false by default)  permits hidden folders to be seen in the job setup remote browser window during job setup.


Efficency Improvement / Fix


This Improvement is aimed at larger installs with more than 50 jobs and or replication servers.  It may be helpful if issues are encountered in other enviroments with less than 50 jobs and or servers.  Now default with this version, real time status messages are sent to the Management Console GUI only when a user opens the live log view for the job.  This reduces a large volume of status messages that were transmitted by all nodes to the management server sometimes causing issues with Management server crashes or overloads. 


Further,  the size and frequency of routine status updates to the management server can be configured. This is also controlled by the file.  The following entries


s2s.gran_file_size=                        (Example: s2s.gran_file_size=10  batches status transmissions to 10MB each)

s2s.gran_file_count=                      (Example: s2s.gran_file_count=100  batches status transmissions to 100 files each)

s2s.gran_folder_count=                 (Example: s2s.gran_folder_count=100  batches status transmissions to 100 folders)


Each of these entires respectively collect and batch transmissons of updates by every n number of MB of processed files (_size=10), or every n number of files (_count=100) processed or on every n number of folders processed. 




Fixed - Replication of large compressed files were corrupted
Fixed - Files\Folders with special characters were not replicated, or caused errors.
Fixed - Real Time was not working on windows when trying to replicate a whole partition from the root
Fixed - Problems with crashing on Mac - solved with updates to native libs used in OSX




Release : 6.3.18   Date: Sept 15, 2010
Operating System Affected : All     Installation: Major Upgrade (New Installation)





 Changes and Fixes for this version



  •  Added s2s property to permit showing of hidden folders in job browser.

  •  Posix lib update to improve replication behavior on various nix systems.

  •  Fix for Real Time on Mac not getting all the files that changed.

  •  Improved FRP internal realtime replication database management. 

  •  Improved FRP internal logging database cleaning routines - logging data was building improperly and causing the database to become very large causing instability in the Management server.

  •  Memory Leak fixed in the Management Server.

  •  Increased base memory available to the Management service

  •  Periodically logging info about used and free memory in the replication logs for use in judging memory needs.

  •  Pagination of pages in the management console to reduce fetch time and load.

  •  Tuning of administrative and logging traffic from the replication servers to the management server.  Reduced frequency and size of data transfers to reduce Management server loading.

  •  ACL problem fixed: when file did not contain ACL entries there were error reported: 'no such file'

  •  Events distribution: when there were two listeners listening to changes on same folder (or any from subfolders) only one of them got the events; this issue was related to OSX version

  •  A lot of exceptions were thrown when tracking JNI calls - exceptions were related to the way castor was serializing/deserializing java objects; this is a performance improvement

  •  Added option to follow symlinks or replicate symlink only to the advanced properties of every job through the GUI.

  •  Added the replication of Mac .rsrc files



 Detail and Functions of new parameters:


 Ability to see hidden folders in the folder browser when creating a job.  This change is needed only on the Management server, the user can create an entry on any empty line in the /FileReplicationPro/etc/ file using a plain text editor as follows:


 and the remote browser will show hidden files. Since this is GUI enhancement this property work only on MS - setting it on replication nodes is not needed.


 We have added pagination of jobs and servers set to default at 30 can be configured to more or less in the /FileReplicationPro/etc/   file.  We have hard coded the defaults at 30 per page and that value does not appear in the file at installation.  We recommend that you leave this unchanged until you have at least established that the management server is stable.


 The properties to control number of elements per page are: 






 To control number of logs per shown per page in the GUI there is property s2s.log.max=   Which is set in /FileReplicationPro/etc/ 


 By default it is set to limit log rows to 120

 It is set in /FileReplicationPro/etc/ 





 We have tuned transmission of logging information from remote servers to reduce size and frequency improving Management console behavior and decreasing network traffic.  The settings have been given hard coded defaults shown below that are deployed at installation.  Each replication server will have to be upgraded as well to get the benefit of these settings at the management server since each server transmits data about jobs, logs, and status back to the management server.


 All the default values for these settings are hard coded.  Only in the case of modification do these values have to be added to the /FileReplicationPro/etc/   file.  They do not appear in the default install.  Each is shown below with the default settings.  We DO NOT recommend you change these settings without consulting with us first.


  The values are in milliseconds.


  RS - Replication server

  MS - Management server

  DB - FRP's internal Data Bases which spools Real time requests and maintains the Logging data for the Management Server GUI display.


 s2s.freq.fetch_model=6000      - how often the Replication Server(RS)  invokes fetch model requests of the Management server (MS), default 60000

 s2s.freq.live_status=15000      - how often the live status information is sent to RS, default 15000 (if there is anything to sent)

 s2s.freq.status=5000              - how often the job status is sent, default 5000 (if there is anything to sent)

 s2s.freq.send_logs=3000         - how often the job logs are sent to MS, default 30000 - but the bigger number here the bigger request might be due to caching so it is best to go lower if there is any doubt.


 Important Factors:


 When RS-es are in debug mode they huge amounts of data with the status, so if stability is a factor it would be good idea to switch all RS-es to INFO or even WARN level Logging.   WARN level stresses the system least so it would be the choice if stability of the Management server is a concern.  The volume of data logging in DEBUG mode is greatest, then INFO, followed by WARN which is smallest.


The above noted size of the logs can be an issue in large installations because first they need to be transmitted from each of the RS servers to the MS via http request, then the MS inserts them to the FRP internal DB to spool them for writing to the txt logs. Secondly the DB stores log summaries for the GUI to display.  


Release : 6.3.0   Date: March 04, 2010
Operating System Affected : All     Installation: Major Upgrade (New Installation)

  • Servers can be sorted by Laptop or Non-Laptop (Mobile) status in server list of GUI
  • Extended logging
  • Online status fix

  • Reverse RPC fixes for roving / laptop machines

  • Source / Initiator server field fix in Job Log, now displays machine name instead of generated machine id.

  • Hidden files on Mac and UX as well as Windows OS no longer display in file browser


Release : 6.2   Date: Jan 07, 2010
Operating System Affected : All     Installation: Major Upgrade (New Installation)


  • Enhanced real time replication efficiency for MAC platform.
  • "Duplicate" and "Save As" Buttons in Job dialog which permit easy creation of many similar jobs.
  • Storage of multiple Management Server IP addresses - provides IP list for roving laptops to "call home".
  • System Settings window now permits the addition of more rows to define all possible Management Server IP addresses.
  • Fixed bug in MACOSX real time replication license checker.
  • Hiding of hidden (Win) folders and starting list with "." [dot] (non-windows OS) folders in the browser pop up for job creation / edit.
  • Replication servers will parse list of management server IPs defined in settings to access management server in addition previous method.
  • If a server should be added to the server list in error as a duplicate it will be marked as -dup for easy removal.
  • Fix to bug in Save Properties
  • Fix to model checker on save and throws exceptions when a corrupted model (Configuration) is being saved.


 INDEX Of All Releases 


 Knowledge Base Home


Home: Knowledge Base Index What's new: Recently changed articles