About long publishing operations

Build 1501 on 14/Nov/2017  This topic last edited on: 28/Oct/2016, at 11:47

When multiple operation on the same publishing destination are scheduled to be executed in background, Back4 uses an internal locking system to execute them serially even when there are multiple processing threads. These locks have a configurable time-out set by SynchTimeOut parameter in Back4Config:

<Back4Config

  xmlns="http://www.teradp.com/schemas/GN4/1/Back4Config.xsd"

  PollingInterval="15000"

  TaskDuration="0"

  NProcThreads="1"

  SynchTimeOut="00:05:00"

  WebUrlPrefix="http://*:7373/"

  LogMaxTime="168"

  MaxCachedLogins="100"

  NProcThreadsHighPriority="0"

  NMinTaskPriorityForHighPriorityThread="5"

  >

  . . . 

if a publishing operation lasts more than the specified time-out other publishing operations waiting for it will fail. The default for this time-out is 5 minutes.

Note that this means that if Back4 executes mostly publishing operations on the same publishing destination the number of processing threads does not really matter: the operations will be executed serially in any case.

Note also that this synchronization system does not work (obviously) if there are multiple Back4 connected to the same database: each Back4 instance will attempt independently to execute the publishing commands - possibly resulting in check-out conflicts.

See also

About publishing process