Make OD publish jobs faster

VinodGoudK

Hi,

 

We are trying to make OD publish jobs faster, we have doubled our RAM and CPU allocation for teamsite server. But we did not noticed much improvement in the OD publish jobs.

 

How do we make OD app to use more RAM ( I mean increase heap size ), so that OD publish jobs gets faster.

 

Increasing the RAM, is this the right way to make OD publish jobs faster?

 

If not please guide me how can we make OD publish jobs faster.

 

Thanks you so much in advance.

 

Thanks
Vinod Goud 

Adam Stoller

I think you need to do a bit more detective work to determine where, if anywhere, the bottleneck(s) are in terms of the speed you are seeing now.  For instance:

• What type of deployment comparison are you doing?
  • source-target comparison
  • source-side area-based comparison
  • file list
• What type of deployment transfer are you doing?
  • single point-to-point
  • fan-out
  • tiered
  • routed
• How much data you are trying to push out?
  • number of files
  • smallest, largets, and average size of files
• Are you running DNRs?
  • source side and/or target side
  • before, during, and/or after deployment
• Source side server specifics
  • RAM
  • CPU
  • disk space available
• Target side server specifics (for all target's involved)
  • RAM
  • CPU
  • disk space available
• Network bandwidth 
  • general speed of network between source and target server(s)
  • actual speed / bandwidth availability between source and target server(s) when the deployment is being run
• How are the deployments being initiated?
  • via a CLT (which one)
  • via OpenDeploy WebServices
  • via the OpenDeploy AdminUI
• Any other potential factors, e.g.:
  • ComparisonRules
    • dateDifferent
    • revert
    • ignoreAcls
    • ignoreUser
    • ignoreGroup
    • checksumCompare
  • TransferRules
    • doDeletes
    • preserveAcls
    • svrTryCount
    • svrTryInterval
    • svrTryDisableOverwrite
    • rmReadOnly
    • compression
    • compressionLevel
    • applySourceFileTime
    • checksumVerify
    • byteIncremental
  • permissionRules
    • omask
    • directory
    • file
    • group
    • user
    • changeAccess
    • setAccess
  • etc.

Any or all of the above might play a role in detemining the speed of the deployment, so you need to determine what your specific parameters are and record (as accurately as possible) the various performance timings, e.g.:

• How long does it take to initiate the deployment
  • i.e.: the time between issuing the command and when the file transfer begins
• How long does it take to perform the deployment
  • i.e.: the time it takes to transfer all the files, the amount of data being transferred per second
• How long does it take to complete the deployment
  • i.e.: the time it takes to clean things up after the files have been transferred upto the time the overall command finishes and control is returned to the shell or calling program.

If doubling the RAM and CPU on the TeamSite server didn't help, then chances are you are looking at the wrong thing and/or in the wrong place.

 

If you're going to ask for more assistance on this, you also need to provide some of the metrics to explain the issue - e.g.: the number and size of files being transferred and the reported transfer rate (from the logs) and the perceived time it takes for performing the deployments - at the very least. Otherwise we have nothing to even use as a reference for our responses.