• Replies 2 replies
  • Subscribers 137 subscribers
  • Views 952 views
  • Users 0 members are here
Options
Related

RE: Publication in SDL

PaulGregory Raj Viyagulam Brema

Hi,

We have a requirement to publish almost 9000 Publication objects, but SDL is able to publish maximum three in parallel.

We don't how long would it take for 9000 Publication objects to complete the publication. We are using API to trigger the Publication.

Is there any way we can speed up this process. Help is much appreciated.

 

 

Thanks,

PaulGregory
Rate translation
×
Suggest better translation
×
Parents

  • Hi PaulGregory,

    Based on "maximum three in parallel" I'm guessing you are on version 2014/11.0.x or earlier... so based on MSMQ/InfoShareConfig.xml

    The default configuration of 3 is to protect your system, it is not about limiting you. If I ask you how long it would take to copy 9000 files, you would ask me how big those files are... so how big are your publications (1000 topics per pub, or 20000 topic per pub)? Are they relying on DITA-OT, or other transformation engines?

    What is the bottleneck you envision (IO, memory, RAM,...). How is your system currently behaving - underused or already under load?

    So unfortunately the only generic answer to offer is

    • If your current backend/batch server or servers have the processing room, you can increase the value from 3 to ... 5...8?
    • r you could add a couple of extra backend/batch servers so you get more throughput.

    Hope this helps,
    Dave
Reply

  • Hi PaulGregory,

    Based on "maximum three in parallel" I'm guessing you are on version 2014/11.0.x or earlier... so based on MSMQ/InfoShareConfig.xml

    The default configuration of 3 is to protect your system, it is not about limiting you. If I ask you how long it would take to copy 9000 files, you would ask me how big those files are... so how big are your publications (1000 topics per pub, or 20000 topic per pub)? Are they relying on DITA-OT, or other transformation engines?

    What is the bottleneck you envision (IO, memory, RAM,...). How is your system currently behaving - underused or already under load?

    So unfortunately the only generic answer to offer is

    • If your current backend/batch server or servers have the processing room, you can increase the value from 3 to ... 5...8?
    • r you could add a couple of extra backend/batch servers so you get more throughput.

    Hope this helps,
    Dave
Children
No Data