Question

Limiting impact upon data source

  • 28 November 2017
  • 2 replies
  • 0 views

Badge

Hi,

I have a production server that is mission critical and I cannot impact performance. Unfortunately there is no mirroring or a secondary server I can access and it's hitting the production server or nothing. In reading from this server, I cannot impact that servers performance in any manner...fortunately I can schedule this in the evening when it's load is minimal, however I cannot impact the performance of this server for that minimal load.

Does FME have any methods of 'running in the background' where it reads from the server only when it's resources are free (or minimally used)? Or diffrent wording...Is there any method I can use to ensure FME doesn't impact production performance while extracting data from a production server?

MSSQL 2008 (source) to a Oracle 12c DW if that makes any difference for this question.

Thanks!


2 replies

Userlevel 2
Badge +12

I would try the extraction without any data conversion, so the reading would be the quickest.

You can do that reading and then writing (as is) to the FFS format (FME's internal format).

Then you can use the FFS files as a source, instead of the database.

You can schedule the FFS extraction during the hours the database is used least.

Hope this helps.

Userlevel 4
Badge +25

I've been thinking about this for a bit and I think if you follow the rules to the letter you can't do anything.

As soon as you read from that server you're impacting its performance. The fraction of a millisecond it takes it to return a feature to you it could be returning a feature to somebody else.

So the short answer is: no.

Having said that, the suggestion that @erik_jan makes is probably the best one to do given these circurmstances. The FME Desktop Advanced Training has some sections on performance that are worth watching/reading.

Reply