Multiple feeders using the same files are loaded as different iterators instead of one

Gatling (Open-Source)
1

I’ve been developing some load tests with Gatling and I could observe that, if I have different scenarios being feeded with the same csv file, for example

val scn =
feed(csv(“users.csv”).random)
.exec( …

each feeder is loaded into memory and, having 10MB csv files loaded multiple times in memory it turns into a potential OOM.
Is there any plan to change the way this works or should I take another approach to solve my memory issue?

Thanks!

2

So why don’t you load it only once if memory is an issue for you?

val myFeeder = csv(“users.csv”).random

val scn =
feed(myFeeder)
.exec( …

3

But in this way, all the user will get the same value. But we do want to have random values for different users…

4

if I have different scenarios being feeded with the same csv file

But in this way, all the user will get the same value.

No. With my solution, all the users will use the same data pool, not the same value.