Freeing up memory from ArrayPools

[Petermarcu]

@okarlsson commented on Thu Apr 05 2018

Hey Guys!

I'm running a .NET Core application in a docker container that runs in a EC2 instance on AWS.

I have a function for resizing profile images before they get uploaded and I'm using a third party library called ImageSharp for this. The library uses ArrayPools when handling the images and my memory usage shoots up to about 1GB after handling a image.

But after the request has completed the memory usage never goes back, but is allocated to these memory pools.

Since I only have 1GB memory in my EC2 instance the entire instance crashes if I try to for example deploy a new version of my app when there is only 70MB memory available.

Is there any way to free up the memory allocated to the ArrayPools after I have finished a request so my containers don't run out of memory?

[RussKeldorph]

@jkotas @Maoni0 Related to #9737, etc.?

[RussKeldorph]

Assuming Future.

[benaadams]

@okarlsson I mentioned this to @JimBobSquarePants and he mentioned there is a article about memory use for ImageSharp; it also has fine grained control over its pooling with the ArrayPoolMemoryManager; including clearing the pools with ReleaseRetainedResources(), using smaller pools with CreateWithModeratePooling() and CreateWithMinimalPooling() or full control via the ArrayPoolMemoryManager constructor.

HTH

[okarlsson]

Awesome, thanks for the tip @benaadams

[antonfirsov]

@okarlsson also make sure to use the latest (beta-3) version of ImageSharp!

Based on your stack trace, you are using an earlier version, where the memory management is immature and sub-optimal!

[okarlsson]

Will do @antonfirsov , thanks!

[jkotas]

Related: https://twitter.com/ladeak87/status/1159092620171337728?s=12

[ghost]

Due to lack of recent activity, this issue has been marked as a candidate for backlog cleanup. It will be closed if no further activity occurs within 14 more days. Any new comment (by anyone, not necessarily the author) will undo this process.

This process is part of our issue cleanup automation.

[ghost]

This issue will now be closed since it had been marked no-recent-activity but received no further activity in the past 14 days. It is still possible to reopen or comment on the issue, but please note that the issue will be locked if it remains inactive for another 30 days.