Objectbox-java: question about logs

[DJafari]

i have too many this logs from my app, is it normal or something is wrong ?

W/Box: Skipping low-level close for read-only cursor (non-creator thread)
W/Box: Destroyed recycled transaction from non-owner thread

[greenrobot]

Do you know which code triggers this?

[DJafari]

@greenrobot unfortunately no, but i can find it, it's important ?

[greenrobot]

No hurry, but would be nice...

[garciagars]

I have this kind of log too

[Yazon2006]

I got it on Android 5.1 during the try to get 23167 entities from DB. Actually application freeze each time when I'm trying to do that. I can provide more details if it is necessary.

[greenrobot]

@Yazon2006 ObjectBox 1.4.x? Or at least 1.3.4? If so, I'd be happy to look into some details.

[DJafari]

@greenrobot for me in 1.4.x same logs exists

[greenrobot]

I assume this happens in short lived threads? Please call BoxStore.closeThreadResources() for Threads that are about to end. Let us know if that makes the logs go away.

Background: it's preferable when resources are closed in the same thread, which is not what happens with Java garbage collection.

[Yazon2006]

@greenrobot Yeah. It looks like during heavy transactions GC trying to clear memory and thats why transaction become recycled from non-owner thread. It seems to me that nothing to worry about cause object box still work as expected. Thank you for assisting.

[greenrobot]

We are thinking about a new method passing a Runnable to ObjectBox. Like you would do with a thread pool, but ObjectBox could clean up its resources automatically. What do you think?

[Yazon2006]

@greenrobot Cleaning up resources automatically it's a good idea. Improving memory consumption is always good)

[avently]

@greenrobot I see this logs often. I have ~30 threads that use one instance of box and boxStore. Threads lives all the time app is running. Can this log indicate about memory leak? If so, how to fix it?

[tatsuyuki25]

I get same logs, how to fix it?

[luffycheung]

it often occur when --> box.put() function operating frequency。

[miguelsalasalegre]

Still happening on 2.0 (Kotlin). The message "Skipped low-level close of cursor" appears every time you modify the database (put or delete, for example)

Example:

val box = store.boxFor(CLASS::class.java)
box.removeAll()
box.put(DATA)

Output:

[WARN ] Skipped low-level close of cursor (write, TX #3 alive)
[WARN ] Skipped low-level close of cursor (write, TX #4 alive)

[rudikiaz]

@greenrobot this warning is very worrying there is some fix already?

[greenrobot]

It should not be a big thing.

Please call BoxStore.closeThreadResources() for Threads that are about to end. Let us know if that makes the logs go away.

Have you tried this?

[rudikiaz]

We get that warning even in our tests which are all run in a single thread. We get this warning for every single put operation.

edit: we have tried closeThreadResources() but the logs keeps appearing

[chrbongardt]

Hi, we've got the same issue over here..

We have tried closeThreadResources() but it doesn't make much sense for us since we're running a single thread. We get the warning for every single put/delete operation. Logs get quite crowded with that warning. I'm wondering why this is happening.

@greenrobot Any more feedback on what could we try or do in order to help solving this issue??

[greenrobot]

Could you provide some simple code to reproduce what's happening on your side?

[Joeywp]

Got the same problem here in v2.1.0. Happens when I put a deep object with relations which have relations which have relations which..., but I don't have any reproducible code at the moment (can't share this project's code unfortunately). Most of the relations have backlinks and the put is happening in a thread from an executor. Also tried putting it in a TX call, but didn't change anything, except for the logs to be grouped together by "chatty" as the TX call keeps the TX ID the same. Will try to see if I can reproduce this in a standalone project, if so, I'll post it here, but so far I've been unable to reproduce it standalone.

Also, closeThreadResources() doesn't seem to do anything to the message

[rudikiaz]

Hello @greenrobot

Sorry for the waiting
Since i cant share the original source code, i created a simple proyect that reproduces this behaviour in the logs

Hope this helps: https://github.com/rudikiaz/ObjectboxLogsTest
As i wrote on the readme, i think its related to the indexs

[greenrobot-team]

Thanks @rudikiaz, could verify. This is also reproduced by our internal tests using indexes.

Cleaned up the Kotlin example to minimal code to reproduce:
196ObjectboxLogsTest.zip

Also reported internally to our development team.

-ut

[Gamezpedia]

whenever i use remove or put i am getting this message , i hope it is not a big deal , so far going good with it.

[greenrobot-team]

@Gamezpedia Your app should be fine. We are investigating this already. -ut

[guqh]

I also encounter this situation here, has been crazy to print this log, the program is frozen can not be used properly.

version：2.3.4
Android：5.1.0

[greenrobot]

@guqh This is probably amplified by the way you have been using Query (#753).

[greenrobot-team]

This was fixed internally and should be available with the next update. Thanks for all the reports.

[greenrobot]

This is also part of the just released version '2.4.0-RC' - would be great if you could give it a try and report back. Thank you!