Robomongo: Memory leaks

[igal1c0de4n]

I would like to report a massive memory leak in Robo3T version 1.2.1. It has been there for as long as I remember, since early versions of Robomongo.

What type of information I can provide?

For now, I can share I see in my system's monitor (macos high sierra) that Robo3T grows from a hundred megs or so at start, to 5GB and (sometimes 10GB) of system memory after a remaining open for a few hours (day?)

I usually quit and relaunch it, but that's inefficient in terms of dev flow since I have no easy way to save the queries and results in the various open tabs.

How can I help to solve it? Is there additional debug info or a Robo3T log I can provide?
Please advise

[danieler1981]

15GB of RAM for me, really hard to use Robo 3T.

[simsekgokhan]

Hi all, thanks a lot for reporting the problem. It seems like a really big problem.

Can you guys tell us how to reproduce this problem? For us the RAM usage is around 70-80 MB. We have never experienced such usage in our labs.

You can share:

• Logs on Robo main window
• DB details (db version), is it Atlas or local, replica set etc...
• Connection settings screenshot

@danieler1981
What is your OS version?

And finally you can try:
Robo 3T 1.3 Beta

[igal1c0de4n]

I can share I get the leaks when having secure connections to a local mongod and/or to a deployment hosted by Compose. Both are using version 3.2.18.

@simsekgokhan are you suggesting we try the beta because you are aware of the issue and 1.3 is expected to address it?

[simsekgokhan]

@igal1c0de4n Thanks a lot for details.
In 1.3, there is no direct change related to memory usage since we have almost no other issues about it and also we have never seen it in our dev environment tests. The reason I suggested is that in 1.3 the MongoDB drivers are updated, so if this problem is from MongoDB code, it can potentially solve your problem.

By secure you mean SSL connection?
So this is how we can reproduce your problem as I see:
Robo 1.2.1, Macos high sierra, SSL, local mongod with version 3.2.18

[igal1c0de4n]

By secure you mean SSL connection?

yes

So this is how we can reproduce your problem as I see:
Robo 1.2.1, Macos high sierra, SSL, local mongod with version 3.2.18

yes. Idk if related, but typically I have connections to more than one db. For example, I can have one connection to the local deployment, and a couple of other connections to the staging and production deployments. IDK if that's useful info. The number of connected deployments may affect the ability to repro

[danieler1981]

I have OSX, usually the problem came after 2/3 hours of working with different connections. LocalDB, remote Database in Atlas. Just try to query or make aggregation massive usage.

[piushkumar]

will it be fixed in the 1.4 ? I am facing the memleaks of size 22GB