Domanda

I'm creating bower package search site (everything is open sourced) and I hit the wall. I have some memory leak (or I think I have) and I honestly don't know why it is there.

You can download it and run on Your own, but simple hint will help me greatly.

I have narrowed it down to this function call here https://github.com/kamilbiela/bowereggs-backend/blob/master/main.js#L14 ( nest.fetchAndSave() ) which is all defined here: https://github.com/kamilbiela/bowereggs-backend/blob/master/lib/nest.js

Basically it downloads a package list from internet, Json.parse and inserts it into database, plus some when.js promises.

Running this function few times creates a 30mb of memory per run, that is not cleaned by garbage collector. Also note that this is my first "real" node.js project, so I'll be really grateful for any tip.

È stato utile?

Soluzione

For anyone having the same problem:

https://github.com/c4milo/node-webkit-agent

After making few heap dumps I discovered that objects are garbage collected and the real memory usage isn't tied to it. I think that real memory usage is bigger because of using mongo and other non node.js stuff. Also real memory usage stabilizes at ~300mb, heap dumps at ~35mb.

Autorizzato sotto: CC-BY-SA insieme a attribuzione
Non affiliato a StackOverflow
scroll top