out of memory in getImageData

[goldenjayr]

I have a nodejs application that transforms images. I run my application with 32gb of ram. When trying to get the image data of a large image with more or less has a width of 20k and height of 20k I am getting this error. I have tried to increase the node js memory limit using --max-old-space-size tried to increase it to 8192 as well as 32678 but it I still get an error.

m[9161:0x41069d0] 18179 ms: Mark-sweep 22.3 (29.2) -> 22.0 (25.2) MB, 23.0 / 0.0 ms (+ 1.5 ms in 5 steps since start of marking, biggest step 0.8 ms, walltime since start of marking 55 ms) (average mu = 0.998, current mu = 0.963) finalize incremental m[9161:0x41069d0] 27013 ms: Mark-sweep 23.2 (25.2) -> 22.2 (25.2) MB, 3.8 / 0.0 ms (+ 0.1 ms in 2 steps since start of marking, biggest step 0.1 ms, walltime since start of marking 168 ms) (average mu = 0.999, current mu = 1.000) finalize incremental m

`<--- JS stacktrace --->

==== JS stack trace =========================================

Security context: 0x21639e1c08d1
0: builtin exit frame: getImageData(this=0x0529133cc801 ,32150,22958,0,0,0x0529133cc801 )

1: getPoints [0x324589ad9b31] [/var/apps/print-export-worker/getPoints.js:11] [bytecode=0x340f496a0d29 offset=86](this=0x2d04988cd319 <JSGlobal Object>,0x0529133cc801 <CanvasRenderingContext2D map = 0x381c...

FATAL ERROR: v8::ArrayBuffer::New Allocation failed - process out of memory
1: 0xa09830 node::Abort() [node]
2: 0xa09c55 node::OnFatalError(char const*, char const*) [node]
3: 0xb7d71e v8::Utils::ReportOOMFailure(v8::internal::Isolate*, char const*, bool) [node]
4: 0xb7da99 v8::internal::V8::FatalProcessOutOfMemory(v8::internal::Isolate*, char const*, bool) [node]
5: 0xb7db7b [node]
6: 0xb9b93e [node]
7: 0x7f7f87de455f Context2d::GetImageData(Nan::FunctionCallbackInfov8::Value const&) [/var/apps/print-export-worker/node_modules/canvas/build/Release/canvas.node]
8: 0x7f7f87dd91f9 [/var/apps/print-export-worker/node_modules/canvas/build/Release/canvas.node]
9: 0xbe96a9 [node]
10: 0xbeb497 v8::internal::Builtin_HandleApiCall(int, unsigned long*, v8::internal::Isolate*) [node]
11: 0x13c5c79 [node]
Aborted`

I think that it's the limit of V8 to store an array bigger than its limit but how do I fix this problem.

[zbjornson]

ArrayBuffers are allocated off of the heap, so setting --max-old-space-size will have no impact. (Try node --max-old-space-size=50 -e "new Uint8Array(550000000).fill(1)" to see for yourself.) Are you calling getImageData in a loop? If so, there's like something retaining those buffers causing a leak.