1451250 - Firefox crashes with OOM | large | NS_ABORT_OOM | mozilla::cyclecollector::HoldJSObjectsImpl ]

Status: RESOLVED DUPLICATE of bug 1216175

(Core :: DOM: Workers, defect, P2)

Description

[Ramnik]

Attachment: Capture5.jpg

User Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/65.0.3325.181 Safari/537.36

Steps to reproduce:

1. Opened one of my web (KVM) application.
2. Kept the windows idle & after some time the window crashes. I could reproduce it several times.


Actual results:

1. I monitored the resource utilisation & could see that as soon as I start the KVM application the physical memory allocation keeps on increasing & once the memory is full, the window crashes.
2. I tried working with the same KVM application on Chrome & IE & it worked well as expected without any crashes.


Expected results:

1. The window should have opened correctly & should not have crashed
Seems there is some memory leakage.

Comment 1

[[:philipp]]

a particular crash report resulting from this memory leak according was bp-43df80b4-1dad-4fa2-8530-d34ef1180329 according to https://support.mozilla.org/questions/1211441

Comment 2

[Ramnik]

https://crash-stats.mozilla.com/report/index/90fad19d-e711-467e-81ef-692840180404 
https://crash-stats.mozilla.com/report/index/bp-a2cddab1-e7b3-4145-bbcb-3e6730180404
https://crash-stats.mozilla.com/report/index/bp-00b56486-8434-4f21-9271-184a00180404
https://crash-stats.mozilla.com/report/index/bp-54d1ba4b-5efc-4d7b-9f5b-f5a270180404
all these are the latest(generated today i.e 4th April 2018) crash reports along with the above-mentioned report.

Comment 3

[Andrew McCreight [:mccr8]]

The OOM allocation size is 3.12 GB, which is quite a lot! The crash stacks are all under WorkerGlobalScopeBinding::setTimeout. Baku, could you take a look? Maybe this is a case where the worker is getting starved so the timeouts are being created but never used up.

Comment 4

[Ramnik]

Any updates on this? Baku??

Comment 5

[Marion Daly [:mdaly]]

This is next in our priority queue. We'll take a look shortly.

Comment 6

[Andrea Marchesini [:baku]]

Ramnik, I need some extra information in order to reproduce the issue.
Do you have a testcase or a webpage I can use?
Thanks!

Comment 7

[Ramnik]

I will not be able to share the webpage because of the policies. What I can do is capture the data you require. You may let me know the test I should carry out & ways to capture the required data. Briefly, it's a Remote KVM application with complete console redirection of one of my systems. Looks like some video sockets are in use.

Comment 8

[Ramnik]

I see some recent comments on Bug 1216175. Are these similar? Could same be tracked together?

Comment 9

[Andrea Marchesini [:baku]]

Yes, it looks the same bug. Thanks!