Memory Leak on 3.8?

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.
7 năm cách đây
Hi,

I've recently updated my website to version 3.8. When I start my website with separate app pool, it uses about 350 MB of memory. That's great and should fluctuate around that number like it did with version 3.4. But in my case the memory raises to +2500 MB (2.5 GB) in 4 hours time, and I'm getting timeout and out of memory errors. I've did a small test and see that Automapper calls/instances is on the top (dotMemory trial). Could It be that Automapper is the cause for the memory raise/leak?

Hardware:

Use VPS - Windows Server 2012 - 64bit
Memory: 16 GB
CPU: 4 x 2.20GHz
7 năm cách đây
Hi,

I don't think it's AutoMapper. But most probably it's related to this forum topic
7 năm cách đây
a.m. wrote:
Hi,

I don't think it's AutoMapper. But most probably it's related to this forum topic


I've already replaced the browsercap file, but without any improvements.My problem is not at startup. Startup is good. But then my memory progressively raises until my site craches.
7 năm cách đây
Check out my comments on this thread.

Because your memory progressively builds even after startup, and we don't think there are any real memory leaks in 3.8 (after this many months, others likely would have noticed), it could be Nops cache engine.  You will see in the thread mentioned, if your not running captcha, blog comments, etc. are vulnerable to easy injection, and these tables get cached in the app pool/w3wp.
So in some cases, you could have 300,000 records in the blog comments and after about 30 to 40 mins your app pool would be about 2+ Gigs (was my experience).

Anyway, hopefully this pinpoints your issue.

Thanks
Chuck
7 năm cách đây
After a long investigation I could conclude that AutoMapper 5.0.2 was a part of the problem. On a test case I've replaced AutoMapper with ExpressMapper. The results where immediately visible and the memory of my app pool stays within normal limits. It now uses +- 400 MB instead of 2500 MB. What a difference. I found some issues on AutoMapper concerning memory, but don't know if that was the problem. This is why I decided to replace AutoMapper and test the new config.

https://github.com/AutoMapper/AutoMapper/issues/1691
https://github.com/AutoMapper/AutoMapper/issues/1759
7 năm cách đây
Nesse wrote:
After a long investigation I could conclude that AutoMapper 5.0.2 was a part of the problem. On a test case I've replaced AutoMapper with ExpressMapper. The results where immediately visible and the memory of my app pool stays within normal limits. It now uses +- 400 MB instead of 2500 MB. What a difference. I found some issues on AutoMapper concerning memory, but don't know if that was the problem. This is why I decided to replace AutoMapper and test the new config.

https://github.com/AutoMapper/AutoMapper/issues/1691
https://github.com/AutoMapper/AutoMapper/issues/1759

Hi Wesley,

Thanks a lot! I haven't played with ExpressMapper yet. But the main concern is that according to its GitHub page it's not actively developed (compare with AutoMapper commit history)

P.S. I've just re-open a work item. But it's still with "on hold / maybe wont" status
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.