[Bugs] [Bug 1593826] Glusterfs being killed due to out of memory situation

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1593826



--- Comment #32 from hari gowtham <hgowtham at redhat.com> ---
(In reply to Amar Tumballi from comment #30)
> @hari, I see that the leak is definitely present in 3.12.x bits, but I am
> not sure why the patch you suggested above (about memcpy) is the reason for
> this. I removed the lines added by the patch
> https://review.gluster.org/#/c/glusterfs/+/16706/. I reverted only that
> patch in the branch, and still see the leak.

Hi Amar,

I came to know that the leak got in with the 3.12 version. Checked the patches
that went it 3.12 after 3.8 in this section. while inspecting the patches the
above patch does a memcpy for gfid which is a char.

It confirms the leak with

[protocol/client.gDBv1-client-0 - usage-type gf_common_mt_char memusage]        
size=311525526                                                                  
num_allocs=12586886                                                             
max_size=311534067                                                              
max_num_allocs=12587217                                                         
total_allocs=226411352

This portion was fixed with https://review.gluster.org/#/c/20437/ 

The portion for 

[protocol/client.gDBv1-client-2 - usage-type gf_common_mt_memdup memusage]      
size=151042232                                                                  
num_allocs=12586936                                                             
max_size=151042348                                                              
max_num_allocs=12587210                                                         
total_allocs=159380212

is yet to be identified.

This might be the reason for leak still seen without the patch.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.