Description of problem: glusterfs process leaking in gf_common_mt_char Additional info: glusterfs not leaking any memory
> I've seen the number of total_allocs increasing in gf_common_mt_char area. Side by side comparison between the first and the the last statedumps: I comment came from the above statement. Total allocs increasing shows that the variable is being used. In general, if there is activity on filesystem, then there will be more increase in 'total_alloc' field. But regarding 'size=' value, we need to work. Not every translator or library utilize 'mt_char' to this high number. We will analyze and get back!
As per Comment#11, and with LRU limit option, this should go away.