retrieve_conf eating unlimited ram and doesnt release it after script end.
Using: CentOS 6 (node) + OVZ Debian 7 (vps with freepbx+asterisk)
slab cache growing after each config reload
dentry at slabtop has huge size.
Active / Total Objects (% used) : 3820317 / 3878206 (98.5%)
Active / Total Slabs (% used) : 231956 / 232058 (100.0%)
Active / Total Caches (% used) : 131 / 245 (53.5%)
Active / Total Size (% used) : 859619.62K / 866908.09K (99.2%)
Minimum / Average / Maximum Object : 0.02K / 0.22K / 4096.00K
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME
3455182 3455107 99% 0.22K 203246 17 812984K dentry
156288 133890 85% 0.10K 4224 37 16896K buffer_head
27874 27850 99% 0.55K 3982 7 15928K radix_tree_node
25718 19105 74% 0.05K 334 77 1336K anon_vma_chain
24864 23905 96% 0.03K 222 112 888K size-32
22876 20664 90% 0.20K 1204 19 4816K vm_area_struct
22698 22691 99% 1.09K 7566 3 30264K ext4_inode_cache
20119 13218 65% 0.06K 341 59 1364K size-64
18620 18549 99% 0.19K 931 20 3724K size-192
15700 15627 99% 0.15K 628 25 2512K sysfs_dir_cache
13013 8965 68% 0.05K 169 77 676K anon_vma
I have ~600 trunks and ringroups
dentry grow +20-30mb per reload. After 15-25 hours i have 25gb used ram, then i need reboot, or OOM kill killing other processes.
How it can be fixed?
Any ideas?