nfs_inode_cache slab cache is very large
Issue
slabtop
shows the nfs_inode_cache slab cache is using a lot of memory:
$ slabtop -o
Active / Total Objects (% used) : 73324430 / 75887955 (96.6%)
Active / Total Slabs (% used) : 15671633 / 15671633 (100.0%)
Active / Total Caches (% used) : 110 / 159 (69.2%)
Active / Total Size (% used) : 48308948.60K / 48829118.25K (98.9%)
Minimum / Average / Maximum Object : 0.02K / 0.64K / 128.00K
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME
41510835 41443838 99% 1.01K 13836945 3 55347780K nfs_inode_cache
29346570 27616754 94% 0.21K 1630365 18 6521460K dentry_cache
3770454 3288427 87% 0.06K 63906 59 255624K size-64
450604 383965 85% 0.52K 64372 7 257488K radix_tree_node
386280 196885 50% 0.09K 9657 40 38628K buffer_head
215948 214254 99% 0.95K 53987 4 215948K ext4_inode_cache
129750 119012 91% 0.12K 4325 30 17300K size-128
14415 14298 99% 0.25K 961 15 3844K size-256
9064 8957 98% 0.09K 206 44 824K sysfs_dir_cache
6292 4102 65% 0.17K 286 22 1144K vm_area_struct
4032 1404 34% 0.02K 28 144 112K anon_vma
3930 3841 97% 0.58K 655 6 2620K proc_inode_cache
[...]
- How can we limit slab memory usage?
- I am concerned about the amount of memory consumed by slab. Can this impact available memory to my application?
Environment
- Redhat Enterprise Linux 5
- Redhat Enterprise Linux 6
- NFS
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.