| Re: Kernel cache dentry leak? [message #45014 is a reply to message #44943] |
Fri, 20 January 2012 19:10   |
insider
Messages: 11
Registered: January 2012
|
Junior Member |
|
|
After a last manual cache clear with echo 2 >/proc/sys/vm/drop_caches there a 2 days passed, "dentry" now holds 15174252 objects and uses 3372056K and keeps incressing...
slabtop command information:
Active / Total Objects (% used) : 15292469 / 15303649 (99.9%)
Active / Total Slabs (% used) : 851679 / 851681 (100.0%)
Active / Total Caches (% used) : 122 / 240 (50.8%)
Active / Total Size (% used) : 3232528.04K / 3234571.11K (99.9%)
Minimum / Average / Maximum Object : 0.02K / 0.21K / 4096.00K
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME
15174252 15174204 99% 0.21K 843014 18 3372056K dentry <<======!!!!!
24790 24759 99% 0.10K 670 37 2680K buffer_head
20048 19762 98% 0.03K 179 112 716K size-32
12528 12373 98% 0.08K 261 48 1044K sysfs_dir_cache
10384 9963 95% 0.06K 176 59 704K size-64
6816 6777 99% 0.62K 1136 6 4544K inode_cache
4928 3144 63% 0.05K 64 77 256K anon_vma_chain
4921 4204 85% 0.20K 259 19 1036K vm_area_struct
4500 4425 98% 0.12K 150 30 600K size-128
3899 3866 99% 0.55K 557 7 2228K radix_tree_node
3411 3404 99% 1.05K 1137 3 4548K ext4_inode_cache
3372 3354 99% 0.83K 843 4 3372K ext3_inode_cache
3205 3184 99% 0.68K 641 5 2564K proc_inode_cache
2862 2694 94% 0.07K 54 53 216K Acpi-Operand
2695 1937 71% 0.05K 35 77 140K anon_vma
1940 1116 57% 0.19K 97 20 388K cred_jar
1740 1714 98% 1.00K 435 4 1740K size-1024
1620 1583 97% 0.19K 81 20 324K size-192
1440 1055 73% 0.25K 96 15 384K filp
1380 1332 96% 0.04K 15 92 60K Acpi-Namespace
1376 1286 93% 0.50K 172 8 688K size-512
945 904 95% 0.84K 105 9 840K shmem_inode_cache
612 571 93% 2.00K 306 2 1224K size-2048
540 345 63% 0.25K 36 15 144K size-256
510 240 47% 0.11K 15 34 60K task_delay_info
468 287 61% 0.31K 39 12 156K skbuff_head_cache
424 56 13% 0.06K 8 53 32K fs_cache
420 222 52% 0.12K 14 30 56K pid
308 298 96% 0.53K 44 7 176K idr_layer_cache
288 231 80% 1.00K 72 4 288K signal_cache
288 29 10% 0.08K 6 48 24K blkdev_ioc
288 256 88% 0.02K 2 144 8K dm_target_io
280 239 85% 0.19K 14 20 56K kmem_cache
280 54 19% 0.13K 10 28 40K cfq_io_context
276 62 22% 0.03K 3 92 12K size-32(UBC)
276 256 92% 0.04K 3 92 12K dm_io
270 233 86% 2.06K 90 3 720K sighand_cache
260 238 91% 2.75K 130 2 1040K task_struct
242 242 100% 4.00K 242 1 968K size-4096
240 182 75% 0.75K 48 5 192K sock_inode_cache
202 2 0% 0.02K 1 202 4K jbd2_revoke_table
202 4 1% 0.02K 1 202 4K revoke_table
187 54 28% 0.69K 17 11 136K files_cache
168 56 33% 0.27K 12 14 48K cfq_queue
162 55 33% 0.81K 18 9 144K task_xstate
159 18 11% 0.06K 3 53 12K size-64(UBC)
153 99 64% 0.81K 17 9 136K UNIX
144 32 22% 0.02K 1 144 4K jbd2_journal_handle
124 74 59% 1.00K 31 4 124K size-1024(UBC)
120 18 15% 0.19K 6 20 24K bio-0
120 45 37% 0.12K 4 30 16K inotify_inode_mark_entry
Is there a way to dump contents of dentry to a file, maybe to inspect and investigate, what this cache contains?
I have tried with "dd" copy from /dev/mem to a files, but it not allows to dump full kernel memory...
Maybe this problem is related to a filesystem?
We have mounted these filesystems:
/dev/md2 on / type ext4 (rw)
proc on /proc type proc (rw)
none on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/md0 on /boot type ext3 (rw)
/dev/mapper/vg0-vz on /vz type ext3 (rw)
none on /proc/sys/fs/binfmt_misc type binfmt_misc (rw)
beancounter on /proc/vz/beancounter type cgroup (rw,name=beancounter)
container on /proc/vz/container type cgroup (rw,name=container)
fairsched on /proc/vz/fairsched type cgroup (rw,name=fairsched)
Searching for a solution or a way to investigate this problem, but still unsuccessful...
Upgraded kernel to 2.6.32-042stab044.17 #1 SMP Fri Jan 13 12:53:58 MSK 2012 x86_64 x86_64 x86_64 GNU/Linux, but this not solved dentry leak problem.
Any thoughts?
Does nobody else got this problem with RHEL 6 63bit 2.6.32 ?
[Updated on: Fri, 20 January 2012 21:23] Report message to a moderator |
|
|
|
OpenVZ Forum
Members
Search
Help
Register
Login
Home
