Hi.
My program : example ‘ceph -s’ often runs for a long time. I find alomost time use to compact_zone in my systemtap:
- coming compact_zone, pid 2331661 *
- 0xffffffff811a7ff0 : compact_zone+0x0/0x430 [kernel]*
- 0xffffffff811a8871 : try_to_compact_pages+0x121/0x1a0 [kernel]*
- 0xffffffff81682526 : __alloc_pages_direct_compact+0xac/0x196 [kernel]*
- 0xffffffff81682cb4 : __alloc_pages_slowpath+0x6a4/0x725 [kernel]*
- 0xffffffff8118b655 : __alloc_pages_nodemask+0x405/0x420 [kernel]*
- 0xffffffffa063d846 : ib_cmem_alloc_contiguous_pages+0x186/0x230 [ib_core]*
- 0xffffffffa094568c : mlx5_ib_exp_contig_mmap+0x3c/0xa0 [mlx5_ib] (inexact)*
- 0xffffffffa092d6e1 : mlx5_ib_mmap+0x51/0x4d0 [mlx5_ib] (inexact)*
- 0xffffffffa043409a : ib_uverbs_mmap+0x4a/0x80 [ib_uverbs] (inexact)*
- 0xffffffff811b7a3a : mmap_region+0x41a/0x640 [kernel] (inexact)*
- 0xffffffff811b7f65 : do_mmap_pgoff+0x305/0x3c0 [kernel] (inexact)*
- 0xffffffff811a081d : vm_mmap_pgoff+0xad/0xe0 [kernel] (inexact)*
- 0xffffffff811b63b1 : sys_mmap_pgoff+0x1f1/0x270 [kernel] (inexact)*
- 0xffffffff8102fb82 : SyS_mmap+0x22/0x30 [kernel] (inexact)*
- 0xffffffff81697809 : system_call_fastpath+0x16/0x1b [kernel] (inexact)*
Has anyone encountered this situation?
I have plenty of memory, so why occur memory compacting? can anyone help me
free -h
total used free shared buff/cache available
Mem: 187G 103G 13G 9.0M 70G 82G
Swap: 119G 15G 104G
Im in centos7.3.
linux kernel version
3.10.0-514.26.2.el7.x86_64
ofed driver is:
mlnx-ofa_kernel-4.1-OFED.4.1.1.0.2.1.gc22af88.rhel7u3.x86_64
libibverbs-41mlnx1-OFED.4.1.0.1.1.41102.x86_64
libmlx5-41mlnx1-OFED.4.1.0.1.5.41102.x86_64
my hca device attrs
hca_id: mlx5_3
transport: InfiniBand (0)
fw_ver: 14.30.1004
node_guid: 08c0:eb03:006e:5c2b
sys_image_guid: 08c0:eb03:006e:5c2a
vendor_id: 0x02c9
vendor_part_id: 4117
hw_ver: 0x0
board_id: MT_2420110034
phys_port_cnt: 1
Device ports:
port: 1
state: PORT_ACTIVE (4)
max_mtu: 4096 (5)
active_mtu: 1024 (3)
sm_lid: 0
port_lid: 0
port_lmc: 0x00
link_layer: Ethernet
thanks