Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 25 Dec 2013 22:02:42 +1100
From:      Mark van der Meulen <mark@fivenynes.com>
To:        "freebsd-net@freebsd.org" <freebsd-net@freebsd.org>, freebsd-hackers@freebsd.org
Subject:   vmstat buckets
Message-ID:  <4260012E-9911-47AB-88B9-FDCD6748E6EE@fivenynes.com>

next in thread | raw e-mail | index | archive | help
Hi All,

I have question about buckets in vmstat -z output which I haven=92t been =
able to find an answer for online=85

I have 6 FreeBSD 9 boxes running as routers and they are seeing varying =
results in vmstat -z. I=92m interested in understanding what the Buckets =
represent/mean and how I can influence usage of them.

Some of the routers show 0 free buckets under 64 or 128 Bucket and so I =
did some research and haven=92t been able to find out what it means - =
most people who have asked online have received an answer that it is =
nothing to worry about, which is meaningless to me.

It seems that the higher the usage of the router, especially when it is =
also running userspace applications the less Bucket availability there =
is. Also I found someone speculating it was related to free kernel =
memory so that was updated on some routers however I saw little =
difference(although not sure if adding it in /etc/sysctl.conf actually =
does anything as some haven=92t had the vm.kmem_size value change at =
all, despite reboots).

Can anyone explain what the buckets (16,32,64,128) do and what they =
represent? Also if you know anything about troubleshooting bucket =
failures? There doesn=92t appear to be anything useful online despite =
many searches.

I am aware that it may not have anything to do with the problems I am =
seeing, but I would still like to understand.

Thanks,

Mark


=
--------------------------------------------------------------------------=
-----

Background:

I have a number of FreeBSD 9 boxes running as routers in a testing =
environment where I am seeing odd latency issues. Most of them have =
different hardware and configurations now just to rule out configuration =
or hardware issues.

What I am seeing is occasional latency increases when routing traffic so =
that on the =93client=94 end the experience is like the link has hung =
for a split second - this happens for both L2TP/PPPoE clients and plain =
ethernet connected clients. I haven=92t been able to identify anything =
obvious like fragmentation, mtu issues, etc.

The setup is:

3 x Dual Quad Core, 8GB RAM running as LAC & LNS w/ Quagga(BGP, OSPF), =
MPD5, Netflow, IPFW Stateless(recently introduced and my problem existed =
before filtering was introduced).
3 x Dual Quad Core, 8GB RAM running as Routers w/ Quagga(BGP, OSPF)

Router 1
=3D=3D=3D=3D=3D=3D=3D=3D

ITEM                   SIZE  LIMIT     USED     FREE      REQ FAIL SLEEP
16 Bucket:              152,      0,     198,       2,     198,   0,   0
32 Bucket:              280,      0,     291,       3,     291,  16,   0
64 Bucket:              536,      0,     221,       3,     221,  58,   0
128 Bucket:            1048,      0,    7993,       2,    7993, 505,   0
vm.kmem_size: 4127952896

Router 2(LNS)
=3D=3D=3D=3D=3D=3D=3D=3D
ITEM                   SIZE  LIMIT     USED     FREE      REQ FAIL SLEEP
16 Bucket:              152,      0,     199,       1,     199,   0,   0
32 Bucket:              280,      0,     329,       7,     333,   0,   0
64 Bucket:              536,      0,     241,       4,     248,  57,   0
128 Bucket:            1048,      0,    7960,       2,    9479,1002,   0
vm.kmem_size: 2147483648

Router 3(LNS)
=3D=3D=3D=3D=3D=3D=3D=3D
ITEM                   SIZE  LIMIT     USED     FREE      REQ FAIL SLEEP
16 Bucket:              152,      0,     194,       6,     194,   0,   0
32 Bucket:              280,      0,     325,      11,     325,  20,   0
64 Bucket:              536,      0,     278,       2,     278,  57,   0
128 Bucket:            1048,      0,    8031,       0,    8031, 955,   0
vm.kmem_size: 4127952896

Router 4(LNS)
=3D=3D=3D=3D=3D=3D=3D=3D
ITEM                   SIZE  LIMIT     USED     FREE      REQ FAIL SLEEP
16 Bucket:              152,      0,     197,       3,     197,   0,   0
32 Bucket:              280,      0,     265,       1,     265,   0,   0
64 Bucket:              536,      0,     227,       4,     227,  57,   0
128 Bucket:            1048,      0,    7824,       0,    7824, 721,   0
vm.kmem_size: 2147483648

Router 5
=3D=3D=3D=3D=3D=3D=3D=3D
ITEM                   SIZE  LIMIT     USED     FREE      REQ FAIL SLEEP
16 Bucket:              152,      0,     204,      21,     204,   0,   0
32 Bucket:              280,      0,     313,       9,     313,   1,   0
64 Bucket:              536,      0,     268,       5,     268,  57,   0
128 Bucket:            1048,      0,    9204,       0,    9204,5723,   0
vm.kmem_size: 4127952896

Router 6
=3D=3D=3D=3D=3D=3D=3D=3D
ITEM                   SIZE  LIMIT     USED     FREE      REQ FAIL SLEEP
16 Bucket:              152,      0,      80,      20,      80,   0,   0
32 Bucket:              280,      0,      86,      12,      86,   0,   0
64 Bucket:              536,      0,      49,       0,      49,  57,   0
128 Bucket:            1048,      0,    8697,       0,    8697,5604,   0
vm.kmem_size: 3108118528







Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4260012E-9911-47AB-88B9-FDCD6748E6EE>