blob: 14395ebc446b01289eae5e24c01b73c2e682270d (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
|
BRAS tuning
===========
Recommendations for BRAS (Broadband Remote Access Server) performance.
Network tuning
--------------
Disable kernel mitigations to maximize performance
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Edit file ``/etc/default/grub``
.. code-block:: sh
GRUB_CMDLINE_LINUX_DEFAULT="intel_idle.max_cstate=0 processor.max_cstate=1 idle=poll quiet mitigations=off"
Additional GRUB CMD arguments:
* ixgbe.allow_unsupported_sfp=1 - Allow to use not original Intel SFP+ modules
* pcie_aspm=off - Disable Active-State Power Management
After saving, please update grub settings
.. code-block:: sh
$sudo update-grub
Warning! Enabling the idle loop (``idle=poll``) parameter can cause 100% CPU utilization on your VM (if you're using virtual enviroments like ProxMox, VMWare, etc.)
Disable NIC offloads
^^^^^^^^^^^^^^^^^^^^
Disable hardware offloads,increase Tx/Rx buffers and queue length on your NICs to prevent speed problems.
Please note, that GSO offload changed to tx-gso-partial in Linux kernels 4.15 and later.
Debian ``/etc/network/interfaces``:
.. code-block:: sh
allow-hotplug eth0
iface eth0 inet manual
up ethtool -K eth0 tso off gso off gro off rxvlan off txvlan off rx-vlan-filter off ntuple on &> /dev/null
up ethtool -K eth0 tx-gso-partial off &> /dev/null
up ethtool -G eth0 rx 4096 tx 4096 &> /dev/null
up ip link set eth0 txqueuelen 10000 &> /dev/null
Please determine your NIC queue and buffers limit before increase:
.. code-block:: sh
ethtool -g eth0
Fix Download speed problem (shaper)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Change shaper from htb to tbf
.. code-block:: sh
[shaper]
…
down-limiter=tbf
Default rate limits (shaper)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
If radius-server Access-Accept answer has no compatible speed attributes - to prevent unlimited session speed you can add default rate-limits (in Kbps).
``nano /etc/accel-ppp.conf``
.. code-block:: sh
[shaper]
rate-limit=888/888
Change PPPoE MTU
^^^^^^^^^^^^^^^^
You can adjust allowed PPPoE min/max MTU/MRU settings:
``nano /etc/accel-ppp.conf``
.. code-block:: sh
[ppp]
verbose=1
min-mtu=1280
mtu=1492
mru=1492
Hotplug optimization
^^^^^^^^^^^^^^^^^^^^
To generate hotplug events on IPoE interfaces (Debian 10):
``nano /lib/udev/ifupdown-hotplug``
.. code-block:: sh
case "$ACTION" in
add)
# these interfaces generate hotplug events *after* they are brought up
case $INTERFACE in
ppp*|ippp*|isdn*|plip*|lo|irda*|ipsec*
just add ``|ipoe*`` after ``|ipsec*``
repeat with file ``/lib/udev/net.agent``
SYSTEMD-UDEV optimizations
^^^^^^^^^^^^^^^^^^^^^^^^^^
1. Delete ``99-default.link`` from ``/lib/systemd/network/`` directory
.. code-block:: sh
rm /lib/systemd/network/99-default.link
2. Change ``/lib/udev/rules.d/99-systemd.rules``
.. code-block:: sh
ACTION=="add", SUBSYSTEM=="net", KERNEL!="lo|ppp*|ipoe*", RUN+="/lib/systemd/systemd-sysctl --prefix=/net/ipv4/conf/$name --prefix=/net/ipv4/neigh/$name --prefix=/net/ipv6/conf/$name --prefix=/net/ipv6/neigh/$name"
Add ``|ppp*|ipoe*`` to ``KERNEL!="lo"``
3. Change ``/lib/udev/rules.d/80-ifupdown.rules``
.. code-block:: sh
SUBSYSTEM=="net", ACTION=="add|remove", KERNEL!="ppp*|ipoe*", RUN+="ifupdown-hotplug"
Add ``KERNEL!="ppp*|ipoe*"``
|
