blob: f2a2488f1bf33d8290384c4823fccee2b39afe36 [file] [log] [blame]
Horms6ce16692006-07-03 19:35:40 -07001/proc/sys/net/ipv4/vs/* Variables:
2
3am_droprate - INTEGER
4 default 10
5
6 It sets the always mode drop rate, which is used in the mode 3
7 of the drop_rate defense.
8
9amemthresh - INTEGER
10 default 1024
11
12 It sets the available memory threshold (in pages), which is
13 used in the automatic modes of defense. When there is no
14 enough available memory, the respective strategy will be
15 enabled and the variable is automatically set to 2, otherwise
16 the strategy is disabled and the variable is set to 1.
17
Simon Horman7e777dd2011-09-09 17:07:43 +090018conntrack - BOOLEAN
19 0 - disabled (default)
20 not 0 - enabled
21
22 If set, maintain connection tracking entries for
23 connections handled by IPVS.
24
25 This should be enabled if connections handled by IPVS are to be
26 also handled by stateful firewall rules. That is, iptables rules
27 that make use of connection tracking. It is a performance
28 optimisation to disable this setting otherwise.
29
30 Connections handled by the IPVS FTP application module
31 will have connection tracking entries regardless of this setting.
32
Simon Horman40cb1f92011-09-29 16:27:37 +090033 Only available when IPVS is compiled with CONFIG_IP_VS_NFCT enabled.
Simon Horman7e777dd2011-09-09 17:07:43 +090034
Horms6ce16692006-07-03 19:35:40 -070035cache_bypass - BOOLEAN
36 0 - disabled (default)
37 not 0 - enabled
38
39 If it is enabled, forward packets to the original destination
40 directly when no cache server is available and destination
41 address is not local (iph->daddr is RTN_UNICAST). It is mostly
42 used in transparent web cache cluster.
43
44debug_level - INTEGER
45 0 - transmission error messages (default)
46 1 - non-fatal error messages
47 2 - configuration
48 3 - destination trash
49 4 - drop entry
50 5 - service lookup
51 6 - scheduling
52 7 - connection new/expire, lookup and synchronization
53 8 - state transition
54 9 - binding destination, template checks and applications
55 10 - IPVS packet transmission
56 11 - IPVS packet handling (ip_vs_in/ip_vs_out)
57 12 or more - packet traversal
58
Simon Horman40cb1f92011-09-29 16:27:37 +090059 Only available when IPVS is compiled with CONFIG_IP_VS_DEBUG enabled.
Horms6ce16692006-07-03 19:35:40 -070060
61 Higher debugging levels include the messages for lower debugging
62 levels, so setting debug level 2, includes level 0, 1 and 2
63 messages. Thus, logging becomes more and more verbose the higher
64 the level.
65
66drop_entry - INTEGER
67 0 - disabled (default)
68
69 The drop_entry defense is to randomly drop entries in the
70 connection hash table, just in order to collect back some
71 memory for new connections. In the current code, the
72 drop_entry procedure can be activated every second, then it
73 randomly scans 1/32 of the whole and drops entries that are in
74 the SYN-RECV/SYNACK state, which should be effective against
75 syn-flooding attack.
76
77 The valid values of drop_entry are from 0 to 3, where 0 means
78 that this strategy is always disabled, 1 and 2 mean automatic
79 modes (when there is no enough available memory, the strategy
80 is enabled and the variable is automatically set to 2,
81 otherwise the strategy is disabled and the variable is set to
82 1), and 3 means that that the strategy is always enabled.
83
84drop_packet - INTEGER
85 0 - disabled (default)
86
87 The drop_packet defense is designed to drop 1/rate packets
88 before forwarding them to real servers. If the rate is 1, then
89 drop all the incoming packets.
90
91 The value definition is the same as that of the drop_entry. In
92 the automatic mode, the rate is determined by the follow
93 formula: rate = amemthresh / (amemthresh - available_memory)
94 when available memory is less than the available memory
95 threshold. When the mode 3 is set, the always mode drop rate
96 is controlled by the /proc/sys/net/ipv4/vs/am_droprate.
97
98expire_nodest_conn - BOOLEAN
99 0 - disabled (default)
100 not 0 - enabled
101
102 The default value is 0, the load balancer will silently drop
103 packets when its destination server is not available. It may
104 be useful, when user-space monitoring program deletes the
105 destination server (because of server overload or wrong
106 detection) and add back the server later, and the connections
107 to the server can continue.
108
109 If this feature is enabled, the load balancer will expire the
110 connection immediately when a packet arrives and its
111 destination server is not available, then the client program
112 will be notified that the connection is closed. This is
113 equivalent to the feature some people requires to flush
114 connections when its destination is not available.
115
116expire_quiescent_template - BOOLEAN
117 0 - disabled (default)
118 not 0 - enabled
119
120 When set to a non-zero value, the load balancer will expire
121 persistent templates when the destination server is quiescent.
122 This may be useful, when a user makes a destination server
123 quiescent by setting its weight to 0 and it is desired that
124 subsequent otherwise persistent connections are sent to a
125 different destination server. By default new persistent
126 connections are allowed to quiescent destination servers.
127
128 If this feature is enabled, the load balancer will expire the
129 persistence template if it is to be used to schedule a new
130 connection and the destination server is quiescent.
131
132nat_icmp_send - BOOLEAN
133 0 - disabled (default)
134 not 0 - enabled
135
136 It controls sending icmp error messages (ICMP_DEST_UNREACH)
137 for VS/NAT when the load balancer receives packets from real
138 servers but the connection entries don't exist.
139
140secure_tcp - INTEGER
141 0 - disabled (default)
142
Simon Horman325aadc2011-09-29 16:14:51 +0900143 The secure_tcp defense is to use a more complicated TCP state
144 transition table. For VS/NAT, it also delays entering the
145 TCP ESTABLISHED state until the three way handshake is completed.
Horms6ce16692006-07-03 19:35:40 -0700146
Simon Horman325aadc2011-09-29 16:14:51 +0900147 The value definition is the same as that of drop_entry and
Horms6ce16692006-07-03 19:35:40 -0700148 drop_packet.
149
150sync_threshold - INTEGER
151 default 3
152
153 It sets synchronization threshold, which is the minimum number
154 of incoming packets that a connection needs to receive before
155 the connection will be synchronized. A connection will be
156 synchronized, every time the number of its incoming packets
157 modulus 50 equals the threshold. The range of the threshold is
158 from 0 to 49.
Simon Horman7e777dd2011-09-09 17:07:43 +0900159
160snat_reroute - BOOLEAN
161 0 - disabled
162 not 0 - enabled (default)
163
164 If enabled, recalculate the route of SNATed packets from
165 realservers so that they are routed as if they originate from the
166 director. Otherwise they are routed as if they are forwarded by the
167 director.
168
169 If policy routing is in effect then it is possible that the route
170 of a packet originating from a director is routed differently to a
171 packet being forwarded by the director.
172
173 If policy routing is not in effect then the recalculated route will
174 always be the same as the original route so it is an optimisation
175 to disable snat_reroute and avoid the recalculation.
176
177sync_version - INTEGER
178 default 1
179
180 The version of the synchronisation protocol used when sending
181 synchronisation messages.
182
183 0 selects the original synchronisation protocol (version 0). This
184 should be used when sending synchronisation messages to a legacy
185 system that only understands the original synchronisation protocol.
186
187 1 selects the current synchronisation protocol (version 1). This
188 should be used where possible.
189
190 Kernels with this sync_version entry are able to receive messages
191 of both version 1 and version 2 of the synchronisation protocol.