1 /proc/sys/net/ipv4/* Variables:
2
3 ip_forward - BOOLEAN
4 0 - disabled (default)
5 not 0 - enabled
6
7 Forward Packets between interfaces.
8
9 This variable is special, its change resets all configuration
10 parameters to their default state (RFC1122 for hosts, RFC1812
11 for routers)
12
13 ip_default_ttl - INTEGER
14 default 64
15
16 ip_no_pmtu_disc - BOOLEAN
17 Disable Path MTU Discovery.
18 default FALSE
19
20 IP Fragmentation:
21
22 ipfrag_high_thresh - INTEGER
23 Maximum memory used to reassemble IP fragments. When
24 ipfrag_high_thresh bytes of memory is allocated for this purpose,
25 the fragment handler will toss packets until ipfrag_low_thresh
26 is reached.
27
28 ipfrag_low_thresh - INTEGER
29 See ipfrag_high_thresh
30
31 ipfrag_time - INTEGER
32 Time in seconds to keep an IP fragment in memory.
33
34 INET peer storage:
35
36 inet_peer_threshold - INTEGER
37 The approximate size of the storage. Starting from this threshold
38 entries will be thrown aggressively. This threshold also determines
39 entries' time-to-live and time intervals between garbage collection
40 passes. More entries, less time-to-live, less GC interval.
41
42 inet_peer_minttl - INTEGER
43 Minimum time-to-live of entries. Should be enough to cover fragment
44 time-to-live on the reassembling side. This minimum time-to-live is
45 guaranteed if the pool size is less than inet_peer_threshold.
46 Measured in jiffies.
47
48 inet_peer_maxttl - INTEGER
49 Maximum time-to-live of entries. Unused entries will expire after
50 this period of time if there is no memory pressure on the pool (i.e.
51 when the number of entries in the pool is very small).
52 Measured in jiffies.
53
54 inet_peer_gc_mintime - INTEGER
55 Minimum interval between garbage collection passes. This interval is
56 in effect under high memory pressure on the pool.
57 Measured in jiffies.
58
59 inet_peer_gc_maxtime - INTEGER
60 Minimum interval between garbage collection passes. This interval is
61 in effect under low (or absent) memory pressure on the pool.
62 Measured in jiffies.
63
64 TCP variables:
65
66 tcp_syn_retries - INTEGER
67 Number of times initial SYNs for an active TCP connection attempt
68 will be retransmitted. Should not be higher than 255. Default value
69 is 5, which corresponds to ~180seconds.
70
71 tcp_synack_retries - INTEGER
72 Number of times SYNACKs for a passive TCP connection attempt will
73 be retransmitted. Should not be higher than 255. Default value
74 is 5, which corresponds to ~180seconds.
75
76 tcp_keepalive_time - INTEGER
77 How often TCP sends out keepalive messages when keepalive is enabled.
78 Default: 2hours.
79
80 tcp_keepalive_probes - INTEGER
81 How many keepalive probes TCP sends out, until it decides that the
82 connection is broken. Default value: 9.
83
84 tcp_keepalive_interval - INTEGER
85 How frequently the probes are send out. Multiplied by
86 tcp_keepalive_probes it is time to kill not responding connection,
87 after probes started. Default value: 75sec i.e. connection
88 will be aborted after ~11 minutes of retries.
89
90 tcp_retries1 - INTEGER
91 How many times to retry before deciding that something is wrong
92 and it is necessary to report this suspection to network layer.
93 Minimal RFC value is 3, it is default, which corresponds
94 to ~3sec-8min depending on RTO.
95
96 tcp_retries2 - INTEGER
97 How may times to retry before killing alive TCP connection.
98 RFC1122 says that the limit should be longer than 100 sec.
99 It is too small number. Default value 15 corresponds to ~13-30min
100 depending on RTO.
101
102 tcp_orphan_retries - INTEGER
103 How may times to retry before killing TCP connection, closed
104 by our side. Default value 7 corresponds to ~50sec-16min
105 depending on RTO. If you machine is loaded WEB server,
106 you should think about lowering this value, such sockets
107 may consume significant resources. Cf. tcp_max_orphans.
108
109 tcp_fin_timeout - INTEGER
110 Time to hold socket in state FIN-WAIT-2, if it was closed
111 by our side. Peer can be broken and never close its side,
112 or even died unexpectedly. Default value is 60sec.
113 Usual value used in 2.2 was 180 seconds, you may restore
114 it, but remember that if your machine is even underloaded WEB server,
115 you risk to overflow memory with kilotons of dead sockets,
116 FIN-WAIT-2 sockets are less dangerous than FIN-WAIT-1,
117 because they eat maximum 1.5K of memory, but they tend
118 to live longer. Cf. tcp_max_orphans.
119
120 tcp_max_tw_buckets - INTEGER
121 Maximal number of timewait sockets held by system simultaneously.
122 If this number is exceeded time-wait socket is immediately destroyed
123 and warning is printed. This limit exists only to prevent
124 simple DoS attacks, you _must_ not lower the limit artificially,
125 but rather increase it (probably, after increasing installed memory),
126 if network conditions require more than default value.
127
128 tcp_tw_recycle - BOOLEAN
129 Enable fast recycling TIME-WAIT sockets. Default value is 1.
130 It should not be changed without advice/request of technical
131 experts.
132
133 tcp_max_orphans - INTEGER
134 Maximal number of TCP sockets not attached to any user file handle,
135 held by system. If this number is exceeded orphaned connections are
136 reset immediately and warning is printed. This limit exists
137 only to prevent simple DoS attacks, you _must_ not rely on this
138 or lower the limit artificially, but rather increase it
139 (probably, after increasing installed memory),
140 if network conditions require more than default value,
141 and tune network services to linger and kill such states
142 more aggressively. Let me to remind again: each orphan eats
143 up to ~64K of unswappable memory.
144
145 tcp_abort_on_overflow - BOOLEAN
146 If listening service is too slow to accept new connections,
147 reset them. Default state is FALSE. It means that if overflow
148 occurred due to a burst, connection will recover. Enable this
149 option _only_ if you are really sure that listening daemon
150 cannot be tuned to accept connections faster. Enabling this
151 option can harm clients of your server.
152
153 tcp_syncookies - BOOLEAN
154 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES
155 Send out syncookies when the syn backlog queue of a socket
156 overflows. This is to prevent against the common 'syn flood attack'
157 Default: FALSE
158
159 Note, that syncookies is fallback facility.
160 It MUST NOT be used to help highly loaded servers to stand
161 against legal connection rate. If you see synflood warnings
162 in your logs, but investigation shows that they occur
163 because of overload with legal connections, you should tune
164 another parameters until this warning disappear.
165 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
166
167 syncookies seriously violate TCP protocol, do not allow
168 to use TCP extensions, can result in serious degradation
169 of some services (f.e. SMTP relaying), visible not by you,
170 but your clients and relays, contacting you. While you see
171 synflood warnings in logs not being really flooded, your server
172 is seriously misconfigured.
173
174 tcp_stdurg - BOOLEAN
175 Use the Host requirements interpretation of the TCP urg pointer field.
176 Most hosts use the older BSD interpretation, so if you turn this on
177 Linux might not communicate correctly with them.
178 Default: FALSE
179
180 tcp_max_syn_backlog - INTEGER
181 Maximal number of remembered connection requests, which are
182 still did not receive an acknowledgement from connecting client.
183 Default value is 1024 for systems with more than 128Mb of memory,
184 and 128 for low memory machines. If server suffers of overload,
185 try to increase this number. Warning! If you make it greater
186 than 1024, it would be better to change TCP_SYNQ_HSIZE in
187 include/net/tcp.h to keep TCP_SYNQ_HSIZE*16<=tcp_max_syn_backlog
188 and to recompile kernel.
189
190 tcp_window_scaling - BOOLEAN
191 Enable window scaling as defined in RFC1323.
192
193 tcp_timestamps - BOOLEAN
194 Enable timestamps as defined in RFC1323.
195
196 tcp_sack - BOOLEAN
197 Enable select acknowledgments (SACKS).
198
199 tcp_fack - BOOLEAN
200 Enable FACK congestion avoidance and fast restransmission.
201 The value is not used, if tcp_sack is not enabled.
202
203 tcp_dsack - BOOLEAN
204 Allows TCP to send "duplicate" SACKs.
205
206 tcp_ecn - BOOLEAN
207 Enable Explicit Congestion Notification in TCP.
208
209 tcp_reordering - INTEGER
210 Maximal reordering of packets in a TCP stream.
211 Default: 3
212
213 tcp_retrans_collapse - BOOLEAN
214 Bug-to-bug compatibility with some broken printers.
215 On retransmit try to send bigger packets to work around bugs in
216 certain TCP stacks.
217
218 tcp_wmem - vector of 3 INTEGERs: min, default, max
219 min: Amount of memory reserved for send buffers for TCP socket.
220 Each TCP socket has rights to use it due to fact of its birth.
221 Default: 4K
222
223 default: Amount of memory allowed for send buffers for TCP socket
224 by default. This value overrides net.core.wmem_default used
225 by other protocols, it is usually lower than net.core.wmem_default.
226 Default: 16K
227
228 max: Maximal amount of memory allowed for automatically selected
229 send buffers for TCP socket. This value does not override
230 net.core.wmem_max, "static" selection via SO_SNDBUF does not use this.
231 Default: 128K
232
233 tcp_rmem - vector of 3 INTEGERs: min, default, max
234 min: Minimal size of receive buffer used by TCP sockets.
235 It is guaranteed to each TCP socket, even under moderate memory
236 pressure.
237 Default: 8K
238
239 default: default size of receive buffer used by TCP sockets.
240 This value overrides net.core.rmem_default used by other protocols.
241 Default: 87380 bytes. This value results in window of 65535 with
242 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
243 less for default tcp_app_win. See below about these variables.
244
245 max: maximal size of receive buffer allowed for automatically
246 selected receiver buffers for TCP socket. This value does not override
247 net.core.rmem_max, "static" selection via SO_RCVBUF does not use this.
248 Default: 87380*2 bytes.
249
250 tcp_mem - vector of 3 INTEGERs: min, pressure, max
251 low: below this number of pages TCP is not bothered about its
252 memory appetite.
253
254 pressure: when amount of memory allocated by TCP exceeds this number
255 of pages, TCP moderates its memory consumption and enters memory
256 pressure mode, which is exited when memory consumtion falls
257 under "low".
258
259 high: number of pages allowed for queueing by all TCP sockets.
260
261 Defaults are calculated at boot time from amount of available
262 memory.
263
264 tcp_app_win - INTEGER
265 Reserve max(window/2^tcp_app_win, mss) of window for application
266 buffer. Value 0 is special, it means that nothing is reserved.
267 Default: 31
268
269 tcp_adv_win_scale - INTEGER
270 Count buffering overhead as bytes/2^tcp_adv_win_scale
271 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
272 if it is <= 0.
273 Default: 2
274
275 ip_local_port_range - 2 INTEGERS
276 Defines the local port range that is used by TCP and UDP to
277 choose the local port. The first number is the first, the
278 second the last local port number. Default value depends on
279 amount of memory available on the system:
280 > 128Mb 32768-61000
281 < 128Mb 1024-4999 or even less.
282 This number defines number of active connections, which this
283 system can issue simultaneously to systems not supporting
284 TCP extensions (timestamps). With tcp_tw_recycle enabled
285 (i.e. by default) range 1024-4999 is enough to issue up to
286 2000 connections per second to systems supporting timestamps.
287
288 icmp_echo_ignore_all - BOOLEAN
289 icmp_echo_ignore_broadcasts - BOOLEAN
290 If either is set to true, then the kernel will ignore either all
291 ICMP ECHO requests sent to it or just those to broadcast/multicast
292 addresses, respectively.
293
294 icmp_destunreach_rate - INTEGER
295 icmp_paramprob_rate - INTEGER
296 icmp_timeexceed_rate - INTEGER
297 icmp_echoreply_rate - INTEGER (not enabled per default)
298 Limit the maximal rates for sending ICMP packets to specific targets.
299 0 to disable any limiting, otherwise the maximal rate in jiffies(1)
300 See the source for more information.
301
302 icmp_ignore_bogus_error_responses - BOOLEAN
303 Some routers violate RFC 1122 by sending bogus responses to broadcast
304 frames. Such violations are normally logged via a kernel warning.
305 If this is set to TRUE, the kernel will not give such warnings, which
306 will avoid log file clutter.
307 Default: FALSE
308
309 (1) Jiffie: internal timeunit for the kernel. On the i386 1/100s, on the
310 Alpha 1/1024s. See the HZ define in /usr/include/asm/param.h for the exact
311 value on your system.
312
313 conf/interface/*:
314 conf/all/* is special and changes the settings for all interfaces.
315 Change special settings per interface.
316
317 log_martians - BOOLEAN
318 Log packets with impossible addresses to kernel log.
319
320 accept_redirects - BOOLEAN
321 Accept ICMP redirect messages.
322 default TRUE (host)
323 FALSE (router)
324
325 forwarding - BOOLEAN
326 Enable IP forwarding on this interface.
327
328 mc_forwarding - BOOLEAN
329 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
330 and a multicast routing daemon is required.
331
332 proxy_arp - BOOLEAN
333 Do proxy arp.
334
335 shared_media - BOOLEAN
336 Send(router) or accept(host) RFC1620 shared media redirects.
337 Overrides ip_secure_redirects.
338 default TRUE
339
340 secure_redirects - BOOLEAN
341 Accept ICMP redirect messages only for gateways,
342 listed in default gateway list.
343 default TRUE
344
345 send_redirects - BOOLEAN
346 Send redirects, if router. Default: TRUE
347
348 bootp_relay - BOOLEAN
349 Accept packets with source address 0.b.c.d destined
350 not to this host as local ones. It is supposed, that
351 BOOTP relay daemon will catch and forward such packets.
352
353 default FALSE
354 Not Implemented Yet.
355
356 accept_source_route - BOOLEAN
357 Accept packets with SRR option.
358 default TRUE (router)
359 FALSE (host)
360
361 rp_filter - BOOLEAN
362 1 - do source validation by reversed path, as specified in RFC1812
363 Recommended option for single homed hosts and stub network
364 routers. Could cause troubles for complicated (not loop free)
365 networks running a slow unreliable protocol (sort of RIP),
366 or using static routes.
367
368 0 - No source validation.
369
370 Default value is 0. Note that some distributions enable it
371 in startip scripts.
372
373 Alexey Kuznetsov.
374 kuznet@ms2.inr.ac.ru
375
376 Updated by:
377 Andi Kleen
378 ak@muc.de
379 $Id: ip-sysctl.txt,v 1.17 2000/11/06 07:15:36 davem Exp $
This page was automatically generated by the
LXR engine.
Visit the LXR main site for more
information.