5 This document explains the meaning of SNMP counters.
9 All layer 4 packets and ICMP packets will change these counters, but
10 these counters won't be changed by layer 2 packets (such as STP) or
14 Defined in `RFC1213 ipInReceives`_
16 .. _RFC1213 ipInReceives: https://tools.ietf.org/html/rfc1213#page-26
18 The number of packets received by the IP layer. It gets increasing at the
19 beginning of ip_rcv function, always be updated together with
20 IpExtInOctets. It will be increased even if the packet is dropped
21 later (e.g. due to the IP header is invalid or the checksum is wrong
22 and so on). It indicates the number of aggregated segments after
26 Defined in `RFC1213 ipInDelivers`_
28 .. _RFC1213 ipInDelivers: https://tools.ietf.org/html/rfc1213#page-28
30 The number of packets delivers to the upper layer protocols. E.g. TCP, UDP,
31 ICMP and so on. If no one listens on a raw socket, only kernel
32 supported protocols will be delivered, if someone listens on the raw
33 socket, all valid IP packets will be delivered.
36 Defined in `RFC1213 ipOutRequests`_
38 .. _RFC1213 ipOutRequests: https://tools.ietf.org/html/rfc1213#page-28
40 The number of packets sent via IP layer, for both single cast and
41 multicast packets, and would always be updated together with
44 * IpExtInOctets and IpExtOutOctets
45 They are Linux kernel extensions, no RFC definitions. Please note,
46 RFC1213 indeed defines ifInOctets and ifOutOctets, but they
47 are different things. The ifInOctets and ifOutOctets include the MAC
48 layer header size but IpExtInOctets and IpExtOutOctets don't, they
49 only include the IP layer header and the IP layer data.
51 * IpExtInNoECTPkts, IpExtInECT1Pkts, IpExtInECT0Pkts, IpExtInCEPkts
52 They indicate the number of four kinds of ECN IP packets, please refer
53 `Explicit Congestion Notification`_ for more details.
55 .. _Explicit Congestion Notification: https://tools.ietf.org/html/rfc3168#page-6
57 These 4 counters calculate how many packets received per ECN
58 status. They count the real frame number regardless the LRO/GRO. So
59 for the same packet, you might find that IpInReceives count 1, but
60 IpExtInNoECTPkts counts 2 or more.
63 Defined in `RFC1213 ipInHdrErrors`_. It indicates the packet is
64 dropped due to the IP header error. It might happen in both IP input
67 .. _RFC1213 ipInHdrErrors: https://tools.ietf.org/html/rfc1213#page-27
70 Defined in `RFC1213 ipInAddrErrors`_. It will be increased in two
71 scenarios: (1) The IP address is invalid. (2) The destination IP
72 address is not a local address and IP forwarding is not enabled
74 .. _RFC1213 ipInAddrErrors: https://tools.ietf.org/html/rfc1213#page-27
77 This counter means the packet is dropped when the IP stack receives a
78 packet and can't find a route for it from the route table. It might
79 happen when IP forwarding is enabled and the destination IP address is
80 not a local address and there is no route for the destination IP
84 Defined in `RFC1213 ipInUnknownProtos`_. It will be increased if the
85 layer 4 protocol is unsupported by kernel. If an application is using
86 raw socket, kernel will always deliver the packet to the raw socket
87 and this counter won't be increased.
89 .. _RFC1213 ipInUnknownProtos: https://tools.ietf.org/html/rfc1213#page-27
91 * IpExtInTruncatedPkts
92 For IPv4 packet, it means the actual data size is smaller than the
93 "Total Length" field in the IPv4 header.
96 Defined in `RFC1213 ipInDiscards`_. It indicates the packet is dropped
97 in the IP receiving path and due to kernel internal reasons (e.g. no
100 .. _RFC1213 ipInDiscards: https://tools.ietf.org/html/rfc1213#page-28
103 Defined in `RFC1213 ipOutDiscards`_. It indicates the packet is
104 dropped in the IP sending path and due to kernel internal reasons.
106 .. _RFC1213 ipOutDiscards: https://tools.ietf.org/html/rfc1213#page-28
109 Defined in `RFC1213 ipOutNoRoutes`_. It indicates the packet is
110 dropped in the IP sending path and no route is found for it.
112 .. _RFC1213 ipOutNoRoutes: https://tools.ietf.org/html/rfc1213#page-29
116 * IcmpInMsgs and IcmpOutMsgs
117 Defined by `RFC1213 icmpInMsgs`_ and `RFC1213 icmpOutMsgs`_
119 .. _RFC1213 icmpInMsgs: https://tools.ietf.org/html/rfc1213#page-41
120 .. _RFC1213 icmpOutMsgs: https://tools.ietf.org/html/rfc1213#page-43
122 As mentioned in the RFC1213, these two counters include errors, they
123 would be increased even if the ICMP packet has an invalid type. The
124 ICMP output path will check the header of a raw socket, so the
125 IcmpOutMsgs would still be updated if the IP header is constructed by
129 | These counters include most of common ICMP types, they are:
130 | IcmpInDestUnreachs: `RFC1213 icmpInDestUnreachs`_
131 | IcmpInTimeExcds: `RFC1213 icmpInTimeExcds`_
132 | IcmpInParmProbs: `RFC1213 icmpInParmProbs`_
133 | IcmpInSrcQuenchs: `RFC1213 icmpInSrcQuenchs`_
134 | IcmpInRedirects: `RFC1213 icmpInRedirects`_
135 | IcmpInEchos: `RFC1213 icmpInEchos`_
136 | IcmpInEchoReps: `RFC1213 icmpInEchoReps`_
137 | IcmpInTimestamps: `RFC1213 icmpInTimestamps`_
138 | IcmpInTimestampReps: `RFC1213 icmpInTimestampReps`_
139 | IcmpInAddrMasks: `RFC1213 icmpInAddrMasks`_
140 | IcmpInAddrMaskReps: `RFC1213 icmpInAddrMaskReps`_
141 | IcmpOutDestUnreachs: `RFC1213 icmpOutDestUnreachs`_
142 | IcmpOutTimeExcds: `RFC1213 icmpOutTimeExcds`_
143 | IcmpOutParmProbs: `RFC1213 icmpOutParmProbs`_
144 | IcmpOutSrcQuenchs: `RFC1213 icmpOutSrcQuenchs`_
145 | IcmpOutRedirects: `RFC1213 icmpOutRedirects`_
146 | IcmpOutEchos: `RFC1213 icmpOutEchos`_
147 | IcmpOutEchoReps: `RFC1213 icmpOutEchoReps`_
148 | IcmpOutTimestamps: `RFC1213 icmpOutTimestamps`_
149 | IcmpOutTimestampReps: `RFC1213 icmpOutTimestampReps`_
150 | IcmpOutAddrMasks: `RFC1213 icmpOutAddrMasks`_
151 | IcmpOutAddrMaskReps: `RFC1213 icmpOutAddrMaskReps`_
153 .. _RFC1213 icmpInDestUnreachs: https://tools.ietf.org/html/rfc1213#page-41
154 .. _RFC1213 icmpInTimeExcds: https://tools.ietf.org/html/rfc1213#page-41
155 .. _RFC1213 icmpInParmProbs: https://tools.ietf.org/html/rfc1213#page-42
156 .. _RFC1213 icmpInSrcQuenchs: https://tools.ietf.org/html/rfc1213#page-42
157 .. _RFC1213 icmpInRedirects: https://tools.ietf.org/html/rfc1213#page-42
158 .. _RFC1213 icmpInEchos: https://tools.ietf.org/html/rfc1213#page-42
159 .. _RFC1213 icmpInEchoReps: https://tools.ietf.org/html/rfc1213#page-42
160 .. _RFC1213 icmpInTimestamps: https://tools.ietf.org/html/rfc1213#page-42
161 .. _RFC1213 icmpInTimestampReps: https://tools.ietf.org/html/rfc1213#page-43
162 .. _RFC1213 icmpInAddrMasks: https://tools.ietf.org/html/rfc1213#page-43
163 .. _RFC1213 icmpInAddrMaskReps: https://tools.ietf.org/html/rfc1213#page-43
165 .. _RFC1213 icmpOutDestUnreachs: https://tools.ietf.org/html/rfc1213#page-44
166 .. _RFC1213 icmpOutTimeExcds: https://tools.ietf.org/html/rfc1213#page-44
167 .. _RFC1213 icmpOutParmProbs: https://tools.ietf.org/html/rfc1213#page-44
168 .. _RFC1213 icmpOutSrcQuenchs: https://tools.ietf.org/html/rfc1213#page-44
169 .. _RFC1213 icmpOutRedirects: https://tools.ietf.org/html/rfc1213#page-44
170 .. _RFC1213 icmpOutEchos: https://tools.ietf.org/html/rfc1213#page-45
171 .. _RFC1213 icmpOutEchoReps: https://tools.ietf.org/html/rfc1213#page-45
172 .. _RFC1213 icmpOutTimestamps: https://tools.ietf.org/html/rfc1213#page-45
173 .. _RFC1213 icmpOutTimestampReps: https://tools.ietf.org/html/rfc1213#page-45
174 .. _RFC1213 icmpOutAddrMasks: https://tools.ietf.org/html/rfc1213#page-45
175 .. _RFC1213 icmpOutAddrMaskReps: https://tools.ietf.org/html/rfc1213#page-46
177 Every ICMP type has two counters: 'In' and 'Out'. E.g., for the ICMP
178 Echo packet, they are IcmpInEchos and IcmpOutEchos. Their meanings are
179 straightforward. The 'In' counter means kernel receives such a packet
180 and the 'Out' counter means kernel sends such a packet.
183 They are IcmpMsgInType[N] and IcmpMsgOutType[N], the [N] indicates the
184 ICMP type number. These counters track all kinds of ICMP packets. The
185 ICMP type number definition could be found in the `ICMP parameters`_
188 .. _ICMP parameters: https://www.iana.org/assignments/icmp-parameters/icmp-parameters.xhtml
190 For example, if the Linux kernel sends an ICMP Echo packet, the
191 IcmpMsgOutType8 would increase 1. And if kernel gets an ICMP Echo Reply
192 packet, IcmpMsgInType0 would increase 1.
195 This counter indicates the checksum of the ICMP packet is
196 wrong. Kernel verifies the checksum after updating the IcmpInMsgs and
197 before updating IcmpMsgInType[N]. If a packet has bad checksum, the
198 IcmpInMsgs would be updated but none of IcmpMsgInType[N] would be updated.
200 * IcmpInErrors and IcmpOutErrors
201 Defined by `RFC1213 icmpInErrors`_ and `RFC1213 icmpOutErrors`_
203 .. _RFC1213 icmpInErrors: https://tools.ietf.org/html/rfc1213#page-41
204 .. _RFC1213 icmpOutErrors: https://tools.ietf.org/html/rfc1213#page-43
206 When an error occurs in the ICMP packet handler path, these two
207 counters would be updated. The receiving packet path use IcmpInErrors
208 and the sending packet path use IcmpOutErrors. When IcmpInCsumErrors
209 is increased, IcmpInErrors would always be increased too.
211 relationship of the ICMP counters
212 -------------------------------
213 The sum of IcmpMsgOutType[N] is always equal to IcmpOutMsgs, as they
214 are updated at the same time. The sum of IcmpMsgInType[N] plus
215 IcmpInErrors should be equal or larger than IcmpInMsgs. When kernel
216 receives an ICMP packet, kernel follows below logic:
218 1. increase IcmpInMsgs
219 2. if has any error, update IcmpInErrors and finish the process
220 3. update IcmpMsgOutType[N]
221 4. handle the packet depending on the type, if has any error, update
222 IcmpInErrors and finish the process
224 So if all errors occur in step (2), IcmpInMsgs should be equal to the
225 sum of IcmpMsgOutType[N] plus IcmpInErrors. If all errors occur in
226 step (4), IcmpInMsgs should be equal to the sum of
227 IcmpMsgOutType[N]. If the errors occur in both step (2) and step (4),
228 IcmpInMsgs should be less than the sum of IcmpMsgOutType[N] plus
234 Defined in `RFC1213 tcpInSegs`_
236 .. _RFC1213 tcpInSegs: https://tools.ietf.org/html/rfc1213#page-48
238 The number of packets received by the TCP layer. As mentioned in
239 RFC1213, it includes the packets received in error, such as checksum
240 error, invalid TCP header and so on. Only one error won't be included:
241 if the layer 2 destination address is not the NIC's layer 2
242 address. It might happen if the packet is a multicast or broadcast
243 packet, or the NIC is in promiscuous mode. In these situations, the
244 packets would be delivered to the TCP layer, but the TCP layer will discard
245 these packets before increasing TcpInSegs. The TcpInSegs counter
246 isn't aware of GRO. So if two packets are merged by GRO, the TcpInSegs
247 counter would only increase 1.
250 Defined in `RFC1213 tcpOutSegs`_
252 .. _RFC1213 tcpOutSegs: https://tools.ietf.org/html/rfc1213#page-48
254 The number of packets sent by the TCP layer. As mentioned in RFC1213,
255 it excludes the retransmitted packets. But it includes the SYN, ACK
256 and RST packets. Doesn't like TcpInSegs, the TcpOutSegs is aware of
257 GSO, so if a packet would be split to 2 by GSO, TcpOutSegs will
261 Defined in `RFC1213 tcpActiveOpens`_
263 .. _RFC1213 tcpActiveOpens: https://tools.ietf.org/html/rfc1213#page-47
265 It means the TCP layer sends a SYN, and come into the SYN-SENT
266 state. Every time TcpActiveOpens increases 1, TcpOutSegs should always
270 Defined in `RFC1213 tcpPassiveOpens`_
272 .. _RFC1213 tcpPassiveOpens: https://tools.ietf.org/html/rfc1213#page-47
274 It means the TCP layer receives a SYN, replies a SYN+ACK, come into
277 * TcpExtTCPRcvCoalesce
278 When packets are received by the TCP layer and are not be read by the
279 application, the TCP layer will try to merge them. This counter
280 indicate how many packets are merged in such situation. If GRO is
281 enabled, lots of packets would be merged by GRO, these packets
282 wouldn't be counted to TcpExtTCPRcvCoalesce.
284 * TcpExtTCPAutoCorking
285 When sending packets, the TCP layer will try to merge small packets to
286 a bigger one. This counter increase 1 for every packet merged in such
287 situation. Please refer to the LWN article for more details:
288 https://lwn.net/Articles/576263/
290 * TcpExtTCPOrigDataSent
291 This counter is explained by `kernel commit f19c29e3e391`_, I pasted the
294 TCPOrigDataSent: number of outgoing packets with original data (excluding
295 retransmission but including data-in-SYN). This counter is different from
296 TcpOutSegs because TcpOutSegs also tracks pure ACKs. TCPOrigDataSent is
297 more useful to track the TCP retransmission rate.
300 This counter is explained by `kernel commit f19c29e3e391`_, I pasted the
303 TCPSynRetrans: number of SYN and SYN/ACK retransmits to break down
304 retransmissions into SYN, fast-retransmits, timeout retransmits, etc.
306 * TCPFastOpenActiveFail
307 This counter is explained by `kernel commit f19c29e3e391`_, I pasted the
310 TCPFastOpenActiveFail: Fast Open attempts (SYN/data) failed because
311 the remote does not accept it or the attempts timed out.
313 .. _kernel commit f19c29e3e391: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=f19c29e3e391a66a273e9afebaf01917245148cd
315 * TcpExtListenOverflows and TcpExtListenDrops
316 When kernel receives a SYN from a client, and if the TCP accept queue
317 is full, kernel will drop the SYN and add 1 to TcpExtListenOverflows.
318 At the same time kernel will also add 1 to TcpExtListenDrops. When a
319 TCP socket is in LISTEN state, and kernel need to drop a packet,
320 kernel would always add 1 to TcpExtListenDrops. So increase
321 TcpExtListenOverflows would let TcpExtListenDrops increasing at the
322 same time, but TcpExtListenDrops would also increase without
323 TcpExtListenOverflows increasing, e.g. a memory allocation fail would
324 also let TcpExtListenDrops increase.
326 Note: The above explanation is based on kernel 4.10 or above version, on
327 an old kernel, the TCP stack has different behavior when TCP accept
328 queue is full. On the old kernel, TCP stack won't drop the SYN, it
329 would complete the 3-way handshake. As the accept queue is full, TCP
330 stack will keep the socket in the TCP half-open queue. As it is in the
331 half open queue, TCP stack will send SYN+ACK on an exponential backoff
332 timer, after client replies ACK, TCP stack checks whether the accept
333 queue is still full, if it is not full, moves the socket to the accept
334 queue, if it is full, keeps the socket in the half-open queue, at next
335 time client replies ACK, this socket will get another chance to move
340 Defined in `RFC1213 tcpEstabResets`_.
342 .. _RFC1213 tcpEstabResets: https://tools.ietf.org/html/rfc1213#page-48
345 Defined in `RFC1213 tcpAttemptFails`_.
347 .. _RFC1213 tcpAttemptFails: https://tools.ietf.org/html/rfc1213#page-48
350 Defined in `RFC1213 tcpOutRsts`_. The RFC says this counter indicates
351 the 'segments sent containing the RST flag', but in linux kernel, this
352 couner indicates the segments kerenl tried to send. The sending
353 process might be failed due to some errors (e.g. memory alloc failed).
355 .. _RFC1213 tcpOutRsts: https://tools.ietf.org/html/rfc1213#page-52
360 When kernel receives a TCP packet, it has two paths to handler the
361 packet, one is fast path, another is slow path. The comment in kernel
362 code provides a good explanation of them, I pasted them below::
364 It is split into a fast path and a slow path. The fast path is
367 - A zero window was announced from us
368 - zero window probing
369 is only handled properly on the slow path.
370 - Out of order segments arrived.
371 - Urgent data is expected.
372 - There is no buffer space left
373 - Unexpected TCP flags/window values/header lengths are received
374 (detected by checking the TCP header against pred_flags)
375 - Data is sent in both directions. The fast path only supports pure senders
376 or pure receivers (this means either the sequence number or the ack
377 value must stay constant)
378 - Unexpected TCP option.
380 Kernel will try to use fast path unless any of the above conditions
381 are satisfied. If the packets are out of order, kernel will handle
382 them in slow path, which means the performance might be not very
383 good. Kernel would also come into slow path if the "Delayed ack" is
384 used, because when using "Delayed ack", the data is sent in both
385 directions. When the TCP window scale option is not used, kernel will
386 try to enable fast path immediately when the connection comes into the
387 established state, but if the TCP window scale option is used, kernel
388 will disable the fast path at first, and try to enable it after kernel
391 * TcpExtTCPPureAcks and TcpExtTCPHPAcks
392 If a packet set ACK flag and has no data, it is a pure ACK packet, if
393 kernel handles it in the fast path, TcpExtTCPHPAcks will increase 1,
394 if kernel handles it in the slow path, TcpExtTCPPureAcks will
398 If a TCP packet has data (which means it is not a pure ACK packet),
399 and this packet is handled in the fast path, TcpExtTCPHPHits will
405 * TcpExtTCPAbortOnData
406 It means TCP layer has data in flight, but need to close the
407 connection. So TCP layer sends a RST to the other side, indicate the
408 connection is not closed very graceful. An easy way to increase this
409 counter is using the SO_LINGER option. Please refer to the SO_LINGER
410 section of the `socket man page`_:
412 .. _socket man page: http://man7.org/linux/man-pages/man7/socket.7.html
414 By default, when an application closes a connection, the close function
415 will return immediately and kernel will try to send the in-flight data
416 async. If you use the SO_LINGER option, set l_onoff to 1, and l_linger
417 to a positive number, the close function won't return immediately, but
418 wait for the in-flight data are acked by the other side, the max wait
419 time is l_linger seconds. If set l_onoff to 1 and set l_linger to 0,
420 when the application closes a connection, kernel will send a RST
421 immediately and increase the TcpExtTCPAbortOnData counter.
423 * TcpExtTCPAbortOnClose
424 This counter means the application has unread data in the TCP layer when
425 the application wants to close the TCP connection. In such a situation,
426 kernel will send a RST to the other side of the TCP connection.
428 * TcpExtTCPAbortOnMemory
429 When an application closes a TCP connection, kernel still need to track
430 the connection, let it complete the TCP disconnect process. E.g. an
431 app calls the close method of a socket, kernel sends fin to the other
432 side of the connection, then the app has no relationship with the
433 socket any more, but kernel need to keep the socket, this socket
434 becomes an orphan socket, kernel waits for the reply of the other side,
435 and would come to the TIME_WAIT state finally. When kernel has no
436 enough memory to keep the orphan socket, kernel would send an RST to
437 the other side, and delete the socket, in such situation, kernel will
438 increase 1 to the TcpExtTCPAbortOnMemory. Two conditions would trigger
439 TcpExtTCPAbortOnMemory:
441 1. the memory used by the TCP protocol is higher than the third value of
442 the tcp_mem. Please refer the tcp_mem section in the `TCP man page`_:
444 .. _TCP man page: http://man7.org/linux/man-pages/man7/tcp.7.html
446 2. the orphan socket count is higher than net.ipv4.tcp_max_orphans
449 * TcpExtTCPAbortOnTimeout
450 This counter will increase when any of the TCP timers expire. In such
451 situation, kernel won't send RST, just give up the connection.
453 * TcpExtTCPAbortOnLinger
454 When a TCP connection comes into FIN_WAIT_2 state, instead of waiting
455 for the fin packet from the other side, kernel could send a RST and
456 delete the socket immediately. This is not the default behavior of
457 Linux kernel TCP stack. By configuring the TCP_LINGER2 socket option,
458 you could let kernel follow this behavior.
460 * TcpExtTCPAbortFailed
461 The kernel TCP layer will send RST if the `RFC2525 2.17 section`_ is
462 satisfied. If an internal error occurs during this process,
463 TcpExtTCPAbortFailed will be increased.
465 .. _RFC2525 2.17 section: https://tools.ietf.org/html/rfc2525#page-50
467 TCP Hybrid Slow Start
469 The Hybrid Slow Start algorithm is an enhancement of the traditional
470 TCP congestion window Slow Start algorithm. It uses two pieces of
471 information to detect whether the max bandwidth of the TCP path is
472 approached. The two pieces of information are ACK train length and
473 increase in packet delay. For detail information, please refer the
474 `Hybrid Slow Start paper`_. Either ACK train length or packet delay
475 hits a specific threshold, the congestion control algorithm will come
476 into the Congestion Avoidance state. Until v4.20, two congestion
477 control algorithms are using Hybrid Slow Start, they are cubic (the
478 default congestion control algorithm) and cdg. Four snmp counters
479 relate with the Hybrid Slow Start algorithm.
481 .. _Hybrid Slow Start paper: https://pdfs.semanticscholar.org/25e9/ef3f03315782c7f1cbcd31b587857adae7d1.pdf
483 * TcpExtTCPHystartTrainDetect
484 How many times the ACK train length threshold is detected
486 * TcpExtTCPHystartTrainCwnd
487 The sum of CWND detected by ACK train length. Dividing this value by
488 TcpExtTCPHystartTrainDetect is the average CWND which detected by the
491 * TcpExtTCPHystartDelayDetect
492 How many times the packet delay threshold is detected.
494 * TcpExtTCPHystartDelayCwnd
495 The sum of CWND detected by packet delay. Dividing this value by
496 TcpExtTCPHystartDelayDetect is the average CWND which detected by the
499 TCP retransmission and congestion control
500 ======================================
501 The TCP protocol has two retransmission mechanisms: SACK and fast
502 recovery. They are exclusive with each other. When SACK is enabled,
503 the kernel TCP stack would use SACK, or kernel would use fast
504 recovery. The SACK is a TCP option, which is defined in `RFC2018`_,
505 the fast recovery is defined in `RFC6582`_, which is also called
508 The TCP congestion control is a big and complex topic. To understand
509 the related snmp counter, we need to know the states of the congestion
510 control state machine. There are 5 states: Open, Disorder, CWR,
511 Recovery and Loss. For details about these states, please refer page 5
512 and page 6 of this document:
513 https://pdfs.semanticscholar.org/0e9c/968d09ab2e53e24c4dca5b2d67c7f7140f8e.pdf
515 .. _RFC2018: https://tools.ietf.org/html/rfc2018
516 .. _RFC6582: https://tools.ietf.org/html/rfc6582
518 * TcpExtTCPRenoRecovery and TcpExtTCPSackRecovery
519 When the congestion control comes into Recovery state, if sack is
520 used, TcpExtTCPSackRecovery increases 1, if sack is not used,
521 TcpExtTCPRenoRecovery increases 1. These two counters mean the TCP
522 stack begins to retransmit the lost packets.
524 * TcpExtTCPSACKReneging
525 A packet was acknowledged by SACK, but the receiver has dropped this
526 packet, so the sender needs to retransmit this packet. In this
527 situation, the sender adds 1 to TcpExtTCPSACKReneging. A receiver
528 could drop a packet which has been acknowledged by SACK, although it is
529 unusual, it is allowed by the TCP protocol. The sender doesn't really
530 know what happened on the receiver side. The sender just waits until
531 the RTO expires for this packet, then the sender assumes this packet
532 has been dropped by the receiver.
534 * TcpExtTCPRenoReorder
535 The reorder packet is detected by fast recovery. It would only be used
536 if SACK is disabled. The fast recovery algorithm detects recorder by
537 the duplicate ACK number. E.g., if retransmission is triggered, and
538 the original retransmitted packet is not lost, it is just out of
539 order, the receiver would acknowledge multiple times, one for the
540 retransmitted packet, another for the arriving of the original out of
541 order packet. Thus the sender would find more ACks than its
542 expectation, and the sender knows out of order occurs.
545 The reorder packet is detected when a hole is filled. E.g., assume the
546 sender sends packet 1,2,3,4,5, and the receiving order is
547 1,2,4,5,3. When the sender receives the ACK of packet 3 (which will
548 fill the hole), two conditions will let TcpExtTCPTSReorder increase
549 1: (1) if the packet 3 is not re-retransmitted yet. (2) if the packet
550 3 is retransmitted but the timestamp of the packet 3's ACK is earlier
551 than the retransmission timestamp.
553 * TcpExtTCPSACKReorder
554 The reorder packet detected by SACK. The SACK has two methods to
555 detect reorder: (1) DSACK is received by the sender. It means the
556 sender sends the same packet more than one times. And the only reason
557 is the sender believes an out of order packet is lost so it sends the
558 packet again. (2) Assume packet 1,2,3,4,5 are sent by the sender, and
559 the sender has received SACKs for packet 2 and 5, now the sender
560 receives SACK for packet 4 and the sender doesn't retransmit the
561 packet yet, the sender would know packet 4 is out of order. The TCP
562 stack of kernel will increase TcpExtTCPSACKReorder for both of the
567 The DSACK is defined in `RFC2883`_. The receiver uses DSACK to report
568 duplicate packets to the sender. There are two kinds of
569 duplications: (1) a packet which has been acknowledged is
570 duplicate. (2) an out of order packet is duplicate. The TCP stack
571 counts these two kinds of duplications on both receiver side and
574 .. _RFC2883 : https://tools.ietf.org/html/rfc2883
576 * TcpExtTCPDSACKOldSent
577 The TCP stack receives a duplicate packet which has been acked, so it
578 sends a DSACK to the sender.
580 * TcpExtTCPDSACKOfoSent
581 The TCP stack receives an out of order duplicate packet, so it sends a
585 The TCP stack receives a DSACK, which indicates an acknowledged
586 duplicate packet is received.
588 * TcpExtTCPDSACKOfoRecv
589 The TCP stack receives a DSACK, which indicate an out of order
590 duplicate packet is received.
592 invalid SACK and DSACK
594 When a SACK (or DSACK) block is invalid, a corresponding counter would
595 be updated. The validation method is base on the start/end sequence
596 number of the SACK block. For more details, please refer the comment
597 of the function tcp_is_sackblock_valid in the kernel source code. A
598 SACK option could have up to 4 blocks, they are checked
599 individually. E.g., if 3 blocks of a SACk is invalid, the
600 corresponding counter would be updated 3 times. The comment of the
601 `Add counters for discarded SACK blocks`_ patch has additional
604 .. _Add counters for discarded SACK blocks: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=18f02545a9a16c9a89778b91a162ad16d510bb32
606 * TcpExtTCPSACKDiscard
607 This counter indicates how many SACK blocks are invalid. If the invalid
608 SACK block is caused by ACK recording, the TCP stack will only ignore
609 it and won't update this counter.
611 * TcpExtTCPDSACKIgnoredOld and TcpExtTCPDSACKIgnoredNoUndo
612 When a DSACK block is invalid, one of these two counters would be
613 updated. Which counter will be updated depends on the undo_marker flag
614 of the TCP socket. If the undo_marker is not set, the TCP stack isn't
615 likely to re-transmit any packets, and we still receive an invalid
616 DSACK block, the reason might be that the packet is duplicated in the
617 middle of the network. In such scenario, TcpExtTCPDSACKIgnoredNoUndo
618 will be updated. If the undo_marker is set, TcpExtTCPDSACKIgnoredOld
619 will be updated. As implied in its name, it might be an old packet.
623 The linux networking stack stores data in sk_buff struct (skb for
624 short). If a SACK block acrosses multiple skb, the TCP stack will try
625 to re-arrange data in these skb. E.g. if a SACK block acknowledges seq
626 10 to 15, skb1 has seq 10 to 13, skb2 has seq 14 to 20. The seq 14 and
627 15 in skb2 would be moved to skb1. This operation is 'shift'. If a
628 SACK block acknowledges seq 10 to 20, skb1 has seq 10 to 13, skb2 has
629 seq 14 to 20. All data in skb2 will be moved to skb1, and skb2 will be
630 discard, this operation is 'merge'.
632 * TcpExtTCPSackShifted
635 * TcpExtTCPSackMerged
638 * TcpExtTCPSackShiftFallback
639 A skb should be shifted or merged, but the TCP stack doesn't do it for
645 The TCP layer receives an out of order packet and has enough memory
649 The TCP layer receives an out of order packet but doesn't have enough
650 memory, so drops it. Such packets won't be counted into
654 The received out of order packet has an overlay with the previous
655 packet. the overlay part will be dropped. All of TcpExtTCPOFOMerge
656 packets will also be counted into TcpExtTCPOFOQueue.
660 PAWS (Protection Against Wrapped Sequence numbers) is an algorithm
661 which is used to drop old packets. It depends on the TCP
662 timestamps. For detail information, please refer the `timestamp wiki`_
663 and the `RFC of PAWS`_.
665 .. _RFC of PAWS: https://tools.ietf.org/html/rfc1323#page-17
666 .. _timestamp wiki: https://en.wikipedia.org/wiki/Transmission_Control_Protocol#TCP_timestamps
669 Packets are dropped by PAWS in Syn-Sent status.
672 Packets are dropped by PAWS in any status other than Syn-Sent.
676 In some scenarios, kernel would avoid sending duplicate ACKs too
677 frequently. Please find more details in the tcp_invalid_ratelimit
678 section of the `sysctl document`_. When kernel decides to skip an ACK
679 due to tcp_invalid_ratelimit, kernel would update one of below
680 counters to indicate the ACK is skipped in which scenario. The ACK
681 would only be skipped if the received packet is either a SYN packet or
684 .. _sysctl document: https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt
686 * TcpExtTCPACKSkippedSynRecv
687 The ACK is skipped in Syn-Recv status. The Syn-Recv status means the
688 TCP stack receives a SYN and replies SYN+ACK. Now the TCP stack is
689 waiting for an ACK. Generally, the TCP stack doesn't need to send ACK
690 in the Syn-Recv status. But in several scenarios, the TCP stack need
691 to send an ACK. E.g., the TCP stack receives the same SYN packet
692 repeately, the received packet does not pass the PAWS check, or the
693 received packet sequence number is out of window. In these scenarios,
694 the TCP stack needs to send ACK. If the ACk sending frequency is higher than
695 tcp_invalid_ratelimit allows, the TCP stack will skip sending ACK and
696 increase TcpExtTCPACKSkippedSynRecv.
699 * TcpExtTCPACKSkippedPAWS
700 The ACK is skipped due to PAWS (Protect Against Wrapped Sequence
701 numbers) check fails. If the PAWS check fails in Syn-Recv, Fin-Wait-2
702 or Time-Wait statuses, the skipped ACK would be counted to
703 TcpExtTCPACKSkippedSynRecv, TcpExtTCPACKSkippedFinWait2 or
704 TcpExtTCPACKSkippedTimeWait. In all other statuses, the skipped ACK
705 would be counted to TcpExtTCPACKSkippedPAWS.
707 * TcpExtTCPACKSkippedSeq
708 The sequence number is out of window and the timestamp passes the PAWS
709 check and the TCP status is not Syn-Recv, Fin-Wait-2, and Time-Wait.
711 * TcpExtTCPACKSkippedFinWait2
712 The ACK is skipped in Fin-Wait-2 status, the reason would be either
713 PAWS check fails or the received sequence number is out of window.
715 * TcpExtTCPACKSkippedTimeWait
716 Tha ACK is skipped in Time-Wait status, the reason would be either
717 PAWS check failed or the received sequence number is out of window.
719 * TcpExtTCPACKSkippedChallenge
720 The ACK is skipped if the ACK is a challenge ACK. The RFC 5961 defines
721 3 kind of challenge ACK, please refer `RFC 5961 section 3.2`_,
722 `RFC 5961 section 4.2`_ and `RFC 5961 section 5.2`_. Besides these
723 three scenarios, In some TCP status, the linux TCP stack would also
724 send challenge ACKs if the ACK number is before the first
725 unacknowledged number (more strict than `RFC 5961 section 5.2`_).
727 .. _RFC 5961 section 3.2: https://tools.ietf.org/html/rfc5961#page-7
728 .. _RFC 5961 section 4.2: https://tools.ietf.org/html/rfc5961#page-9
729 .. _RFC 5961 section 5.2: https://tools.ietf.org/html/rfc5961#page-11
733 * TcpExtTCPWantZeroWindowAdv
734 Depending on current memory usage, the TCP stack tries to set receive
735 window to zero. But the receive window might still be a no-zero
736 value. For example, if the previous window size is 10, and the TCP
737 stack receives 3 bytes, the current window size would be 7 even if the
738 window size calculated by the memory usage is zero.
740 * TcpExtTCPToZeroWindowAdv
741 The TCP receive window is set to zero from a no-zero value.
743 * TcpExtTCPFromZeroWindowAdv
744 The TCP receive window is set to no-zero value from zero.
749 The TCP Delayed ACK is a technique which is used for reducing the
750 packet count in the network. For more details, please refer the
753 .. _Delayed ACK wiki: https://en.wikipedia.org/wiki/TCP_delayed_acknowledgment
756 A delayed ACK timer expires. The TCP stack will send a pure ACK packet
757 and exit the delayed ACK mode.
759 * TcpExtDelayedACKLocked
760 A delayed ACK timer expires, but the TCP stack can't send an ACK
761 immediately due to the socket is locked by a userspace program. The
762 TCP stack will send a pure ACK later (after the userspace program
763 unlock the socket). When the TCP stack sends the pure ACK later, the
764 TCP stack will also update TcpExtDelayedACKs and exit the delayed ACK
767 * TcpExtDelayedACKLost
768 It will be updated when the TCP stack receives a packet which has been
769 ACKed. A Delayed ACK loss might cause this issue, but it would also be
770 triggered by other reasons, such as a packet is duplicated in the
773 Tail Loss Probe (TLP)
775 TLP is an algorithm which is used to detect TCP packet loss. For more
776 details, please refer the `TLP paper`_.
778 .. _TLP paper: https://tools.ietf.org/html/draft-dukkipati-tcpm-tcp-loss-probe-01
780 * TcpExtTCPLossProbes
781 A TLP probe packet is sent.
783 * TcpExtTCPLossProbeRecovery
784 A packet loss is detected and recovered by TLP.
791 Run the ping command against the public dns server 8.8.8.8::
793 nstatuser@nstat-a:~$ ping 8.8.8.8 -c 1
794 PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
795 64 bytes from 8.8.8.8: icmp_seq=1 ttl=119 time=17.8 ms
797 --- 8.8.8.8 ping statistics ---
798 1 packets transmitted, 1 received, 0% packet loss, time 0ms
799 rtt min/avg/max/mdev = 17.875/17.875/17.875/0.000 ms
803 nstatuser@nstat-a:~$ nstat
813 IcmpMsgOutType8 1 0.0
815 IpExtOutOctets 84 0.0
816 IpExtInNoECTPkts 1 0.0
818 The Linux server sent an ICMP Echo packet, so IpOutRequests,
819 IcmpOutMsgs, IcmpOutEchos and IcmpMsgOutType8 were increased 1. The
820 server got ICMP Echo Reply from 8.8.8.8, so IpInReceives, IcmpInMsgs,
821 IcmpInEchoReps and IcmpMsgInType0 were increased 1. The ICMP Echo Reply
822 was passed to the ICMP layer via IP layer, so IpInDelivers was
823 increased 1. The default ping data size is 48, so an ICMP Echo packet
824 and its corresponding Echo Reply packet are constructed by:
826 * 14 bytes MAC header
828 * 16 bytes ICMP header
829 * 48 bytes data (default value of the ping command)
831 So the IpExtInOctets and IpExtOutOctets are 20+16+48=84.
835 On server side, we run::
837 nstatuser@nstat-b:~$ nc -lknv 0.0.0.0 9000
838 Listening on [0.0.0.0] (family 0, port 9000)
840 On client side, we run::
842 nstatuser@nstat-a:~$ nc -nv 192.168.122.251 9000
843 Connection to 192.168.122.251 9000 port [tcp/*] succeeded!
845 The server listened on tcp 9000 port, the client connected to it, they
846 completed the 3-way handshake.
848 On server side, we can find below nstat output::
850 nstatuser@nstat-b:~$ nstat | grep -i tcp
851 TcpPassiveOpens 1 0.0
854 TcpExtTCPPureAcks 1 0.0
856 On client side, we can find below nstat output::
858 nstatuser@nstat-a:~$ nstat | grep -i tcp
863 When the server received the first SYN, it replied a SYN+ACK, and came into
864 SYN-RCVD state, so TcpPassiveOpens increased 1. The server received
865 SYN, sent SYN+ACK, received ACK, so server sent 1 packet, received 2
866 packets, TcpInSegs increased 2, TcpOutSegs increased 1. The last ACK
867 of the 3-way handshake is a pure ACK without data, so
868 TcpExtTCPPureAcks increased 1.
870 When the client sent SYN, the client came into the SYN-SENT state, so
871 TcpActiveOpens increased 1, the client sent SYN, received SYN+ACK, sent
872 ACK, so client sent 2 packets, received 1 packet, TcpInSegs increased
873 1, TcpOutSegs increased 2.
879 nstatuser@nstat-b:~$ nc -lkv 0.0.0.0 9000
880 Listening on [0.0.0.0] (family 0, port 9000)
884 nstatuser@nstat-a:~$ nc -v nstat-b 9000
885 Connection to nstat-b 9000 port [tcp/*] succeeded!
887 Input a string in the nc client ('hello' in our example)::
889 nstatuser@nstat-a:~$ nc -v nstat-b 9000
890 Connection to nstat-b 9000 port [tcp/*] succeeded!
893 The client side nstat output::
895 nstatuser@nstat-a:~$ nstat
902 TcpExtTCPPureAcks 1 0.0
903 TcpExtTCPOrigDataSent 1 0.0
905 IpExtOutOctets 58 0.0
906 IpExtInNoECTPkts 1 0.0
908 The server side nstat output::
910 nstatuser@nstat-b:~$ nstat
918 IpExtOutOctets 52 0.0
919 IpExtInNoECTPkts 1 0.0
921 Input a string in nc client side again ('world' in our exmaple)::
923 nstatuser@nstat-a:~$ nc -v nstat-b 9000
924 Connection to nstat-b 9000 port [tcp/*] succeeded!
928 Client side nstat output::
930 nstatuser@nstat-a:~$ nstat
937 TcpExtTCPHPAcks 1 0.0
938 TcpExtTCPOrigDataSent 1 0.0
940 IpExtOutOctets 58 0.0
941 IpExtInNoECTPkts 1 0.0
944 Server side nstat output::
946 nstatuser@nstat-b:~$ nstat
953 TcpExtTCPHPHits 1 0.0
955 IpExtOutOctets 52 0.0
956 IpExtInNoECTPkts 1 0.0
958 Compare the first client-side nstat and the second client-side nstat,
959 we could find one difference: the first one had a 'TcpExtTCPPureAcks',
960 but the second one had a 'TcpExtTCPHPAcks'. The first server-side
961 nstat and the second server-side nstat had a difference too: the
962 second server-side nstat had a TcpExtTCPHPHits, but the first
963 server-side nstat didn't have it. The network traffic patterns were
964 exactly the same: the client sent a packet to the server, the server
965 replied an ACK. But kernel handled them in different ways. When the
966 TCP window scale option is not used, kernel will try to enable fast
967 path immediately when the connection comes into the established state,
968 but if the TCP window scale option is used, kernel will disable the
969 fast path at first, and try to enable it after kerenl receives
970 packets. We could use the 'ss' command to verify whether the window
971 scale option is used. e.g. run below command on either server or
974 nstatuser@nstat-a:~$ ss -o state established -i '( dport = :9000 or sport = :9000 )
975 Netid Recv-Q Send-Q Local Address:Port Peer Address:Port
976 tcp 0 0 192.168.122.250:40654 192.168.122.251:9000
977 ts sack cubic wscale:7,7 rto:204 rtt:0.98/0.49 mss:1448 pmtu:1500 rcvmss:536 advmss:1448 cwnd:10 bytes_acked:1 segs_out:2 segs_in:1 send 118.2Mbps lastsnd:46572 lastrcv:46572 lastack:46572 pacing_rate 236.4Mbps rcv_space:29200 rcv_ssthresh:29200 minrtt:0.98
979 The 'wscale:7,7' means both server and client set the window scale
980 option to 7. Now we could explain the nstat output in our test:
982 In the first nstat output of client side, the client sent a packet, server
983 reply an ACK, when kernel handled this ACK, the fast path was not
984 enabled, so the ACK was counted into 'TcpExtTCPPureAcks'.
986 In the second nstat output of client side, the client sent a packet again,
987 and received another ACK from the server, in this time, the fast path is
988 enabled, and the ACK was qualified for fast path, so it was handled by
989 the fast path, so this ACK was counted into TcpExtTCPHPAcks.
991 In the first nstat output of server side, fast path was not enabled,
992 so there was no 'TcpExtTCPHPHits'.
994 In the second nstat output of server side, the fast path was enabled,
995 and the packet received from client qualified for fast path, so it
996 was counted into 'TcpExtTCPHPHits'.
998 TcpExtTCPAbortOnClose
1000 On the server side, we run below python script::
1007 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
1008 s.bind(('0.0.0.0', port))
1010 sock, addr = s.accept()
1014 This python script listen on 9000 port, but doesn't read anything from
1017 On the client side, we send the string "hello" by nc::
1019 nstatuser@nstat-a:~$ echo "hello" | nc nstat-b 9000
1021 Then, we come back to the server side, the server has received the "hello"
1022 packet, and the TCP layer has acked this packet, but the application didn't
1023 read it yet. We type Ctrl-C to terminate the server script. Then we
1024 could find TcpExtTCPAbortOnClose increased 1 on the server side::
1026 nstatuser@nstat-b:~$ nstat | grep -i abort
1027 TcpExtTCPAbortOnClose 1 0.0
1029 If we run tcpdump on the server side, we could find the server sent a
1030 RST after we type Ctrl-C.
1032 TcpExtTCPAbortOnMemory and TcpExtTCPAbortOnTimeout
1033 -----------------------------------------------
1034 Below is an example which let the orphan socket count be higher than
1035 net.ipv4.tcp_max_orphans.
1036 Change tcp_max_orphans to a smaller value on client::
1038 sudo bash -c "echo 10 > /proc/sys/net/ipv4/tcp_max_orphans"
1040 Client code (create 64 connection to server)::
1042 nstatuser@nstat-a:~$ cat client_orphan.py
1046 server = 'nstat-b' # server address
1051 connection_list = []
1054 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
1055 s.connect((server, port))
1056 connection_list.append(s)
1057 print("connection_count: %d" % len(connection_list))
1062 Server code (accept 64 connection from client)::
1064 nstatuser@nstat-b:~$ cat server_orphan.py
1071 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
1072 s.bind(('0.0.0.0', port))
1074 connection_list = []
1076 sock, addr = s.accept()
1077 connection_list.append((sock, addr))
1078 print("connection_count: %d" % len(connection_list))
1080 Run the python scripts on server and client.
1084 python3 server_orphan.py
1088 python3 client_orphan.py
1090 Run iptables on server::
1092 sudo iptables -A INPUT -i ens3 -p tcp --destination-port 9000 -j DROP
1094 Type Ctrl-C on client, stop client_orphan.py.
1096 Check TcpExtTCPAbortOnMemory on client::
1098 nstatuser@nstat-a:~$ nstat | grep -i abort
1099 TcpExtTCPAbortOnMemory 54 0.0
1101 Check orphane socket count on client::
1103 nstatuser@nstat-a:~$ ss -s
1104 Total: 131 (kernel 0)
1105 TCP: 14 (estab 1, closed 0, orphaned 10, synrecv 0, timewait 0/0), ports 0
1107 Transport Total IP IPv6
1115 The explanation of the test: after run server_orphan.py and
1116 client_orphan.py, we set up 64 connections between server and
1117 client. Run the iptables command, the server will drop all packets from
1118 the client, type Ctrl-C on client_orphan.py, the system of the client
1119 would try to close these connections, and before they are closed
1120 gracefully, these connections became orphan sockets. As the iptables
1121 of the server blocked packets from the client, the server won't receive fin
1122 from the client, so all connection on clients would be stuck on FIN_WAIT_1
1123 stage, so they will keep as orphan sockets until timeout. We have echo
1124 10 to /proc/sys/net/ipv4/tcp_max_orphans, so the client system would
1125 only keep 10 orphan sockets, for all other orphan sockets, the client
1126 system sent RST for them and delete them. We have 64 connections, so
1127 the 'ss -s' command shows the system has 10 orphan sockets, and the
1128 value of TcpExtTCPAbortOnMemory was 54.
1130 An additional explanation about orphan socket count: You could find the
1131 exactly orphan socket count by the 'ss -s' command, but when kernel
1132 decide whither increases TcpExtTCPAbortOnMemory and sends RST, kernel
1133 doesn't always check the exactly orphan socket count. For increasing
1134 performance, kernel checks an approximate count firstly, if the
1135 approximate count is more than tcp_max_orphans, kernel checks the
1136 exact count again. So if the approximate count is less than
1137 tcp_max_orphans, but exactly count is more than tcp_max_orphans, you
1138 would find TcpExtTCPAbortOnMemory is not increased at all. If
1139 tcp_max_orphans is large enough, it won't occur, but if you decrease
1140 tcp_max_orphans to a small value like our test, you might find this
1141 issue. So in our test, the client set up 64 connections although the
1142 tcp_max_orphans is 10. If the client only set up 11 connections, we
1143 can't find the change of TcpExtTCPAbortOnMemory.
1145 Continue the previous test, we wait for several minutes. Because of the
1146 iptables on the server blocked the traffic, the server wouldn't receive
1147 fin, and all the client's orphan sockets would timeout on the
1148 FIN_WAIT_1 state finally. So we wait for a few minutes, we could find
1149 10 timeout on the client::
1151 nstatuser@nstat-a:~$ nstat | grep -i abort
1152 TcpExtTCPAbortOnTimeout 10 0.0
1154 TcpExtTCPAbortOnLinger
1155 ---------------------
1156 The server side code::
1158 nstatuser@nstat-b:~$ cat server_linger.py
1164 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
1165 s.bind(('0.0.0.0', port))
1167 sock, addr = s.accept()
1171 The client side code::
1173 nstatuser@nstat-a:~$ cat client_linger.py
1177 server = 'nstat-b' # server address
1180 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
1181 s.setsockopt(socket.SOL_SOCKET, socket.SO_LINGER, struct.pack('ii', 1, 10))
1182 s.setsockopt(socket.SOL_TCP, socket.TCP_LINGER2, struct.pack('i', -1))
1183 s.connect((server, port))
1186 Run server_linger.py on server::
1188 nstatuser@nstat-b:~$ python3 server_linger.py
1190 Run client_linger.py on client::
1192 nstatuser@nstat-a:~$ python3 client_linger.py
1194 After run client_linger.py, check the output of nstat::
1196 nstatuser@nstat-a:~$ nstat | grep -i abort
1197 TcpExtTCPAbortOnLinger 1 0.0
1199 TcpExtTCPRcvCoalesce
1201 On the server, we run a program which listen on TCP port 9000, but
1202 doesn't read any data::
1207 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
1208 s.bind(('0.0.0.0', port))
1210 sock, addr = s.accept()
1214 Save the above code as server_coalesce.py, and run::
1216 python3 server_coalesce.py
1218 On the client, save below code as client_coalesce.py::
1223 s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
1224 s.connect((server, port))
1228 nstatuser@nstat-a:~$ python3 -i client_coalesce.py
1230 We use '-i' to come into the interactive mode, then a packet::
1235 Send a packet again::
1240 On the server, run nstat::
1242 ubuntu@nstat-b:~$ nstat
1249 TcpExtTCPRcvCoalesce 1 0.0
1250 IpExtInOctets 110 0.0
1251 IpExtOutOctets 104 0.0
1252 IpExtInNoECTPkts 2 0.0
1254 The client sent two packets, server didn't read any data. When
1255 the second packet arrived at server, the first packet was still in
1256 the receiving queue. So the TCP layer merged the two packets, and we
1257 could find the TcpExtTCPRcvCoalesce increased 1.
1259 TcpExtListenOverflows and TcpExtListenDrops
1260 ----------------------------------------
1261 On server, run the nc command, listen on port 9000::
1263 nstatuser@nstat-b:~$ nc -lkv 0.0.0.0 9000
1264 Listening on [0.0.0.0] (family 0, port 9000)
1266 On client, run 3 nc commands in different terminals::
1268 nstatuser@nstat-a:~$ nc -v nstat-b 9000
1269 Connection to nstat-b 9000 port [tcp/*] succeeded!
1271 The nc command only accepts 1 connection, and the accept queue length
1272 is 1. On current linux implementation, set queue length to n means the
1273 actual queue length is n+1. Now we create 3 connections, 1 is accepted
1274 by nc, 2 in accepted queue, so the accept queue is full.
1276 Before running the 4th nc, we clean the nstat history on the server::
1278 nstatuser@nstat-b:~$ nstat -n
1280 Run the 4th nc on the client::
1282 nstatuser@nstat-a:~$ nc -v nstat-b 9000
1284 If the nc server is running on kernel 4.10 or higher version, you
1285 won't see the "Connection to ... succeeded!" string, because kernel
1286 will drop the SYN if the accept queue is full. If the nc client is running
1287 on an old kernel, you would see that the connection is succeeded,
1288 because kernel would complete the 3 way handshake and keep the socket
1289 on half open queue. I did the test on kernel 4.15. Below is the nstat
1292 nstatuser@nstat-b:~$ nstat
1297 TcpExtListenOverflows 4 0.0
1298 TcpExtListenDrops 4 0.0
1299 IpExtInOctets 240 0.0
1300 IpExtInNoECTPkts 4 0.0
1302 Both TcpExtListenOverflows and TcpExtListenDrops were 4. If the time
1303 between the 4th nc and the nstat was longer, the value of
1304 TcpExtListenOverflows and TcpExtListenDrops would be larger, because
1305 the SYN of the 4th nc was dropped, the client was retrying.
1307 IpInAddrErrors, IpExtInNoRoutes and IpOutNoRoutes
1308 ----------------------------------------------
1309 server A IP address: 192.168.122.250
1310 server B IP address: 192.168.122.251
1311 Prepare on server A, add a route to server B::
1313 $ sudo ip route add 8.8.8.8/32 via 192.168.122.251
1315 Prepare on server B, disable send_redirects for all interfaces::
1317 $ sudo sysctl -w net.ipv4.conf.all.send_redirects=0
1318 $ sudo sysctl -w net.ipv4.conf.ens3.send_redirects=0
1319 $ sudo sysctl -w net.ipv4.conf.lo.send_redirects=0
1320 $ sudo sysctl -w net.ipv4.conf.default.send_redirects=0
1322 We want to let sever A send a packet to 8.8.8.8, and route the packet
1323 to server B. When server B receives such packet, it might send a ICMP
1324 Redirect message to server A, set send_redirects to 0 will disable
1327 First, generate InAddrErrors. On server B, we disable IP forwarding::
1329 $ sudo sysctl -w net.ipv4.conf.all.forwarding=0
1331 On server A, we send packets to 8.8.8.8::
1335 On server B, we check the output of nstat::
1340 IpInAddrErrors 3 0.0
1341 IpExtInOctets 180 0.0
1342 IpExtInNoECTPkts 3 0.0
1344 As we have let server A route 8.8.8.8 to server B, and we disabled IP
1345 forwarding on server B, Server A sent packets to server B, then server B
1346 dropped packets and increased IpInAddrErrors. As the nc command would
1347 re-send the SYN packet if it didn't receive a SYN+ACK, we could find
1348 multiple IpInAddrErrors.
1350 Second, generate IpExtInNoRoutes. On server B, we enable IP
1353 $ sudo sysctl -w net.ipv4.conf.all.forwarding=1
1355 Check the route table of server B and remove the default route::
1358 default via 192.168.122.1 dev ens3 proto static
1359 192.168.122.0/24 dev ens3 proto kernel scope link src 192.168.122.251
1360 $ sudo ip route delete default via 192.168.122.1 dev ens3 proto static
1362 On server A, we contact 8.8.8.8 again::
1365 nc: connect to 8.8.8.8 port 53 (tcp) failed: Network is unreachable
1367 On server B, run nstat::
1374 IcmpOutDestUnreachs 1 0.0
1375 IcmpMsgOutType3 1 0.0
1376 IpExtInNoRoutes 1 0.0
1377 IpExtInOctets 60 0.0
1378 IpExtOutOctets 88 0.0
1379 IpExtInNoECTPkts 1 0.0
1381 We enabled IP forwarding on server B, when server B received a packet
1382 which destination IP address is 8.8.8.8, server B will try to forward
1383 this packet. We have deleted the default route, there was no route for
1384 8.8.8.8, so server B increase IpExtInNoRoutes and sent the "ICMP
1385 Destination Unreachable" message to server A.
1387 Third, generate IpOutNoRoutes. Run ping command on server B::
1390 connect: Network is unreachable
1392 Run nstat on server B::
1398 We have deleted the default route on server B. Server B couldn't find
1399 a route for the 8.8.8.8 IP address, so server B increased
1402 TcpExtTCPACKSkippedSynRecv
1403 ------------------------
1404 In this test, we send 3 same SYN packets from client to server. The
1405 first SYN will let server create a socket, set it to Syn-Recv status,
1406 and reply a SYN/ACK. The second SYN will let server reply the SYN/ACK
1407 again, and record the reply time (the duplicate ACK reply time). The
1408 third SYN will let server check the previous duplicate ACK reply time,
1409 and decide to skip the duplicate ACK, then increase the
1410 TcpExtTCPACKSkippedSynRecv counter.
1412 Run tcpdump to capture a SYN packet::
1414 nstatuser@nstat-a:~$ sudo tcpdump -c 1 -w /tmp/syn.pcap port 9000
1415 tcpdump: listening on ens3, link-type EN10MB (Ethernet), capture size 262144 bytes
1417 Open another terminal, run nc command::
1419 nstatuser@nstat-a:~$ nc nstat-b 9000
1421 As the nstat-b didn't listen on port 9000, it should reply a RST, and
1422 the nc command exited immediately. It was enough for the tcpdump
1423 command to capture a SYN packet. A linux server might use hardware
1424 offload for the TCP checksum, so the checksum in the /tmp/syn.pcap
1425 might be not correct. We call tcprewrite to fix it::
1427 nstatuser@nstat-a:~$ tcprewrite --infile=/tmp/syn.pcap --outfile=/tmp/syn_fixcsum.pcap --fixcsum
1429 On nstat-b, we run nc to listen on port 9000::
1431 nstatuser@nstat-b:~$ nc -lkv 9000
1432 Listening on [0.0.0.0] (family 0, port 9000)
1434 On nstat-a, we blocked the packet from port 9000, or nstat-a would send
1437 nstatuser@nstat-a:~$ sudo iptables -A INPUT -p tcp --sport 9000 -j DROP
1439 Send 3 SYN repeatly to nstat-b::
1441 nstatuser@nstat-a:~$ for i in {1..3}; do sudo tcpreplay -i ens3 /tmp/syn_fixcsum.pcap; done
1443 Check snmp cunter on nstat-b::
1445 nstatuser@nstat-b:~$ nstat | grep -i skip
1446 TcpExtTCPACKSkippedSynRecv 1 0.0
1448 As we expected, TcpExtTCPACKSkippedSynRecv is 1.
1450 TcpExtTCPACKSkippedPAWS
1451 ----------------------
1452 To trigger PAWS, we could send an old SYN.
1454 On nstat-b, let nc listen on port 9000::
1456 nstatuser@nstat-b:~$ nc -lkv 9000
1457 Listening on [0.0.0.0] (family 0, port 9000)
1459 On nstat-a, run tcpdump to capture a SYN::
1461 nstatuser@nstat-a:~$ sudo tcpdump -w /tmp/paws_pre.pcap -c 1 port 9000
1462 tcpdump: listening on ens3, link-type EN10MB (Ethernet), capture size 262144 bytes
1464 On nstat-a, run nc as a client to connect nstat-b::
1466 nstatuser@nstat-a:~$ nc -v nstat-b 9000
1467 Connection to nstat-b 9000 port [tcp/*] succeeded!
1469 Now the tcpdump has captured the SYN and exit. We should fix the
1472 nstatuser@nstat-a:~$ tcprewrite --infile /tmp/paws_pre.pcap --outfile /tmp/paws.pcap --fixcsum
1474 Send the SYN packet twice::
1476 nstatuser@nstat-a:~$ for i in {1..2}; do sudo tcpreplay -i ens3 /tmp/paws.pcap; done
1478 On nstat-b, check the snmp counter::
1480 nstatuser@nstat-b:~$ nstat | grep -i skip
1481 TcpExtTCPACKSkippedPAWS 1 0.0
1483 We sent two SYN via tcpreplay, both of them would let PAWS check
1484 failed, the nstat-b replied an ACK for the first SYN, skipped the ACK
1485 for the second SYN, and updated TcpExtTCPACKSkippedPAWS.
1487 TcpExtTCPACKSkippedSeq
1488 --------------------
1489 To trigger TcpExtTCPACKSkippedSeq, we send packets which have valid
1490 timestamp (to pass PAWS check) but the sequence number is out of
1491 window. The linux TCP stack would avoid to skip if the packet has
1492 data, so we need a pure ACK packet. To generate such a packet, we
1493 could create two sockets: one on port 9000, another on port 9001. Then
1494 we capture an ACK on port 9001, change the source/destination port
1495 numbers to match the port 9000 socket. Then we could trigger
1496 TcpExtTCPACKSkippedSeq via this packet.
1498 On nstat-b, open two terminals, run two nc commands to listen on both
1499 port 9000 and port 9001::
1501 nstatuser@nstat-b:~$ nc -lkv 9000
1502 Listening on [0.0.0.0] (family 0, port 9000)
1504 nstatuser@nstat-b:~$ nc -lkv 9001
1505 Listening on [0.0.0.0] (family 0, port 9001)
1507 On nstat-a, run two nc clients::
1509 nstatuser@nstat-a:~$ nc -v nstat-b 9000
1510 Connection to nstat-b 9000 port [tcp/*] succeeded!
1512 nstatuser@nstat-a:~$ nc -v nstat-b 9001
1513 Connection to nstat-b 9001 port [tcp/*] succeeded!
1515 On nstat-a, run tcpdump to capture an ACK::
1517 nstatuser@nstat-a:~$ sudo tcpdump -w /tmp/seq_pre.pcap -c 1 dst port 9001
1518 tcpdump: listening on ens3, link-type EN10MB (Ethernet), capture size 262144 bytes
1520 On nstat-b, send a packet via the port 9001 socket. E.g. we sent a
1521 string 'foo' in our example::
1523 nstatuser@nstat-b:~$ nc -lkv 9001
1524 Listening on [0.0.0.0] (family 0, port 9001)
1525 Connection from nstat-a 42132 received!
1528 On nstat-a, the tcpdump should have caputred the ACK. We should check
1529 the source port numbers of the two nc clients::
1531 nstatuser@nstat-a:~$ ss -ta '( dport = :9000 || dport = :9001 )' | tee
1532 State Recv-Q Send-Q Local Address:Port Peer Address:Port
1533 ESTAB 0 0 192.168.122.250:50208 192.168.122.251:9000
1534 ESTAB 0 0 192.168.122.250:42132 192.168.122.251:9001
1536 Run tcprewrite, change port 9001 to port 9000, chagne port 42132 to
1539 nstatuser@nstat-a:~$ tcprewrite --infile /tmp/seq_pre.pcap --outfile /tmp/seq.pcap -r 9001:9000 -r 42132:50208 --fixcsum
1541 Now the /tmp/seq.pcap is the packet we need. Send it to nstat-b::
1543 nstatuser@nstat-a:~$ for i in {1..2}; do sudo tcpreplay -i ens3 /tmp/seq.pcap; done
1545 Check TcpExtTCPACKSkippedSeq on nstat-b::
1547 nstatuser@nstat-b:~$ nstat | grep -i skip
1548 TcpExtTCPACKSkippedSeq 1 0.0