1SS(8) System Manager's Manual SS(8)
2
6 ss - another utility to investigate sockets
7
9 ss [options] [ FILTER ]
10
12 ss is used to dump socket statistics. It allows showing information
13 similar to netstat. It can display more TCP and state information than
14 other tools.
15
18 When no option is used ss displays a list of open non-listening sockets
19 (e.g. TCP/UNIX/UDP) that have established connection.
20 -h, --help
22 Show summary of options.
23 -V, --version
25 Output version information.
26 -H, --no-header
28 Suppress header line.
29 -O, --oneline
31 Print each socket's data on a single line.
32 -n, --numeric
34 Do not try to resolve service names. Show exact bandwidth val‐
35 ues, instead of human-readable.
36 -r, --resolve
38 Try to resolve numeric address/ports.
39 -a, --all
41 Display both listening and non-listening (for TCP this means es‐
42 tablished connections) sockets.
43 -l, --listening
45 Display only listening sockets (these are omitted by default).
46 -o, --options
48 Show timer information. For TCP protocol, the output format is:
49 timer:(<timer_name>,<expire_time>,<retrans>)
51 <timer_name>
53 the name of the timer, there are five kind of timer
54 names:
55 on : means one of these timers: TCP retrans timer, TCP
57 early retrans timer and tail loss probe timer
58 keepalive: tcp keep alive timer
60 timewait: timewait stage timer
62 persist: zero window probe timer
64 unknown: none of the above timers
66 <expire_time>
68 how long time the timer will expire
69 <retrans>
71 how many times the retransmission occurred
72 -e, --extended
74 Show detailed socket information. The output format is:
75 uid:<uid_number> ino:<inode_number> sk:<cookie>
77 <uid_number>
79 the user id the socket belongs to
80 <inode_number>
82 the socket's inode number in VFS
83 <cookie>
85 an uuid of the socket
86 -m, --memory
88 Show socket memory usage. The output format is:
89 skmem:(r<rmem_alloc>,rb<rcv_buf>,t<wmem_alloc>,tb<snd_buf>,
91 f<fwd_alloc>,w<wmem_queued>,o<opt_mem>,
92 bl<back_log>,d<sock_drop>)
93 <rmem_alloc>
95 the memory allocated for receiving packet
96 <rcv_buf>
98 the total memory can be allocated for receiving packet
99 <wmem_alloc>
101 the memory used for sending packet (which has been sent
102 to layer 3)
103 <snd_buf>
105 the total memory can be allocated for sending packet
106 <fwd_alloc>
108 the memory allocated by the socket as cache, but not used
109 for receiving/sending packet yet. If need memory to
110 send/receive packet, the memory in this cache will be
111 used before allocate additional memory.
112 <wmem_queued>
114 The memory allocated for sending packet (which has not
115 been sent to layer 3)
116 <ropt_mem>
118 The memory used for storing socket option, e.g., the key
119 for TCP MD5 signature
120 <back_log>
122 The memory used for the sk backlog queue. On a process
123 context, if the process is receiving packet, and a new
124 packet is received, it will be put into the sk backlog
125 queue, so it can be received by the process immediately
126 <sock_drop>
128 the number of packets dropped before they are de-multi‐
129 plexed into the socket
130 -p, --processes
132 Show process using socket.
133 -i, --info
135 Show internal TCP information. Below fields may appear:
136 ts show string "ts" if the timestamp option is set
138 sack show string "sack" if the sack option is set
140 ecn show string "ecn" if the explicit congestion notification
142 option is set
143 ecnseen
145 show string "ecnseen" if the saw ecn flag is found in re‐
146 ceived packets
147 fastopen
149 show string "fastopen" if the fastopen option is set
150 cong_alg
152 the congestion algorithm name, the default congestion al‐
153 gorithm is "cubic"
154 wscale:<snd_wscale>:<rcv_wscale>
156 if window scale option is used, this field shows the send
157 scale factor and receive scale factor
158 rto:<icsk_rto>
160 tcp re-transmission timeout value, the unit is millisec‐
161 ond
162 backoff:<icsk_backoff>
164 used for exponential backoff re-transmission, the actual
165 re-transmission timeout value is icsk_rto << icsk_backoff
166 rtt:<rtt>/<rttvar>
168 rtt is the average round trip time, rttvar is the mean
169 deviation of rtt, their units are millisecond
170 ato:<ato>
172 ack timeout, unit is millisecond, used for delay ack mode
173 mss:<mss>
175 max segment size
176 cwnd:<cwnd>
178 congestion window size
179 pmtu:<pmtu>
181 path MTU value
182 ssthresh:<ssthresh>
184 tcp congestion window slow start threshold
185 bytes_acked:<bytes_acked>
187 bytes acked
188 bytes_received:<bytes_received>
190 bytes received
191 segs_out:<segs_out>
193 segments sent out
194 segs_in:<segs_in>
196 segments received
197 send <send_bps>bps
199 egress bps
200 lastsnd:<lastsnd>
202 how long time since the last packet sent, the unit is
203 millisecond
204 lastrcv:<lastrcv>
206 how long time since the last packet received, the unit is
207 millisecond
208 lastack:<lastack>
210 how long time since the last ack received, the unit is
211 millisecond
212 pacing_rate <pacing_rate>bps/<max_pacing_rate>bps
214 the pacing rate and max pacing rate
215 rcv_space:<rcv_space>
217 a helper variable for TCP internal auto tuning socket re‐
218 ceive buffer
219 tcp-ulp-mptcp flags:[MmBbJjecv] token:<rem_token(rem_id)/loc_to‐
221 ken(loc_id)> seq:<sn> sfseq:<ssn> ssnoff:<off> maplen:<maplen>
222 MPTCP subflow information
223 --tos Show ToS and priority information. Below fields may appear:
225 tos IPv4 Type-of-Service byte
227 tclass IPv6 Traffic Class byte
229 class_id
231 Class id set by net_cls cgroup. If class is zero this
232 shows priority set by SO_PRIORITY.
233 --cgroup
235 Show cgroup information. Below fields may appear:
236 cgroup Cgroup v2 pathname. This pathname is relative to the
238 mount point of the hierarchy.
239 -K, --kill
241 Attempts to forcibly close sockets. This option displays sockets
242 that are successfully closed and silently skips sockets that the
243 kernel does not support closing. It supports IPv4 and IPv6 sock‐
244 ets only.
245 -s, --summary
247 Print summary statistics. This option does not parse socket
248 lists obtaining summary from various sources. It is useful when
249 amount of sockets is so huge that parsing /proc/net/tcp is
250 painful.
251 -E, --events
253 Continually display sockets as they are destroyed
254 -Z, --context
256 As the -p option but also shows process security context.
257 For netlink(7) sockets the initiating process context is dis‐
259 played as follows:
260 1. If valid pid show the process context.
262 2. If destination is kernel (pid = 0) show kernel ini‐
264 tial context.
265 3. If a unique identifier has been allocated by the ker‐
267 nel or netlink user, show context as "unavailable".
268 This will generally indicate that a process has more
269 than one netlink socket active.
270 -z, --contexts
272 As the -Z option but also shows the socket context. The socket
273 context is taken from the associated inode and is not the actual
274 socket context held by the kernel. Sockets are typically labeled
275 with the context of the creating process, however the context
276 shown will reflect any policy role, type and/or range transition
277 rules applied, and is therefore a useful reference.
278 -N NSNAME, --net=NSNAME
280 Switch to the specified network namespace name.
281 -b, --bpf
283 Show socket classic BPF filters (only administrators are allowed
284 to get these information).
285 -4, --ipv4
287 Display only IP version 4 sockets (alias for -f inet).
288 -6, --ipv6
290 Display only IP version 6 sockets (alias for -f inet6).
291 -0, --packet
293 Display PACKET sockets (alias for -f link).
294 -t, --tcp
296 Display TCP sockets.
297 -u, --udp
299 Display UDP sockets.
300 -d, --dccp
302 Display DCCP sockets.
303 -w, --raw
305 Display RAW sockets.
306 -x, --unix
308 Display Unix domain sockets (alias for -f unix).
309 -S, --sctp
311 Display SCTP sockets.
312 --vsock
314 Display vsock sockets (alias for -f vsock).
315 --xdp Display XDP sockets (alias for -f xdp).
317 --inet-sockopt
319 Display inet socket options.
320 -f FAMILY, --family=FAMILY
322 Display sockets of type FAMILY. Currently the following fami‐
323 lies are supported: unix, inet, inet6, link, netlink, vsock,
324 xdp.
325 -A QUERY, --query=QUERY, --socket=QUERY
327 List of socket tables to dump, separated by commas. The follow‐
328 ing identifiers are understood: all, inet, tcp, udp, raw, unix,
329 packet, netlink, unix_dgram, unix_stream, unix_seqpacket,
330 packet_raw, packet_dgram, dccp, sctp, vsock_stream, vsock_dgram,
331 xdp Any item in the list may optionally be prefixed by an excla‐
332 mation mark (!) to exclude that socket table from being dumped.
333 -D FILE, --diag=FILE
335 Do not display anything, just dump raw information about TCP
336 sockets to FILE after applying filters. If FILE is - stdout is
337 used.
338 -F FILE, --filter=FILE
340 Read filter information from FILE. Each line of FILE is inter‐
341 preted like single command line option. If FILE is - stdin is
342 used.
343 FILTER := [ state STATE-FILTER ] [ EXPRESSION ]
345 Please take a look at the official documentation for details re‐
346 garding filters.
347
350 STATE-FILTER allows to construct arbitrary set of states to match. Its
351 syntax is sequence of keywords state and exclude followed by identifier
352 of state.
353 Available identifiers are:
355 All standard TCP states: established, syn-sent, syn-recv, fin-
357 wait-1, fin-wait-2, time-wait, closed, close-wait, last-ack,
358 listening and closing.
359 all - for all the states
361 connected - all the states except for listening and closed
363 synchronized - all the connected states except for syn-sent
365 bucket - states, which are maintained as minisockets, i.e.
367 time-wait and syn-recv
368 big - opposite to bucket
370
373 EXPRESSION allows filtering based on specific criteria. EXPRESSION
374 consists of a series of predicates combined by boolean operators. The
375 possible operators in increasing order of precedence are or (or | or
376 ||), and (or & or &&), and not (or !). If no operator is between con‐
377 secutive predicates, an implicit and operator is assumed. Subexpres‐
378 sions can be grouped with "(" and ")".
379 The following predicates are supported:
381 {dst|src} [=] HOST
384 Test if the destination or source matches HOST. See HOST SYNTAX
385 for details.
386 {dport|sport} [OP] [FAMILY:]:PORT
388 Compare the destination or source port to PORT. OP can be any of
389 "<", "<=", "=", "!=", ">=" and ">". Following normal arithmetic
390 rules. FAMILY and PORT are as described in HOST SYNTAX below.
391 dev [=|!=] DEVICE
393 Match based on the device the connection uses. DEVICE can either
394 be a device name or the index of the interface.
395 fwmark [=|!=] MASK
397 Matches based on the fwmark value for the connection. This can
398 either be a specific mark value or a mark value followed by a
399 "/" and a bitmask of which bits to use in the comparison. For
400 example "fwmark = 0x01/0x03" would match if the two least sig‐
401 nificant bits of the fwmark were 0x01.
402 cgroup [=|!=] PATH
404 Match if the connection is part of a cgroup at the given path.
405 autobound
407 Match if the port or path of the source address was automati‐
408 cally allocated (rather than explicitly specified).
409 Most operators have aliases. If no operator is supplied "=" is assumed.
411 Each of the following groups of operators are all equivalent:
412 • = == eq
414 • != ne neq
416 • > gt
418 • < lt
420 • >= ge geq
422 • <= le leq
424 • ! not
426 • | || or
428 • & && and
430
432 The general host syntax is [FAMILY:]ADDRESS[:PORT].
433 FAMILY must be one of the families supported by the -f option. If not
435 given it defaults to the family given with the -f option, and if that
436 is also missing, will assume either inet or inet6. Note that all host
437 conditions in the expression should either all be the same family or be
438 only inet and inet6. If there is some other mixture of families, the
439 results will probably be unexpected.
440 The form of ADDRESS and PORT depends on the family used. "*" can be
442 used as a wildcard for either the address or port. The details for each
443 family are as follows:
444 unix ADDRESS is a glob pattern (see fnmatch(3)) that will be matched
446 case-insensitively against the unix socket's address. Both path
447 and abstract names are supported. Unix addresses do not support
448 a port, and "*" cannot be used as a wildcard.
449 link ADDRESS is the case-insensitive name of an Ethernet protocol to
451 match. PORT is either a device name or a device index for the
452 desired link device, as seen in the output of ip link.
453 netlink
455 ADDRESS is a descriptor of the netlink family. Possible values
456 come from /etc/iproute2/nl_protos. PORT is the port id of the
457 socket, which is usually the same as the owning process id. The
458 value "kernel" can be used to represent the kernel (port id of
459 0).
460 vsock ADDRESS is an integer representing the CID address, and PORT is
462 the port.
463 inet and inet6
465 ADDRESS is an ip address (either v4 or v6 depending on the fam‐
466 ily) or a DNS hostname that resolves to an ip address of the re‐
467 quired version. An ipv6 address must be enclosed in "[" and "]"
468 to disambiguate the port separator. The address may additionally
469 have a prefix length given in CIDR notation (a slash followed by
470 the prefix length in bits). PORT is either the numerical socket
471 port, or the service name for the port to match.
472
475 ss -t -a
476 Display all TCP sockets.
477 ss -t -a -Z
479 Display all TCP sockets with process SELinux security contexts.
480 ss -u -a
482 Display all UDP sockets.
483 ss -o state established '( dport = :ssh or sport = :ssh )'
485 Display all established ssh connections.
486 ss -x src /tmp/.X11-unix/*
488 Find all local processes connected to X server.
489 ss -o state fin-wait-1 '( sport = :http or sport = :https )' dst
491 193.233.7/24
492 List all the tcp sockets in state FIN-WAIT-1 for our apache to
493 network 193.233.7/24 and look at their timers.
494 ss -a -A 'all,!tcp'
496 List sockets in all states from all socket tables but TCP.
497
499 ip(8),
500 RFC 793 - https://tools.ietf.org/rfc/rfc793.txt (TCP states)
501
504 ss was written by Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>.
505 This manual page was written by Michael Prokop <mika@grml.org> for the
507 Debian project (but may be used by others).
508 SS(8)