1SYSTEMD-COREDUMP(8) systemd-coredump SYSTEMD-COREDUMP(8)
2
6 systemd-coredump, systemd-coredump.socket, systemd-coredump@.service -
7 Acquire, save and process core dumps
8
10 /usr/lib/systemd/systemd-coredump
11 /usr/lib/systemd/systemd-coredump --backtrace
13 systemd-coredump@.service
15 systemd-coredump.socket
17
19 systemd-coredump@.service is a system service to process core dumps. It
20 will log a summary of the event to systemd-journald.service(8),
21 including information about the process identifier, owner, the signal
22 that killed the process, and the stack trace if possible. It may also
23 save the core dump for later processing. See the "Information about the
24 crashed process" section below.
25 The behavior of a specific program upon reception of a signal is
27 governed by a few factors which are described in detail in core(5). In
28 particular, the core dump will only be processed when the related
29 resource limits are sufficient.
30 Core dumps can be written to the journal or saved as a file. In both
32 cases, they can be retrieved for further processing, for example in
33 gdb(1). See coredumpctl(1), in particular the list and debug verbs.
34 By default, systemd-coredump will log the core dump to the journal,
36 including a backtrace if possible, and store the core dump (an image of
37 the memory contents of the process) itself in an external file in
38 /var/lib/systemd/coredump. These core dumps are deleted after a few
39 days by default; see /usr/lib/tmpfiles.d/systemd.conf for details. Note
40 that the removal of core files from the file system and the purging of
41 journal entries are independent, and the core file may be present
42 without the journal entry, and journal entries may point to
43 since-removed core files. Some metadata is attached to core files in
44 the form of extended attributes, so the core files are useful for some
45 purposes even without the full metadata available in the journal entry.
46 Invocation of systemd-coredump
48 The systemd-coredump executable does the actual work. It is invoked
49 twice: once as the handler by the kernel, and the second time in the
50 systemd-coredump@.service to actually write the data to the journal and
51 process and save the core file.
52 When the kernel invokes systemd-coredump to handle a core dump, it runs
54 in privileged mode, and will connect to the socket created by the
55 systemd-coredump.socket unit, which in turn will spawn an unprivileged
56 systemd-coredump@.service instance to process the core dump. Hence
57 systemd-coredump.socket and systemd-coredump@.service are helper units
58 which do the actual processing of core dumps and are subject to normal
59 service management.
60 It is also possible to invoke systemd-coredump with --backtrace option.
62 In this case, systemd-coredump expects a journal entry in the journal
63 Journal Export Format[1] on standard input. The entry should contain a
64 MESSAGE= field and any additional metadata fields the caller deems
65 reasonable. systemd-coredump will append additional metadata fields in
66 the same way it does for core dumps received from the kernel. In this
67 mode, no core dump is stored in the journal.
68
70 For programs started by systemd, process resource limits can be set by
71 directive LimitCORE=, see systemd.exec(5).
72 In order to be used by the kernel to handle core dumps,
74 systemd-coredump must be configured in sysctl(8) parameter
75 kernel.core_pattern. The syntax of this parameter is explained in
76 core(5). systemd installs the file /usr/lib/sysctl.d/50-coredump.conf
77 which configures kernel.core_pattern accordingly. This file may be
78 masked or overridden to use a different setting following normal
79 sysctl.d(5) rules. If the sysctl configuration is modified, it must be
80 updated in the kernel before it takes effect, see sysctl(8) and
81 systemd-sysctl(8).
82 In order to be used in the --backtrace mode, an appropriate backtrace
84 handler must be installed on the sender side. For example, in case of
85 python(1), this means a sys.excepthook must be installed, see
86 systemd-coredump-python[2].
87 The behavior of systemd-coredump itself is configured through the
89 configuration file /etc/systemd/coredump.conf and corresponding
90 snippets /etc/systemd/coredump.conf.d/*.conf, see coredump.conf(5). A
91 new instance of systemd-coredump is invoked upon receiving every core
92 dump. Therefore, changes in these files will take effect the next time
93 a core dump is received.
94 Resources used by core dump files are restricted in two ways.
96 Parameters like maximum size of acquired core dumps and files can be
97 set in files /etc/systemd/coredump.conf and snippets mentioned above.
98 In addition the storage time of core dump files is restricted by
99 systemd-tmpfiles, corresponding settings are by default in
100 /usr/lib/tmpfiles.d/systemd.conf. The default is to delete core dumps
101 after a few days; see the above file for details.
102 Disabling coredump processing
104 To disable potentially resource-intensive processing by
105 systemd-coredump, set
106 Storage=none ProcessSizeMax=0
108 in coredump.conf(5).
110
112 coredumpctl(1) can be used to retrieve saved core dumps independently
113 of their location, to display information, and to process them e.g. by
114 passing to the GNU debugger (gdb).
115 Data stored in the journal can be also viewed with journalctl(1) as
117 usual (or from any other process, using the sd-journal(3) API). The
118 relevant messages have MESSAGE_ID=fc2e22bc6ee647b6b90729ab34a250b1:
119 $ journalctl MESSAGE_ID=fc2e22bc6ee647b6b90729ab34a250b1 -o verbose
121 ...
122 MESSAGE_ID=fc2e22bc6ee647b6b90729ab34a250b1
123 COREDUMP_PID=552351
124 COREDUMP_UID=1000
125 COREDUMP_GID=1000
126 COREDUMP_SIGNAL_NAME=SIGSEGV
127 COREDUMP_SIGNAL=11
128 COREDUMP_TIMESTAMP=1614342930000000
129 COREDUMP_COMM=Web Content
130 COREDUMP_EXE=/usr/lib64/firefox/firefox
131 COREDUMP_USER_UNIT=app-gnome-firefox-552136.scope
132 COREDUMP_CMDLINE=/usr/lib64/firefox/firefox -contentproc -childID 5 -isForBrowser ...
133 COREDUMP_CGROUP=/user.slice/user-1000.slice/user@1000.service/app.slice/app-....scope
134 COREDUMP_FILENAME=/var/lib/systemd/coredump/core.Web....552351.....zst
135 ...
136 The following fields are saved (if known) with the journal entry
139 COREDUMP_UID=, COREDUMP_PID=, COREDUMP_GID=
141 The process number (PID), owner user number (UID), and group number
142 (GID) of the crashed process.
143 When the crashed process was part of a container (or in a process
145 or user namespace in general), those are the values as seen
146 outside, in the namespace where systemd-coredump is running.
147 COREDUMP_TIMESTAMP=
149 The time of the crash as reported by the kernel (in µs since the
150 epoch).
151 COREDUMP_RLIMIT=
153 The core file size soft resource limit, see getrlimit(2).
154 COREDUMP_UNIT=, COREDUMP_SLICE=
156 The system unit and slice names.
157 When the crashed process was in container, those are the units
159 names outside, in the main system manager.
160 COREDUMP_CGROUP=
162 Control group information in the format used in /proc/self/cgroup.
163 On systems with the unified cgroup hierarchy, this is a single path
164 prefixed with "0::", and multiple paths prefixed with controller
165 numbers on legacy systems.
166 When the crashed process was in a container, this is the full path,
168 as seen outside of the container.
169 COREDUMP_OWNER_UID=, COREDUMP_USER_UNIT=
171 The numerical UID of the user owning the login session or systemd
172 user unit of the crashed process, and the user manager unit. Both
173 fields are only present for user processes.
174 When the crashed process was in container, those are the values
176 outside, in the main system.
177 COREDUMP_SIGNAL_NAME=, COREDUMP_SIGNAL=
179 The terminating signal name (with the "SIG" prefix [3]) and
180 numerical value. (Both are included because signal numbers vary by
181 architecture.)
182 COREDUMP_CWD=, COREDUMP_ROOT=
184 The current working directory and root directory of the crashed
185 process.
186 When the crashed process is in a container, those paths are
188 relative to the root of the container's mount namespace.
189 COREDUMP_OPEN_FDS=
191 Information about open file descriptors, in the following format:
192 fd:/path/to/file
194 pos: ...
195 flags: ...
196 ...
197 fd:/path/to/file
199 pos: ...
200 flags: ...
201 ...
202 The first line contains the file descriptor number fd and the path,
205 while subsequent lines show the contents of /proc/pid/fdinfo/fd.
206 COREDUMP_EXE=
208 The destination of the /proc/pid/exe symlink.
209 When the crashed process is in a container, that path is relative
211 to the root of the container's mount namespace.
212 COREDUMP_COMM=, COREDUMP_PROC_STATUS=, COREDUMP_PROC_MAPS=,
214 COREDUMP_PROC_LIMITS=, COREDUMP_PROC_MOUNTINFO=, COREDUMP_ENVIRON=
215 Fields that map the per-process entries in the /proc/ filesystem:
216 /proc/pid/comm (the command name associated with the process),
217 /proc/pid/exe (the filename of the executed command),
218 /proc/pid/status (various metadata about the process),
219 /proc/pid/maps (memory regions visible to the process and their
220 access permissions), /proc/pid/limits (the soft and hard resource
221 limits), /proc/pid/mountinfo (mount points in the process's mount
222 namespace), /proc/pid/environ (the environment block of the crashed
223 process).
224 See proc(5) for more information.
226 COREDUMP_HOSTNAME=
228 The system hostname.
229 When the crashed process was in container, this is the container
231 hostname.
232 COREDUMP_CONTAINER_CMDLINE=
234 For processes running in a container, the commandline of the
235 process spawning the container (the first parent process with a
236 different mount namespace).
237 COREDUMP=
239 When the core is stored in the journal, the core image itself.
240 COREDUMP_FILENAME=
242 When the core is stored externally, the path to the core file.
243 COREDUMP_TRUNCATED=
245 Set to "1" when the saved coredump was truncated. (A partial core
246 image may still be processed by some tools, though obviously not
247 all information is available.)
248 COREDUMP_PACKAGE_NAME=, COREDUMP_PACKAGE_VERSION=,
250 COREDUMP_PACKAGE_JSON=
251 If the executable contained .package metadata ELF notes, they will
252 be parsed and attached. The package and packageVersion of the
253 'main' ELF module (ie: the executable) will be appended
254 individually. The JSON-formatted content of all modules will be
255 appended as a single JSON object, each with the module name as the
256 key. For more information about this metadata format and content,
257 see the coredump metadata spec[4].
258 MESSAGE=
260 The message generated by systemd-coredump that includes the
261 backtrace if it was successfully generated. When systemd-coredump
262 is invoked with --backtrace, this field is provided by the caller.
263 Various other fields exist in the journal entry, but pertain to the
265 logging process, i.e. systemd-coredump, not the crashed process. See
266 systemd.journal-fields(7).
267 The following fields are saved (if known) with the external file listed
269 in COREDUMP_FILENAME= as extended attributes:
270 user.coredump.pid, user.coredump.uid, user.coredump.gid,
272 user.coredump.signal, user.coredump.timestamp, user.coredump.rlimit,
273 user.coredump.hostname, user.coredump.comm, user.coredump.exe
274 Those are the same as COREDUMP_PID=, COREDUMP_UID=, COREDUMP_GID=,
275 COREDUMP_SIGNAL=, COREDUMP_TIMESTAMP=, COREDUMP_RLIMIT=,
276 COREDUMP_HOSTNAME=, COREDUMP_COMM=, and COREDUMP_EXE=, described
277 above.
278 Those can be viewed using getfattr(1). For the core file described in
280 the journal entry shown above:
281 $ getfattr --absolute-names -d /var/lib/systemd/coredump/core.Web....552351.....zst
283 # file: /var/lib/systemd/coredump/core.Web....552351.....zst
284 user.coredump.pid="552351"
285 user.coredump.uid="1000"
286 user.coredump.gid="1000"
287 user.coredump.signal="11"
288 user.coredump.timestamp="1614342930000000"
289 user.coredump.comm="Web Content"
290 user.coredump.exe="/usr/lib64/firefox/firefox"
291 ...
292
295 coredump.conf(5), coredumpctl(1), systemd-journald.service(8), systemd-
296 tmpfiles(8), core(5), sysctl.d(5), systemd-sysctl.service(8).
297
299 1. Journal Export Format
300 https://systemd.io/JOURNAL_EXPORT_FORMATS#journal-export-format
301 2. systemd-coredump-python
303 https://github.com/systemd/systemd-coredump-python
304 3. kill(1) expects signal names without the prefix; kill(2) uses the
306 prefix; all systemd tools accept signal names both with and without
307 the prefix.
308 4. the coredump metadata spec
310 https://systemd.io/COREDUMP_PACKAGE_METADATA/
311systemd 250 SYSTEMD-COREDUMP(8)