1earlyoom(1) General Commands Manual earlyoom(1)
2
6 earlyoom - Early OOM Daemon
7
9 earlyoom [OPTION]...
10
12 The oom-killer generally has a bad reputation among Linux users. One
13 may have to sit in front of an unresponsive system, listening to the
14 grinding disk for minutes, and press the reset button to quickly get
15 back to what one was doing after running out of patience.
16 earlyoom checks the amount of available memory and free swap up to 10
18 times a second (less often if there is a lot of free memory). If both
19 memory and swap are below 10%, it will kill the largest process (high‐
20 est oom_score). The percentage value is configurable via command line
21 arguments.
22 If there is a failure when trying to kill a process, earlyoom sleeps
24 for 1 second to limit log spam due to recurring errors.
25
27 -m PERCENT[,KILL_PERCENT]
28 set available memory minimum to PERCENT of total (default 10 %).
29 earlyoom starts sending SIGTERM once both memory and swap are below
31 their respective PERCENT setting. It sends SIGKILL once both are below
32 their respective KILL_PERCENT setting (default PERCENT/2).
33 Use the same value for PERCENT and KILL_PERCENT if you always want to
35 use SIGKILL.
36 Examples:
38 earlyoom # sets PERCENT=10, KILL_PERCENT=5
40 earlyoom -m 30 # sets PERCENT=30, KILL_PERCENT=15
41 earlyoom -m 20,18 # sets PERCENT=20, KILL_PERCENT=18
42 -s PERCENT[,KILL_PERCENT]
44 set free swap minimum to PERCENT of total (default 10 %). Send SIGKILL
45 if at or below KILL_PERCENT (default PERCENT/2), otherwise SIGTERM.
46 You can use -s 100 to have earlyoom effectively ignore swap usage: Pro‐
48 cesses are killed once available memory drops below the configured min‐
49 imum, no matter how much swap is free.
50 Use the same value for PERCENT and KILL_PERCENT if you always want to
52 use SIGKILL.
53 -M SIZE[,KILL_SIZE]
55 As an alternative to specifying a percentage of total memory, -M sets
56 the available memory minimum to SIZE KiB. The value is internally con‐
57 verted to a percentage. If you pass both -M and -m, the lower value is
58 used. Example: Reserve 10% of RAM but at most 1 GiB:
59 earlyoom -m 10 -M 1048576
61 earlyoom sends SIGKILL if at or below KILL_SIZE (default SIZE/2), oth‐
63 erwise SIGTERM.
64 -S SIZE[,KILL_SIZE]
66 As an alternative to specifying a percentage of total swap, -S sets the
67 free swap minimum to SIZE KiB. The value is internally converted to a
68 percentage. If you pass both -S and -s, the lower value is used.
69 Send SIGKILL if at or below KILL_SIZE (default SIZE/2), otherwise
71 SIGTERM.
72 -k
74 removed in earlyoom v1.2, ignored for compatibility
75 -i
77 removed in earlyoom v1.7, ignored for compatibility
78 -d
80 enable debugging messages
81 -v
83 print version information and exit
84 -r INTERVAL
86 Time between printing periodic memory reports, in seconds (default
87 1.0). A memory report looks like this:
88 mem avail: 21790 of 23909 MiB (91.14%), swap free: 0 of 0 MiB ( 0.00%)
90 Set to 3600 to print a report every hour, to 86400 to print once a day
92 etc. Set to 0 to disable printing periodic memory reports. Free memo‐
93 ry monitoring and low-memory killing runs independently of this option
94 at an adaptive poll rate that only depends on free memory. Due to the
95 adaptive poll rate, when there is a lot of free memory, the actual in‐
96 terval may be up to 1 second longer than the setting.
97 -p
99 Increase earlyoom’s priority: set niceness of earlyoom to -20 and
100 oom_score_adj to -100.
101 When earlyoom is run through its default systemd service, the -p switch
103 doesn’t work. To achieve the same effect, enter the following three
104 lines into sudo systemctl edit earlyoom:
105 [Service]
107 OOMScoreAdjust=-100
108 Nice=-20
109 -n
111 Enable notifications via d-bus.
112 To actually see the notifications in your GUI session, you need to have
114 systembus-notify (https://github.com/rfjakob/systembus-notify) running
115 as your user.
116 -N /PATH/TO/SCRIPT
118 Run the given script for each process killed. Must be an absolute
119 path.
120 Within the script, information about the killed process can be obtained
122 via the following environment variables:
123 EARLYOOM_PID Process PID
125 EARLYOOM_NAME Process name truncated to 16 bytes (as reported in /proc/PID/comm)
126 EARLYOOM_UID UID of the user running the process
127 WARNING: EARLYOOM_NAME can contain spaces, newlines, special characters
129 and is controlled by the user, or it can be empty! Make sure that your
130 notification script can handle that!
131 -g
133 Kill all processes that have same process group id (PGID) as the
134 process with excessive memory usage.
135 For example, with this flag turned on, the whole application will be
137 killed when one of its subprocess consumes too much memory (as long as
138 they all have the same PGID, of course).
139 Enable this flag when completely cleaning up the “entire application”
141 is more desirable, and you are sure that the application puts all its
142 processes in the same PGID.
143 Note that some desktop environments (GNOME, for example) put all desk‐
145 top application in the same process group as gnome-shell. earlyoom
146 might kill all such processes including gnome-shell when this flag is
147 turned on.
148 Be sure to check how your environment behaves beforehand. Use
150 pstree -gT
152 to show all processes with the PGID in brackets.
154 --prefer REGEX
156 prefer killing processes matching REGEX (adds 300 to oom_score)
157 --avoid REGEX
159 avoid killing processes matching REGEX (subtracts 300 from oom_score)
160 --dryrun
162 dry run (do not kill any processes)
163 -h, --help
165 this help text
166
168 0: Successful program execution.
169 1: Other error - inspect message for details
171 2: Switch conflict.
173 4: Could not cd to /proc
175 5: Could not open proc
177 7: Could not open /proc/sysrq-trigger
179 13: Unknown options.
181 14: Wrong parameters for other options.
183 15: Wrong parameters for memory threshold.
185 16: Wrong parameters for swap threshold.
187 102: Could not open /proc/meminfo
189 103: Could not read /proc/meminfo
191 104: Could not find a specific entry in /proc/meminfo
193 105: Could not convert number when parse the contents of /proc/meminfo
195
197 Earlyoom does not use echo f > /proc/sysrq-trigger because the Chrome
198 people made their browser always be the first (innocent!) victim by
199 setting oom_score_adj very high. Instead, earlyoom finds out itself by
200 reading through /proc/*/status (actually /proc/*/statm, which contains
201 the same information but is easier to parse programmatically).
202 Additionally, in recent kernels (tested on 4.0.5), triggering the ker‐
204 nel oom killer manually may not work at all. That is, it may only free
205 some graphics memory (that will be allocated immediately again) and not
206 actually kill any process.
207
209 About 2 MiB VmRSS. All memory is locked using mlockall() to make sure
210 earlyoom does not slow down in low memory situations.
211
213 If there is zero total swap on earlyoom startup, any -S (uppercase “S”)
214 values are ignored, a warning is printed, and default swap percentages
215 are used.
216 For processes matched by --prefer, negative oom_score_adj values are
218 not taken into account, and the process gets an effective oom_score of
219 at least 300. See https://github.com/rfjakob/earlyoom/issues/159 for
220 details.
221
223 The author of earlyoom is Jakob Unterwurzacher ⟨jakobunt@gmail.com⟩.
224 This manual page was written by Yangfl ⟨mmyangfl@gmail.com⟩, for the
226 Debian project (and may be used by others).
227 earlyoom(1)