1
2mcelog.conf(5) File Formats Manual mcelog.conf(5)
3
4
5
7 mcelog.conf - mcelog.conf reference
8
10 /etc/mcelog.conf
11
13 /etc/mcelog.conf is the main configuration file for mcelog(8). This is
14 configuration file separated into sections including a default section.
15
16
17 General format
18
19 optionname = value
20
21 White space is not allowed in value currently, except at the end where
22 it is dropped
23
24
25 In general all command line options that are not commands work here.
26 See man mcelog or mcelog --help for a list. e.g. to enable the --no-
27 syslog option use
28
29 no-syslog = yes (or no to disable)
30
31 When the option has a argument
32
33 logfile = /tmp/logfile
34
35 Below are the options which are not command line options.
36
37 Set cpu type for which mcelog decodes events:
38
39 cpu = type
40
41 For valid values for type please see mcelog --help. If this value is
42 set incorrectly the decoded output will be likely incorrect. By
43 default when this parameter is not set mcelog uses the CPU it is run‐
44 ning on on very new kernels the mcelog events reported by the kernel
45 also carry the CPU type which is used too when available and not over‐
46 ridden.
47
48 Enable daemon mode:
49
50 daemon = yes
51
52 By default mcelog just processes the currently pending events and
53 exits. In daemon mode it will keep running as a daemon in the back‐
54 ground and poll the kernel for events and then decode them.
55
56 Filter out known broken events by default.
57
58 filter = yes
59
60 Don't log memory errors individually. They still get accounted if that
61 is enabled.
62
63 filter-memory-errors = yes
64
65 Output in undecoded raw format to be easier machine readable (default
66 is decoded).
67
68 raw = yes
69
70 Set cpu mhz to decode uptime from time stamp counter (output unreli‐
71 able, not needed on new kernels which report the event time directly. A
72 lot of systems don't have a linear time stamp clock and the output is
73 wrong then. Normally mcelog tries to figure out if it the TSC is reli‐
74 able and only uses the current frequency then. Setting a frequency
75 forces timestamp decoding. This setting is obsolete with modern ker‐
76 nels which report the time directly.
77
78 cpumhz = 1800.00
79
80 Log output options Log decoded machine checks in syslog (default stdout
81 or syslog for daemon)
82
83 syslog = yes
84
85 Log decoded machine checks in syslog with error level
86
87 syslog-error = yes
88
89 Never log anything to syslog
90
91 no-syslog = yes
92
93 Append log output to logfile instead of stdout. only when no syslog
94 logging is active
95
96 logfile = filename
97
98 Use smbios information to decode dimms (needs root). This function is
99 not recommended to use right now and generally not needed. The excep‐
100 tion is memdb prepopulation, which is configured separately below.
101
102 dmi = no
103
104 When in daemon mode run as this user after set up. Note that the trig‐
105 gers will run as this user too. Setting this to non root will mean
106 that triggers cannot take some corrective action, like offlining
107 objects.
108
109 run-credentials-user = root
110
111 Group to run as daemon with default to the group of the run-creden‐
112 tials-user
113
114 run-credentials-group = nobody
115
116 The server config section
117 User allowed to access client socket. when set to * match any root is
118 always allowed to access. default: root only
119
120 client-user = root
121
122 Group allowed to access mcelog When no group is configured any group
123 matches (but still user checking). when set to * match any
124
125 client-group = root
126
127 Path to the unix socket for client<->server communication. When no
128 socket-path is configured the server will not start
129
130 socket-path = /var/run/mcelog-client
131
132 When mcelog starts it checks if a server is already running. this con‐
133 figures the timeout for this check.
134
135 initial-ping-timeout = 2
136
137 The dimm config section
138 Is the in memory dimm error tracking enabled? Only works on systems
139 with integrated memory controller and which are supported. Only takes
140 effect in daemon mode.
141
142 dimm-tracking-enabled = yes
143
144 Use dmi information from the bios to prepopulate dimm database. Note
145 this might not work with all BIOS and requires mcelog to run as root.
146 Alternative is to let mcelog create DIMM objects on demand.
147
148 dmi-prepopulate = yes
149
150 Execute these triggers when the rate of corrected or uncorrected Errors
151 per DIMM exceeds the threshold. Note when the hardware does not report
152 DIMMs this might also be per channel. The default of 10/24h is reason‐
153 able for server quality DDR3 DIMMs as of 2009/10.
154
155 uc-error-trigger = dimm-error-trigger
156
157 uc-error-threshold = 1 / 24h
158
159 ce-error-trigger = dimm-error-trigger
160
161 ce-error-threshold = 10 / 24h
162
163 The socket config section
164 Enable memory error accounting per socket.
165
166 socket-tracking-enabled = yes
167
168 Threshold and trigger for uncorrected memory errors on a socket. mem-
169 uc-error-trigger = socket-memory-error-trigger
170
171 mem-uc-error-threshold = 100 / 24h
172
173 Trigger script for corrected memory errors on a socket.
174
175 mem-ce-error-trigger = socket-memory-error-trigger
176
177 Threshold on when to trigger a correct error for the socket.
178
179 mem-ce-error-threshold = 100 / 24h
180
181 log socket error threshold explicitly?
182
183 mem-ce-error-log = yes
184
185 Trigger script for uncorrected bus error events
186
187 bus-uc-threshold-trigger = bus-error-trigger
188
189 Trigger script for uncorrected iomca erors
190
191 iomca-threshold-trigger = iomca-error-trigger
192
193 Trigger script for other uncategorized errors
194
195 unknown-threshold-trigger = unknown-error-trigger
196
197 The cache config section
198 Processing of cache error thresholds reported by intel cpus.
199
200 cache-threshold-trigger = cache-error-trigger
201
202 Should cache threshold events be logged explicitly?
203
204 cache-threshold-log = yes
205
206 The page config section
207 Memory error accouting per 4k memory page. Threshold for the correct
208 memory errors trigger script.
209
210 memory-ce-threshold = 10 / 24h
211
212 Trigger script for corrected errors. memory-ce-trigger = page-error-
213 trigger
214
215 Should page threshold events be logged explicitly?
216
217 memory-ce-log = yes
218
219 Specify the internal action in mcelog to exceeding a page error thresh‐
220 old this is done in addition to executing the trigger script if avail‐
221 able off no action account only account errors soft try to
222 soft-offline page without killing any processes
223 This requires an uptodate kernel. Might not be successfull.
224 hard try to hard-offline page by killing processes
225 Requires an uptodate kernel. Might not be successfull. soft-
226 then-hard First try to soft offline, then try hard offlining
227
228 memory-ce-action = off|account|soft|hard|soft-then-hard
229
230 memory-ce-action = soft
231
232 The trigger config section
233 Maximum number of running triggers
234
235 children-max = 2
236
237 Execute triggers in this directory
238
239 directory = /etc/mcelog
240
242 mcelog(8) , http://www.mcelog.org
243
244
245
246
247 mcelog mcelog.conf(5)