<?xml version="1.0" encoding="utf-8" ?> <!DOCTYPE chapter SYSTEM "chapter.dtd"> <chapter> <header> <copyright> <year>1997</year><year>2013</year> <holder>Ericsson AB. All Rights Reserved.</holder> </copyright> <legalnotice> The contents of this file are subject to the Erlang Public License, Version 1.1, (the "License"); you may not use this file except in compliance with the License. You should have received a copy of the Erlang Public License along with this software. If not, it can be retrieved online at http://www.erlang.org/. Software distributed under the License is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License for the specific language governing rights and limitations under the License. </legalnotice> <title>SASL Error Logging</title> <prepared>Magnus Fröberg</prepared> <responsible>Bjarne Däcker</responsible> <docno></docno> <approved></approved> <checked></checked> <date>1999-04-13</date> <rev>B</rev> <file>error_logging.xml</file> </header> <p>The SASL application introduces three types of reports:</p> <list type="bulleted"> <item>supervisor report</item> <item>progress report</item> <item>crash report.</item> </list> <p>When the SASL application is started, it adds a handler that formats and writes these reports, as specified in the configuration parameters for sasl, i.e the environment variables in the SASL application specification, which is found in the <c>.app</c> file of SASL. See <seealso marker="sasl_app">sasl(Application)</seealso>, and app(File) in the Kernel Reference Manual for the details.</p> <section> <title>Supervisor Report</title> <p>A supervisor report is issued when a supervised child terminates in an unexpected way. A supervisor report contains the following items:</p> <taglist> <tag>Supervisor.</tag> <item>The name of the reporting supervisor.</item> <tag>Context.</tag> <item>Indicates in which phase the child terminated from the supervisor's point of view. This can be <c>start_error</c>, <c>child_terminated</c>, or <c>shutdown_error</c>.</item> <tag>Reason.</tag> <item>The termination reason.</item> <tag>Offender.</tag> <item>The start specification for the child.</item> </taglist> </section> <section> <title>Progress Report</title> <p>A progress report is issued whenever a supervisor starts or restarts. A progress report contains the following items:</p> <taglist> <tag>Supervisor.</tag> <item>The name of the reporting supervisor.</item> <tag>Started.</tag> <item>The start specification for the successfully started child.</item> </taglist> <marker id="CRASH"></marker> </section> <section> <title>Crash Report</title> <p>Processes started with the <c>proc_lib:spawn</c> or <c>proc_lib:spawn_link</c> functions are wrapped within a <c>catch</c>. A crash report is issued whenever such a process terminates with an unexpected reason, which is any reason other than <c>normal</c> or <c>shutdown</c>. Processes using the <c>gen_server</c> and <c>gen_fsm</c> behaviours are examples of such processes. A crash report contains the following items:</p> <taglist> <tag>Crasher.</tag> <item>Information about the crashing process is reported, such as initial function call, exit reason, and message queue.</item> <tag>Neighbours.</tag> <item>Information about processes which are linked to the crashing process and do not trap exits. These processes are the neighbours which will terminate because of this process crash. The information gathered is the same as the information for Crasher, shown in the previous item.</item> </taglist> <section> <title>An Example</title> <p>The following example shows the reports which are generated when a process crashes. The example process is an <c>permanent</c> process supervised by the <c>test_sup</c> supervisor. A division by zero is executed and the error is first reported by the faulty process. A crash report is generated as the process was started using the <c>proc_lib:spawn/3</c> function. The supervisor generates a supervisor report showing the process that has crashed, and then a progress report is generated when the process is finally re-started.</p> <pre> =ERROR REPORT==== 27-May-1996::13:38:56 === <0.63.0>: Divide by zero ! =CRASH REPORT==== 27-May-1996::13:38:56 === crasher: pid: <0.63.0> registered_name: [] error_info: {badarith,{test,s,[]}} initial_call: {test,s,[]} ancestors: [test_sup,<0.46.0>] messages: [] links: [<0.47.0>] dictionary: [] trap_exit: false status: running heap_size: 128 stack_size: 128 reductions: 348 neighbours: =SUPERVISOR REPORT==== 27-May-1996::13:38:56 === Supervisor: {local,test_sup} Context: child_terminated Reason: {badarith,{test,s,[]}} Offender: [{pid,<0.63.0>}, {name,test}, {mfa,{test,t,[]}}, {restart_type,permanent}, {shutdown,200}, {child_type,worker}] =PROGRESS REPORT==== 27-May-1996::13:38:56 === Supervisor: {local,test_sup} Started: [{pid,<0.64.0>}, {name,test}, {mfa,{test,t,[]}}, {restart_type,permanent}, {shutdown,200}, {child_type,worker}] </pre> </section> </section> <section> <title>Multi-File Error Report Logging</title> <p>Multi-file error report logging is used to store error messages, which are received by the <c>error_logger</c>. The error messages are stored in several files and each file is smaller than a specified amount of kilobytes, and no more than a specified number of files exist at the same time. The logging is very fast because each error message is written as a binary term.</p> <p>Refer to <c>sasl</c> application in the Reference Manual for more details.</p> </section> <section> <title>Report Browser</title> <p>The report browser is used to browse and format error reports written by the error logger handler <c>log_mf_h</c> defined in <c>stdlib</c>.</p> <p>The <c>log_mf_h</c> handler writes all reports to a report logging directory. This directory is specified when configuring the SASL application.</p> <p>If the report browser is used off-line, the reports can be copied to another directory which is specified when starting the browser. If no such directory is specified, the browser reads reports from the SASL <c>error_logger_mf_dir</c>.</p> <section> <title>Starting the Report Browser</title> <p>Start the <c>rb_server</c> with the function <c>rb:start([Options])</c> as shown in the following example:</p> <pre> 5><input>rb:start([{max, 20}]).</input> rb: reading report...done. rb: reading report...done. rb: reading report...done. rb: reading report...done. </pre> </section> <section> <title>On-line Help</title> <p>Enter the command <em>rb:help().</em> to access the report browser on-line help system.</p> </section> <section> <title>List Reports in the Server</title> <p>The function <c>rb:list()</c> lists all loaded reports:</p> <pre> 4><input>rb:list().</input> No Type Process Date Time == ==== ======= ==== ==== 20 progress <0.17.0> 1996-10-16 16:14:54 19 progress <0.14.0> 1996-10-16 16:14:55 18 error <0.15.0> 1996-10-16 16:15:02 17 progress <0.14.0> 1996-10-16 16:15:06 16 progress <0.38.0> 1996-10-16 16:15:12 15 progress <0.17.0> 1996-10-16 16:16:14 14 progress <0.17.0> 1996-10-16 16:16:14 13 progress <0.17.0> 1996-10-16 16:16:14 12 progress <0.14.0> 1996-10-16 16:16:14 11 error <0.17.0> 1996-10-16 16:16:21 10 error <0.17.0> 1996-10-16 16:16:21 9 crash_report release_handler 1996-10-16 16:16:21 8 supervisor_report <0.17.0> 1996-10-16 16:16:21 7 progress <0.17.0> 1996-10-16 16:16:21 6 progress <0.17.0> 1996-10-16 16:16:36 5 progress <0.17.0> 1996-10-16 16:16:36 4 progress <0.17.0> 1996-10-16 16:16:36 3 progress <0.14.0> 1996-10-16 16:16:36 2 error <0.15.0> 1996-10-16 16:17:04 1 progress <0.14.0> 1996-10-16 16:17:09 ok </pre> </section> <section> <title>Show Reports</title> <p>To show details of a specific report, use the function <c>rb:show(Number)</c>:</p> <pre> 10> <input>rb:show(1).</input> 7> <input>rb:show(4).</input> PROGRESS REPORT <0.20.0> 1996-10-16 16:16:36 =============================================================================== supervisor {local,sasl_sup} started [{pid,<0.24.0>}, {name,release_handler}, {mfa,{release_handler,start_link,[]}}, {restart_type,permanent}, {shutdown,2000}, {child_type,worker}] ok 8> rb:show(9). CRASH REPORT <0.24.0> 1996-10-16 16:16:21 =============================================================================== Crashing process pid <0.24.0> registered_name release_handler error_info {undef,{release_handler,mbj_func,[]}} initial_call {gen,init_it, [gen_server, <0.20.0>, <0.20.0>, {erlang,register}, release_handler, release_handler, [], []]} ancestors [sasl_sup,<0.18.0>] messages [] links [<0.23.0>,<0.20.0>] dictionary [] trap_exit false status running heap_size 610 stack_size 142 reductions 54 ok </pre> </section> <section> <title>Search the Reports</title> <p>It is possible to show all reports which contain a common pattern. Suppose a process crashes because it tries to call a non-existing function <c>release_handler:mbj_func.</c> We could then show reports as follows:</p> <pre> 12><input>rb:grep("mbj_func").</input> Found match in report number 11 ERROR REPORT <0.24.0> 1996-10-16 16:16:21 =============================================================================== ** undefined function: release_handler:mbj_func[] ** Found match in report number 10 ERROR REPORT <0.24.0> 1996-10-16 16:16:21 =============================================================================== ** Generic server release_handler terminating ** Last message in was {unpack_release,hej} ** When Server state == {state,[], "/home/dup/otp2/otp_beam_sunos5_p1g_7", [{release, "OTP APN 181 01", "P1G", undefined, [], permanent}], undefined} ** Reason for termination == ** {undef,{release_handler,mbj_func,[]}} Found match in report number 9 CRASH REPORT <0.24.0> 1996-10-16 16:16:21 =============================================================================== Crashing process pid <0.24.0> registered_name release_handler error_info {undef,{release_handler,mbj_func,[]}} initial_call {gen,init_it, [gen_server, <0.20.0>, <0.20.0>, {erlang,register}, release_handler, release_handler, [], []]} ancestors [sasl_sup,<0.18.0>] messages [] links [<0.23.0>,<0.20.0>] dictionary [] trap_exit false status running heap_size 610 stack_size 142 reductions 54 Found match in report number 8 SUPERVISOR REPORT <0.20.0> 1996-10-16 16:16:21 =============================================================================== Reporting supervisor {local,sasl_sup} Child process errorContext child_terminated reason {undef,{release_handler,mbj_func,[]}} pid <0.24.0> name release_handler start_function {release_handler,start_link,[]} restart_type permanent shutdown 2000 child_type worker ok </pre> </section> <section> <title>Stop the Server</title> <p>Stop the <c>rb_server</c> with the function <c>rb:stop()</c>:</p> <pre> 13><input>rb:stop().</input> ok </pre> </section> </section> </chapter>