Log Files For Offline Analysis and Defect Reporting
Current releases of the Oracle Communications Session Border Controller collect by default a large amount of information which is saved in various system logs and available after a system crash. This information is useful when debugging and submitting defects to Oracle.
The following table lists and describes the Application crash log files and includes a description and the output location of the crash log file when an HDD/SDD is present in the system or not.
| file name | description | Output path on HDD/SSD | Output path with flash drive only (no HDD/SSD) |
|---|---|---|---|
crashlogs.gz
|
compressed subset of /opt/logs/*
|
/opt/logs
|
/code/crash
|
dump.dpwd-datapath
|
Trace files for datapath watchdog initiated crash | /opt/crash
|
/code/crash
|
trace.<timestamp>.<task-name>
|
Crashing thread stack trace (text) | /opt/crash
/code/crash
|
/code/crash
|
core.<timestamp>.<task-name>.<process
ID>
|
Full process core dump | /opt/crash
|
not saved |
The following table lists and describes the Kernel crash log files and includes a description and the output path of the crash log file when an HDD/SDD is present in the system or not.
| file name | description | Output path on HDD/SSD | Output path with flash drive only (no HDD/SSD) |
|---|---|---|---|
vmcore.<datestamp>.dmesg
|
Copy of console logs from crashed kernel | /opt/crash
|
/code/crash
|
vmcore.<datestamp>.dump.*
|
Copy of the crashed kernel system memory | /opt/crash
|
not saved |
The following table lists and describes the operating system log files and includes a description and hte output path of the crash log file when an HDD/SDD is present in the system or not.
| file name | description | Output path on HDD/SSD | Output path with flash drive only (no HDD/SSD) |
|---|---|---|---|
kernel_lgob.log
|
Copied from /opt/logs/kernel.log at reboot
|
/opt/crash
|
/code/crash
|
dmesg_lgob.log
|
Copied from /opt/logs/dmesg.log at reboot
|
/opt/crash
|
/code/crash
|
For systems without SSDs or HDDs installed, persistent files are written to
flash in /code/crash. For systems with formatted SSDs or HDDs, the core and dump
files are written to /opt/crash on HDD. Regardless of whether an HDD/SSD is present, check
both flash and HDD locations on all platforms for all available files.
lgob files are overwritten at each reboot and crashlogs are overwritten at each crash. This data must be collected immediately after system recovers from an outage. Because trace and core files are deleted automatically at reboot to free disk space, they should be moved off-box at the earliest opportunity.
For systems with HDDs/SSDs, some minidump and trace files will be saved to
/code/crash and
some will be saved to /opt/crash. Please check both locations.
In the event of an uncontrolled reboot (i.e. hardware reset, power cycle,
watchdog trigger) the lgobs corresponding to the crash timestamp will be missing, so
absence of this file on /code/crash is an indication of an uncontrolled reboot event.
Note:
Large core files (core, vmcore dumps) may be split into multiple files (.aa, .ab etc), and on some platforms these files are also compressed. Be advised to look for .gz versions of the core files.