CODEWITHSUNDEEP
kernel
Sriram
Sriram

Reputation: 150

Kernel panic - not syncing: Fatal exception in interrupt

How can I alter this behavior, even if it is risky? In some cases, the default behaviour is an obstacle, particularly, when the devices are crashing in the field and we need to know what went wrong, as a post mortem. So the /var/log/syslog should be updated. The question is how to easily change this behavior? any help will be greatly appreciated

Upvotes: 2

Views: 12063

Answers (2)

Terry Wang
Terry Wang

Reputation: 13920

If there is a kernel crash dump generated, that'll help to identify the root cause.

Even the call trace on the physical console will help to narrow down the issue.

I've seen many of this type of kernel panic due to NIC drivers, Cisco enic and various broadcom NICs in particular, hope it helps.

If it happens again and you want to do RCA, get a screen shot (photo) at least ;-D

Upvotes: 0

sqr163
sqr163

Reputation: 1184

Kernel panic means it decided to go into infinite loop to avoid execution of any other code (because it may result in unpredicted behavior). And normally you would not want to change this behavior.

Ensure you get all the messages from kernel: look into /etc/syslogd.conf and find string like this:

kern.*   -/var/log/kern.log

Ensure it contains the star * which means that all kernel messages (even debug) goes into the log file.

You may try configure syslogd to send the messages to a remote host - to read them even when the target machine becomes completely unavailable.

Upvotes: 1

Related Questions