BMC capture journal

[debmc]

While debugging a host checkstop we needed to review the bmc journal to see when watchdog timers were being expired and related logging to help identify root cause (still working root cause). We are having to recreate and immediately capture the journal to get the data captured at the time of the checkstop.

It would be good to 'journalctl --no-pager' and add that file to the obmc_dumps collected so we can have valuable data both during debug as well as field issues.

I do understand the concern for keeping the dump small, but helpfulness and valuable data will be good strategic objectives as well.

Maybe identify which conditions would warrant capturing the journal, etc.

[ojayanth]

Due to space limitation, curently collecting journal log related the elog PID associated to the Checkstop.
@devenrao is working new dump type related to checkstop type.
https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-debug-collector/+/16440/

[debmc]

collecting journal

There are other non-checkstop paths where we are hitting the timer expired and need the journal log.

Dec 11 19:49:53 witherspoon phosphor-watchdog[982]: watchdog: Timed out