The following commit addresses the graceful unmounting of file
system and graceful shutdown of dockers before calling a
cold reboot which will cause a power cycle of SSD. This ensures
orderly shutdown and no corruption of files systems because
of the power cycle to SSD.
This commit will use the existing systemd-reboot service scripts
and override the configuration to do cold reboot for S6100 and
Z9100.
Unit tested the fix and graceful shutdown of file system and
dockers are done with cold reboot.
Signed-off-by: Harish Venkatraman <harish_venkatraman@dell.com>
- What I did
Added Daemon to Log LPC bus degradation in Intel C2000 processor. Intel Rangeley C2000 processors with revision less than or equal to 2 have issue where LPC bus degrades over time in some processors. To identify the problem and to notify the issue, a daemon has been added which will log on encountering the issue.
- How I did it
Added a daemon which validates the CPLD scratch(0x102) and SMF scratch(0x202) registers by writing and reading values on regular polling intervals (300 seconds). If there is a discrepancy between read and write, a critical log will be thrown.
- How to verify it
The infra is verify by simulating the issue where between write and read, the value in register is modified and the log appearance is checked.
- Description for the changelog
Added Daemon to identify LPC bus degradation issue and notify using syslog in Dell S6100 and Z9100 platforms. This daemon will only run on processors with revision less than or equal to 2.
This service (weekly) will let SSD firmware to do the garbage collection
after file-system deleted files. It could avoid slowness or
even READ-ONLY error due to SSD not being able to free the pages
even though the file system thinks there was a lot of space left.
Signed-off-by: Zhenggen Xu <zxu@linkedin.com>
