An instability counter is used to detect system instability.

• If 16 consecutive reboots are detected due to a watchdog, a “backup restore” operation is triggered.
• If 32 consecutive reboots are detected due to a watchdog, a “stock restore” operation is triggered.

The counter is reset if the system is up during at least 10 minutes. This feature cannot be disabled as it avoids bricking the gateway in case of software failure.

The counter is displayed at boot time only if the reboot is due to a watchdog. Bootcause information are available here.

U-Boot SPL 2015.04 (Apr 12 2017 - 16:03:25)
SEC: RNG instantiated
[rpmb] not configuring the rpmb while in non-secure mode
[rpmb] the secure storage will be installed in the USER partition
[rpmb] rng vector just fine avg:126 var:7
[rpmb] setup complete
spl bootfail=2
hab fuse not enabled
Authenticating TZSW succeeded
hab fuse not enabled
Authenticating u-boot.img succeeded


U-Boot 2015.04 (Apr 12 2017 - 16:03:25)

CPU:   Freescale i.MX6SX rev1.3 at 792 MHz
CPU:   Temperature 51 C
Reset cause: WDOG
       Watchdog enabled
DRAM:  240 MiB
MMC:   FSL_SDHC: 0
Using default environment

In:    serial
Out:   serial
Err:   serial
bootcause = WDOG
bootfail = 2
bootmode = nominal
Net:   FEC [PRIME]
Normal Boot

The updates that are being installed during an update request can make the system unstable so by default they are not included immediately in the backup.
Before each update, a “snapshot” of the firmware is done. It contains all the previously applied updates that are considered as “stable”. This “snapshot” will be used if a “backup restore” is triggered.
A “backup restore” will reinstall all the packets previously installed on the gateway except the last one. The files contained in /user/* are not removed. To trigger a backup restore execute the following commands:

kerosd -b
reboot

The stock firmware is backuped at production time (or when updated with a KerOS liveburner.ipk file).
A “stock restore” will factory reset the station. All data will be lost. To trigger a stock restore execute the following commands:

kerosd -s 
reboot

The goal of this feature is to provide a simple method to keep some local configuration which could be erased otherwise during some particular operations on the devices. Also this configuration could be stored, retrieved on a computer and applied on other devices.

To store the global configuration use the command :

store-config

If no file list is given, then /user/login and /user/rootfs_rw/* files are stored. Help menu can be displayed with the command store-config -h.

root@klk-lpbs-04018B:~ # store-config -h

    Syntax:
        /usr/sbin/store-config -h : this help
        /usr/sbin/store-config -l : list of files to be stored
        /usr/sbin/store-config -p : store the password file
        
    Store the files given as argument in an archive for later use.
    If no file list is given, then automatically /user//login /user//rootfs_rw files are stored.
    The counterpart of this command is restore_config, to restore the files.
root@klk-lpbs-04018B:~ # 

After this, a file named saved-config.tar should be created in the directory /user/.backup/. This file is just a basic tar archive without compression nor encryption.

Example with firmware v3.3.3:

# store-config -l /user/test1 /user/test2
Archiving file : tar: removing leading '/' from member names
Done

After this, a file named saved-config.tar should be created in the directory /user/.backup/ This file is just a basic tar archive without compression nor encryption.

# ll /user/.backup/
drwxr-xr-x    2 root     root        4.0K Feb 21 08:54 .
drwxr-xr-x    7 root     root        4.0K Feb 20 16:22 ..
-rw-r--r--    1 root     root        2.0K Feb 21 08:54 saved-config.tar

# tar -tf /user/.backup/saved-config.tar
user/test1
user/test2

To restore the global configuration please use the command :

restore-config

All previously stored files will be restored in their original places, and overwrite the files already in place if any.

A particular case is when we restore a configuration, but in the meanwhile an upgrade has been done and new users has been added. Then in such a case, an automatic merge of the password file is applied.