I'm reviving this thread because I have taken delivery of another RB1200, and it has the exact same problem with intermittent lockups during reboot as the last one I bought: Typically, it will successfully reboot two to four times in a row before it fails to complete the boot process--at which point the only way to recover is to physically cycle the power.
As before, there is no crash output or error message on the console; it simply stops printing to the console after "jumping to kernel code". (Sometimes, upon being sent the reboot command, it locks up after printing the "Restarting system." message to the console, and never even gets back to the boot loader.)
This pattern holds whether running ROS 6.7, 5.26 or 5.18 (and their corresponding boot loaders). Sometimes it even fails on the first try (initial power-up), and will even fail within a minute of powering up from room temperature, when the heat sink is still barely warm; i.e., it doesn't seem to be heat-related.
On those occasions when it does reach the login prompt, everything seems to be fine; i.e., no crashes or sudden reboots or watchdog timeouts, just normal operation.
Oddities when running ROS 6.7/3.10:
ONCE I got the following message after "jumping to kernel code":
Initramfs unpacking failed: no cpio magic
Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)
Rebooting in 180 seconds..
TWICE, after locking up, it jumped on its own back to the boot loader after a few seconds, and completed the boot on the second attempt. On every other occasion that it has stopped indicating progress through the boot, it has failed to recover on its own, even after sitting for over 50 minutes on one occasion.
The serial number of this unit is 2D72011DF4B9/121, and the MAC range is 00:0C:42:C0:AC:0E - 17. (It isn't the same one that I bought before, as I posted earlier. That's why I came back to this thread in the first place: to make sure I hadn't been sold the same lemon as last time!) I have yet another RB1200 on its way, which should be arriving toward the end of the week, and will receive the same degree of scrutiny. If it has the same problems, that will mark the end of my involvement with RB1200s, at any price.
I had hoped that MT would have found it worthwhile to try to identify the problem with these particular boards, and either release (software) or describe (hardware) the fix, or pronounce them irreparable and put an end to the mystery once and for all. Best of all would be to identify a bad run of boards or components, so we could identify the bad ones without having to buy/test/return them. Instead, the main thing I see is that all mention of the RB1200 is scrubbed from routerboard.com, even from the archive section, where older hardware is still listed.
Cape Ann Communications, Inc.
Gloucester, MA, USA