View Single Post
Old 2020-07-30, 01:20   #10
P90 years forever!
Prime95's Avatar
Aug 2002
Yeehaw, FL

2·5·23·31 Posts

What I've discovered thusfar: This is a 5-CPU system. One CPU has a disk, the other four netboot off the first CPU. They all connect to the same network switch, which is crammed inside the CPU case.

No problems if main CPU and netboot CPUs #1 and #4 are running.
Adding netboot CPU#2 fails.
Adding netboot CPU#3 fails.
Now I way have used the same internet cable for #2 and #3 because the cable ordinarily used for #3 isn't working so I choose a different one at random (random being one of 3 other cables as there used to be 6 netboot CPUs).

Now testing #3 with a different cable.

New working theory: Recent work on the machine jostled or loosened one (or more) of the Internet cables causing noise on one of the lines. The OS is trying to process the noise and falling behind, until 19 minutes later a buffer overflows, and the OS dies. /var/log/syslog has no entries right before each reboot.
Prime95 is offline   Reply With Quote