Karl Katzke

We’ve had problems recently with our Dell 1650s at work. After five or six years of steady service, in the past three weeks we’ve had two bad fans on two separate machines, another machine has a voltage wobbly on it’s drive backplane, and another one failed with a rather dramatic detonation of some copper wire coiled around an insulator or magnet or something and mounted to the motherboard.

In every case but the last (which was a sudden failure) we had some good warning via the orange light — but one of the machines was running Netware, and the other two had BSD on them. No way to install OMSA and access the event log to figure out why the light was going wrong.

To solve it, I booted to a CentOS LiveCD, downloaded the ipmitool rpm and installed the Dell OMSA RPMS to the ramdisk, and waited till the orange light blinked. Then I checked the logs, got the results, fixed the problem, and rebooted back to the hard drive partitions.

This many failures within the course of weeks greatly worries me, since we still have three or four production systems in place using 1650s, not counting the two that are out of service (the one with the bad backplane and the one with the bad motherboard), which means I need to accelerate our migration to new hardware.