| Sun Events: |
|
10/97 to 11/97 Systems Hang due to under-designed power supply. All 200 Sparc 4's affected. Fault isolated in 2 weeks. Remedy by replacing Power Supply when system hangs up. |
| SGI Events: |
|
3/95 to 8/95 133 MHz system replaces 100 MHz systems. IRIX 5.2 will not install. SGI purchase nearly cancelled. Remedy: systems with Secondary Cache work with IRIX 5.2. |
|
8/96 to 2/97 Leader of SGI project can't read mail on SGI system. Two months of fault isolation, and guess work. After an additional month of effort, convinced SGI service that the fault was in the R5000 chip. (Using a web page erratta sheet supplied by a colleague.) Remedy: Kernel patch to work around R5000 CPU chip bug. |
|
7/97 to 8/97 1995 vintage hardware fault in Indigo2 shows up in our Indys. Sporadic fault noticed over several months. Two months full time effort to diagnose. Nearly all Indy motherboards need replacement. |
|
7/98 to 8/98 Fileservers flooded with client packets. AFS bug, not SGI's bug. Sporadic fault noticed over months. Eventually isolated. Remedy: AFS patch. |
|
9/98 to 10/98 All O2's fall off the net. Remedy: Friend from SGI remembered an enet driver patch which is hurredly installed. 1 lucky day to diagnose. 1 week to integrate and test the patch. |
|
11/98 to 2/99 Large packet loss prevents reliable operation of Biosym application in courses taught in O2 classroom. Told by SGI to change to newer rev of Asante repeaters. Remedy: Change network for room full of O2's. Remember where we can't deploy O2's. Three weeks |
|
9/99 to ?/?? Systems hang in the field and require re-install. Athena patch release obviated need to reinstall. |
