jona@iscp.Bellcore.COM (Jon Alperin) (04/20/91)
The following information is provided so that we can double check a problem that IBM claims it cannot reproduce. The consequences of what I am about to present can be extremely dangerous, especially in a multi-user system. DO NOT TRY THIS ON ANYTHING OTHER THAN A STANDALONE OR SINGLE USER SYSTEM. We are running 3003/3004 and have managed to panic our machines consistently by performing simple series of commands. One such series is: set | grep PWD | grep u as well as: tail -200 | grep <somestring> | more SO: 1. Can anyone reproduce this (or a similar sequence)? 2. Does anyone have any ideas what caused this panic? We have narrowed it down to panicing under ksh (not csh), and having no effect whether we have mount points or not. Furthermore, it will not always fail on the first invocation of these commands. Sometimes it takes several invocations before the PANIC, while others will PANIC on the first try. Please e-mail responses, and I will pursue with IBM. Since we can't seem to figure out why these seemingly innocuous commands cause a panic, I figured that you system admins might want to be aware. The 888 signals are 0c9 888 102 700 0c4, and we are trying to get IBM a copy of the dump. -- Jon Alperin Bell Communications Research ---> Internet: jona@iscp.bellcore.com ---> Voicenet: (908) 699-8674 ---> UUNET: uunet!bcr!jona * All opinions and stupid questions are my own *