> scroll lock on on the vconsole with syslog spewing things. Buffer fills, > syslog hangs, anything trying to use syslog hangs. Chaos ensues. <snip> > Or use some kind of tail on a log file rather than directly dumping to a > console. Though this strikes me as a rather serious bug. Something as > vital as syslog shouldn't hang if it can't write somewhere, should it? > ;P something like this was brought up on the Linux for S/390 mailing list some time ago; the tradeoff between thorough logging (esp. kernel logging), and managing the amount of data logged so you don't DoS yourself via syslog and klog. the upshot was... there is no good answer. especially for kernel logging, where things *have* to be done quickly, or else it can be a nontrivial use of system resources. the mainframe people are used to logging *everything*, down to microcode problems (I belive S/390s have a special hardware coprocessor just for hardware fault logging; so even if everything bricks up, there's still a log of exactly what happened). however, especially in a large scale debugging environment (and let's face it, mainframes are built for _Large Scale_), the number of log messages (both kernel messages and userspace messages) can become difficult to manage, and even generate problems like Callum experienced. feel free to correct me if I'm wrong.. I'm still a bit shaky on how this all works. :) Carl Soderstrom -- Network Engineer Real-Time Enterprises (952) 943-8700 --------------------------------------------------------------------- To unsubscribe, e-mail: tclug-list-unsubscribe at mn-linux.org For additional commands, e-mail: tclug-list-help at mn-linux.org