On Mon, 2005-03-21 at 11:23, Aleksandar Milivojevic wrote: > David Curry wrote: > > Discussion in this thread frequently reflected an unwarranted, > > underlying assumption. Namely, that linux/unix experts are intalling > > the OS, know how the system will be used, and act immediately after > > installation to reset default installation resource limits to > > appropriate levels. It is obvious to me from thread discussion that that > > assumption is invalid. Rather the discussion suggests many/some Yes the assumption is invalid now that Linux is beginning to make inroads onto the desktop world. Having some kind of limit is IMHO preferable to having no limit. It is better to have a user process fail than to have the entire system fail. Those that need more resources can raise the limits. Selecting a reasonable ulimit based on the systems available resources should not be difficult and should cover 90% of the user base. The other 10% are probably tweaking their systems due to the types of things they use them for anyway. > Linux does not protect user space > processes from each other. That statement is incorrect. Linux and Unix in general have done a better job of this than Windows ever did. I think what you mean is that without setting appropriate ulimits there is nothing to keep a user process from using all available resources on a system. This in turn can impact other users since they may not be able to get resources from the system as needed and ultimately it can impact the entire system if the kernel is unable to get resources as well. As Linux becomes more main stream the assumption has to be that users won't have the expertise to tune a system. As such reasonable defaults and limits should be put in place to protect the user and the system. Those that have requirements that exceed these limits should be in the 10% range if the defaults and limits are well selected. -- Scot L. Harris webid@xxxxxxxxxx DEC diagnostics would run on a dead whale. -- Mel Ferentz
