David wrote: >Apology in advance for the multitude of questions. They've gotta be really simple, >and slightly related. > > >1) What would be the procedure for flushing out the cache in my swap? Rebooting is >not an option in this case. > >2) Also, what are zombie processes? And sleeping processes? I'm thinking the >sleeping ones are daemons? What causes stopped jobs? How are they re-started? > >3) And while I'm at it, what is that entry for 0.0% nice under CPU states? >(Re-worded)--What is the CPU state of _nice_? > > > >output from top (captured at an abnormally active moment) is: > >[dave@kracker-jack dave]$ top >11:29pm up 10 days, 9:53, 1 user, load average: 0.71, 0.46, 0.49 >69 processes: 62 sleeping, 5 running, 2 zombie, 0 stopped >CPU states: 18.7% user, 7.4% system, 0.0% nice, 73.7% idle >Mem: 255828K av, 241164K used, 14664K free, 0K shrd, 7876K buff >Swap: 401584K av, 0K used, 401584K free 133156K cached > > >A URL would suffice for any worthwhile reading. > > > TIA >같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같� >David L. Steiner Registered Linux User #262493 >Mandrake 8.2 Enlightenment 0.16.5 Sylpheed 0.7.4claws >Email: [EMAIL PROTECTED] Homepage: www.davidlsteiner.com >같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같같� > > > > > > > >------------------------------------------------------------------------ > >Want to buy your Pack or Services from MandrakeSoft? >Go to http://www.mandrakestore.com > You never need to clean the cache from swap.
Zombie processes are what you don't want.. they take space and resources but don't do amything. Use kill -9 <pid of zombie process>. Sleeping processes are just that--either waiting for a slice of time and an interrupt or event to start them again or deliberately set to sleep until a scheduled awakening to perform a specific job (you'd be amazed--do you know what your computer is doing every day at 4 am?) daemons are either sleeping or running as the case may be. What differentiates a daemon is that 1) It waits for specific conditions to do a job 2) It checks for its own integrity 3) It is spawned from a master 4) After a cetrtain (programmable) number of runs, or if it finds itself corrupt, it suicides, requesting the master to spawn another. 5) Several copies of the same daemon may be available at the same time. Often this is configurable as with the apache daemon where you want to have a few spare copies around to attach to a flurry of incoming new connections. So a daemon is a programming style that is somewhat exploit, channel error and even (slightly) bug tolerant. It will still get the job done through all of that. Civileme
Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
