On 7/2/07, Doug Scott <[EMAIL PROTECTED]> wrote: > Alberto Ruiz wrote: > > > > 2007/7/2, Peter Tribble <[EMAIL PROTECTED] > > <mailto:[EMAIL PROTECTED]>>: > > > > No, we should ship top and give users what they want. > > > > So, let's ship Microsoft Office with wine facilities. It's what the > > people want ;-)
Yes, but we don't have the rights to that :-( > Maybe you should actually give a good argument why top should not be > shipped, rather than just try to argue for the sake of it. > Arguing against 'top' is actually quite easy. i.e. prstat 'can' give > more accurate information for less load on the system. Maybe the answer > is to add an option to prstat to output 'top' like formatted data, and > hardlink 'top' to prstat. Arguing why top should be shipped is quite easy, given the relatively poor quality of prstat in many circumstances. (Not all, I use prstat extensively too.) It shows the time, which means you can see at a glance how well it's updating and at what interval. It shows the last pid, which shows you whether you've got a lot of process creation going on, and the rough rate. It gives you a rough cpu summary (idle/user/kernel) It gives you a quick memory summary. You can not show idle processes, which can reduce the visual clutter. You can select a user to show without exiting and restarting. You can sort without exiting and restating. You can change the update rate without exiting and restarting. You can zap an errant process directly from top without having move to another window and type in something afresh. The TIME column is more readable. The CPU column has an extra decimal place. I prefer having LWP as a separate column - putting the process name and number of LWPs together makes it harder to read. I find top output much more readable - I'm not sure why, but prstat always seems more dense and cluttered. It may just be the order of the columns and the spaces breaks it up a bit. As for resource utilization, that's pretty mixed. I would expect top to be slightly heavier as it's displaying a wider range of information, but sometimes find top to be lighter weight. (My last test showed top consuming 30% less cpu than prstat, but is does vary a bit - they seem pretty close on sparc.) But I'm digressing wildly. Just because it's the official dogma on Solaris doesn't make it right, and offering a choice of the best available tools has to be a win for everybody. -- -Peter Tribble http://www.petertribble.co.uk/ - http://ptribble.blogspot.com/ _______________________________________________ indiana-discuss mailing list [email protected] http://opensolaris.org/mailman/listinfo/indiana-discuss
