>> >> This is for review. >> goto : http://www.blastwave.org/docs/index.html >> select doc BLS-0061 >> A lot of the pages were written when I was getting >> tired and there will >> be errors. At the very least some dry humour and >> possibly rambling. >> The topic or index page is still being written. It >> needs to point to >> all the other pages or at least to main topics. >> >> Yes, its a tad long. I count 170+ pages of html and >> I took I don't know >> how many snapshot images. >> >> bash-3.1$ find . -type f | grep png | wc -l >> 357 >> y .. a lot. >> >> Please let me know what needs to be fixed. >> I want to add in branches for jumpstart paths to >> follow as well as a >> whole writeup for recent Solaris Nevada. > > You did exactly what I was afraid you did: you sliced up the disk!
oh no !! ... [ spank spank spank ] > Of all the things to avoid, it's teaching the newbies to slice up the disk > into different filesystems. That's a topic "seasoned" sysadmins don't get > right, and a newb is sure to get lost. [ spank slap smack boom spank spank ] > Not to mention the disk is not used optimally when it is carved out into [ smack spank ] > filesystems. It's not 1991, and we do not have 465MB Maxtors and Conners to [ insert whimpering noises here ] > deal with any more, nor is the Boot PROM revision 2.x, so that it couldn't > boot a slice greater than 1GB... [ yes Sir .. I'll fix that .. sorry .. won't happen again ] > Point: > > s0: flags wm, /, sizeof(s2 - (s1 + s7) > s1: flags wu, swap, sizeof(RAM + 64MB) > s2: don't touch > s7: flags wu, 64.00MB (for metadb) > > And you're done. And the disk space is used optimally. I like s4 for my metadb area. Personal preference from way way back. Easy to fix^H^H^Hmodify. > Carving up the disk like you did is for advanced sysadmins who have a deep > understanding of Solaris file systems and want to specifically customize > them. And in your case, I argue that you are doing anything but using that > disk space in an optimal way. [ spank smack whimper ] but but .. it was just an example ! [ whammo ] > The "whole disk root" approach worked for IRIX. And for HP-UX. And by golly, > it sure works for Solaris as well, beautifully. And if the same old tired > and worn out argument is going to be "filling up the / fs", then you have a > much bigger problem, like lack of a system monitoring infrastructure in > place. On top of that somehow I doubt that a newb will fill up his/her 500GB > disk so fast, especially on Solaris. > > I argued this point with our Sun Solaris 10 instructor in class. He claimed > that slicing up the disk "is more flexible". I told him I'll have some of > whatever he's smoking too, please. I'd like some pain killers now. Percocets ? please ? :-) Dennis _______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
