I'm with Mark here.
We run 24/7 with somewhere around 170,000 law enforcement agencies relying on this system. We have stringent configuration control and even though we can get around that (if we have to) temporarily (for emergency fixes), we ALWAYS back end the fixes back into configuration management. I suspect any large shop will do exactly the same things. We cannot afford to have people monkeying around with kernels etc. Kevin -----Original Message----- From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On Behalf Of Erik N Johnson Sent: Monday, July 14, 2008 11:17 PM To: LINUX-390@VM.MARIST.EDU Subject: Re: Reiser > Sometimes, I see people on RHEL lists talking about building their own > kernel. My advice generally goes something like > 1. Check your support contract (think it says "unsupported," it would if > I were the vendor). > 2. Use CentOS (or Scientific Linux) on that system. > > Sometimes I might also suggest checking the supply of CentOS kernels. Or roll your own completely from scratch. If you don't have a reason to build a kernel, don't. But if you have a computer capable of amazing feats of virtualization and there's something to be gained by making a change IPL a whole new kernel. Make your changes in that one. It's sand-boxed, right? Nothing is permanent here. Nobody is asking you to put your entire clientele at the mercy of some experiment. If this kind of talk makes you wince then you're right to think "I should just go with the program that comes in the box that my vendor provides and not ask any questions or poke any holes in anything. If you don't have people actually programming for a mainframe on a mainframe this doesn't make any sense either. I guess my question would have to be: In an environment where changes cost people money and people are making unnecessary changes, why are you worried about reining them in? FIRE THEM. You don't praise accountants for creativity. If somebody is trying to use your very expensive machinery to solve a problem in a new and interesting way that is going to make your company money though, that's when I'm confused about why you would ever rein in the creative process. Programmers DO create things, after all. Erik Johnson On Mon, Jul 14, 2008 at 7:40 PM, John Summerfield <[EMAIL PROTECTED]> wrote: > Alan Altmark wrote: >> >> On Monday, 07/14/2008 at 05:55 EDT, Erik N Johnson <[EMAIL PROTECTED]> >> wrote: >>> >>> Building a kernel is not a herculean task by any measure. It is >>> completely automated and the configuration can easily be done >>> graphically if you have an X11 server. You probably need to go >>> looking for some literature before you try to boot up a machine as >>> expensive as a z10 on a homebrew kernel, but scads of PC Linux users >>> build their own kernel with every new release. The benefit is perhaps >>> to be questioned on big iron, bearing in mind that the folks like SuSE >>> that provide those default builds also provide lots of the actual >>> kernel code. Besides, the peripherals on a mainframe are much less >>> numerous and klugey, eliminating another big reason to roll your own. >>> It's not hard, just not that useful. >> >> It's not a question of difficulty. When you build your own kernel, the >> support you get from the distributors evaporates. Corporate customers >> need someone to flog in case things go bad, ergo no custom kernels by >> policy. >> >> Perhaps the distributors are more tolerant of custom kernels on other >> platforms - I don't know. > > Sometimes, I see people on RHEL lists talking about building their own > kernel. My advice generally goes something like > 1. Check your support contract (think it says "unsupported," it would if > I were the vendor). > 2. Use CentOS (or Scientific Linux) on that system. > > Sometimes I might also suggest checking the supply of CentOS kernels. > > > > -- > > Cheers > John > > -- spambait > [EMAIL PROTECTED] [EMAIL PROTECTED] > -- Advice > http://webfoot.com/advice/email.top.php > http://www.catb.org/~esr/faqs/smart-questions.html > http://support.microsoft.com/kb/555375 > > You cannot reply off-list:-) > > ---------------------------------------------------------------------- > For LINUX-390 subscribe / signoff / archive access instructions, > send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or > visit > http://www.marist.edu/htbin/wlvindex?LINUX-390 > ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390 ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
