Making world but no kernel

[Michal Varga]

On Tue, 2011-07-26 at 08:34 -0700, Chuck Swiger wrote:
> On Jul 26, 2011, at 8:19 AM, Michal Varga wrote:
> > Well, that depends. Probably every single time I've seen someone
> > touching sysinstal in a post-install environment, that OS was instantly
> > rendered as much as good for a complete reinstall. It's just "one of
> > those things" that shouldn't be present in any rescue scenario.
> 
> I've been using sysinstall at dozens of customer sites since the
> 1990's without ever once running into a problem resembling what you've
> described.  Considering the absence of specific details, I think
> you're exaggerating well past the point of credibility.

This is an opinion you're entitled to and I have no particular need to
start proving you otherwise.

As far as my concerns go, you're free to update your source trees by
whatever means that suit you best, I was just pointing the OP a possible
point of failure that is easy to avoid by a simple and easily controlled
procedure, as there is perfectly nothing that can go wrong with csup. 

A vital part of any one-shot rescue operation is minimizing things that
may go wrong, beforehand. Sysinstall is everything but 'minimal'.


> Anyway, if someone does something bad using sysinstall and needs to
> fix it, restoring from backups should be all that is needed.  When
> people talk about doing a complete reinstall, it implies to me that
> they don't have backups in the first place.

Yes, and interestingly this is, in my own experience, very common among
people who use sysinstall to manage their systems (usually only about
once or twice).

Regards,
m.


-- 
Michal Varga,
Stonehenge (Gmail account)