Re: Sun killing Solaris?
From: Martin Paul (map_at_par.univie.ac.at)
Date: 07/28/05
- Next message: Martin Paul: "Re: Sun killing Solaris?"
- Previous message: mr_peter_stevenson_at_hotmail.com: "Re: Routing/IP question: Is there any way to do this?"
- In reply to: Lion-O: "Re: Sun killing Solaris?"
- Next in thread: Casper H.S. ***: "Re: Sun killing Solaris?"
- Reply: Casper H.S. ***: "Re: Sun killing Solaris?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 28 Jul 2005 08:50:53 GMT
Lion-O <nosp@m.catslair.org> wrote:
>> Has anybody tried the examples listed under "Using wget" ?
>
> My guess is that they're trying to limit the options to only a few which
> consist of using their own software. You can see that much on Solaris 10
> which features 'smpatch' to update the system.
Hm, but that doesn't explain why they list examples on how to do it
with wget which then don't work.
> So those users will probably rely on those tools, and in a way they
> maybe more reliable and safer too. Note; I'm not sure since I never used
> ftp nor wget to download patches.
The tools may be nice for users who run Solaris on their private
desktop PC. But for any semi-serious administrator with multiple
Sun machines (with different versions of Solaris, stripped down
OS installations, machines which might not be connected to the
Internet) none of Sun's tools are useable.
For me smpatch is just too opaque to trust it on a production machine.
It's like running "make install" as root when installing software,
which is a big no-no. I want to know what happens and which files will
be modified.
> But I wonder if people using anonymous ftp only did so to get hold of
> them or also automated the install part. Because the latter part may
> imply problems with patches which need an immediate reboot and/or single
> user mode.
In fact there aren't many patches which really need an immediate
reboot or single user mode. I have often installed all missing
patches in one go (either in multi-user or single-user mode) with
one reboot following the installation.
Real administrators will read the README of any patch anyway, and
decided on their own whether a reboot is needed or not. Often patches
are for unused services, or they say reboot although a restart of
this particular service is enough.
> If you compare that to smpatch the advantages become clear quick;
> policies to allow you to define what to install and what not. And
> several options to download, download and install or only report new
> available patches.
PCA ( http://www.par.univie.ac.at/solaris/pca/ ) is doing most of what
smpatch does and it's just 25k of plain perl code, running on any
Solaris version, so it's not that hard to come up with something
that fits your local needs much better than anything Sun provides.
I just want to have a choice - and killing patch downloads via ftp/http
makes this impossible, and is just plain unnecessary.
mp.
-- Systems Administrator | Institute of Scientific Computing | Univ. of Vienna
- Next message: Martin Paul: "Re: Sun killing Solaris?"
- Previous message: mr_peter_stevenson_at_hotmail.com: "Re: Routing/IP question: Is there any way to do this?"
- In reply to: Lion-O: "Re: Sun killing Solaris?"
- Next in thread: Casper H.S. ***: "Re: Sun killing Solaris?"
- Reply: Casper H.S. ***: "Re: Sun killing Solaris?"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]