[SunRescue] SS5 upgrade

[Dave Reader]

On Wed, 16 May 2001, David Cantrell wrote:

> "Paul Khoury" <pkhoury2 at loop.com> wrote:
> > Reply-To: rescue at sunhelp.org
> > 
> > I ordered an SS5 on eBay, which I should be getting tomorrow.
> > It's intended to replace my SS2 webserver.

> >                                              How much work do I need
> > to do to get everything working on it, without reinstalling the OS?
> > I know I need a new kernel,
> 
> Ha!  So Linux/Sparc is good for at least one thing then :-)  No messing
> around with kernel architectures, as 4c, 4d and 4m are all treated the
> same.
> 
> And yes, my SS2 is running Linux.  With only 64Mb I deemed Solaris to be
> silly.  And I was *far* more comfortable with Linux at the time.

I have a Linux SS2/64Mb co-located. It's a nice stable machine, and a
breeze to maintain. It's not doing any heavy work, but working it is.

I suspect that solaris old enough to run on that hardware wouldn't be
worth my time. I do have 2.5.1 media - last time i installed it (on an
SS2) it took all night and then some (_just_ to run the installer). of
course it would then want endless additions and patches and nappies
changing before it was grown-up enough to go near the 'net.

> The SS1000 will, needless to say, be running Solaris, as none of the 
> *BSD or Linux crowd have adequate SMP support.

I've installed linux on an SMP SS10, but not really given it anything to
do yet. Even with only a single processor, it had a good go at putting the
SS5/110 to shame.

Is it just >2 processor which sucks, or should i expect my SS10 to
underperform too? I'm still using 2.2 kernels.

I should play with Sol8 a bit more on some of that hardware I guess.

As an interesting (or maybe not) deviation from the topic, a silly thread
came up on another list i take today.. about the size of /bin/true (and
how it really is bloated for something that does nothing). It was noted
that SunOS and Solaris /bin/true is a shell script which just exits, and
hence fires up a shell to do it. My linux boxen use a binary, so I looked
at how it varies by platform...

dave at dr:~$ file /bin/true
/bin/true: ELF 32-bit LSB executable, Intel 80386, version 1, dynamically
linked (uses shared libs), stripped
dave at dr:~$ ls -al /bin/true
-rwxr-xr-x    1 root     root         4416 Nov 22 21:04 /bin/true


dave at alpha:/space$ file /bin/true
/bin/true: ELF 64-bit LSB executable, Alpha (unofficial), version 1,
dynamically linked (uses shared libs), stripped
dave at alpha:/space$ ls -al /bin/true
-rwxr-xr-x    1 root     root         6760 Jan 29 17:35 /bin/true
dave at alpha:/space$ 


dave at serv2:~$ file /bin/true
/bin/true: ELF 32-bit MSB executable, SPARC, version 1, dynamically linked
(uses shared libs), stripped
8dave at serv2:~$ ls -al /bin/true
-rwxr-xr-x    1 root     root         9772 Oct 22  2000 /bin/true
dave at serv2:~$ ldd /bin/true
        libc.so.6 => /lib/libc.so.6 (0xe002b000)
        /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0xe0000000)



Is this typical of the differences in instruction sets, or would there be
poor compilation showing through there?

Those are all debian 2.2 linux boxen, so those will be built in pretty
much the same way from the same source in similar conditions. I'll have to
take a look at the MIPS and HPPA boxes when i get around to getting them
installed...

I wonder how that same source would compile with the native tools for
Solaris / Digital Unix (or whatever it's called these days).

d.