What am I going to replace my L1000(x4) and rp5430 (x1) Servers with??

[Leon Allen]

Our crop of old PA-RISC L100 series servers are ear-maked for elimination this financial year.

I'm interested in hearing peoples opinions on what to replace them with.

More PA-RISC; Itanium; Xenon; HPUX; RedHat; 32 bit; 64 bit; Windows Servers 2003 64; VMWare; and what's this vista thing?????????

Hot contender at the moment are a couple of dual core Xenon based servers running windows server 2003 64 bit.

Anybody got a tale to tell?

Thanks!

Leon

[Luk Vandenbussche]

Leon,

First check your application.
On which os and hardware are they supported.

[Leon Allen]

Prabably the best tip to be sure, and one I have already undertaken to do.

Thx

[rick jones · Accepted solution]

Don't underestimate the value of your built-up institutional knowledge of HP-UX.

In terms of maintaining the principle of least change, go with an rp34XX or rp44XX and HP-UX.

You could go instead to Itanium (rxNNNN) and HP-UX and likely as not could still run legacy PA-RISC applications under the Aries emulator. You also then have the option of going to Linux, Windows Server 2003 or OpenVMS on the same hardware. (There is also FreeBSD for Itanium but that isn't officially supported by HP, and I think the Integrity NonStop systems are different boxes).

The Itanium boxes would allow you to run HP VM - HP's virtual machine software for Itanium systems. Guest OSes would include UX, Linux, Windows and I suspect OpenVMS - might be a staggered set of dates for support.


FWIW it is Xeon, not Xenon :)

[Leon Allen]

I like the bit about 'don't underestimate built-up institutuional knowledge'. I'll bear that in mind.

The Aries emulator is something I didn't know about, and could ensure current apps continue to run on Itanium.

And HP VM sounds like it's worth investigating to.

Thanks for your input Rick.

[Bill Hassell]

And to further expand the consideration for non-HP-UX solutions: Don't underestimate the labor and time costs for conversion. Depending on how well (or poorly) your applications are written, conversion could be a matter of weeks to a matter of months (or more). Even Unix-to-Unix (as in HP-UX to Linux-something) may require weeks of coding, testing and perhaps even several hardware changes. It is far too easy to look at only hardware and OS costs but overlook the hidden costs of conversion. Then there's issues with stability, supportability, security...