Well I wanted to catch up on the mail-list prior to responding (2 weeks gets
a lot of posts!) but this particular thread is of great interest to me.  I
agree with many of the prior posts and wish to add my $0.02:
*    We are currently running 2 zSeries CECs, each with an IFL; other than
the IFLs they are OS/390 (soon to be z/OS) running multiple LPARs in a
data-sharing geoplex.  Each IFL has a z/VM license to support running Linux.
*    Primary reason for going to L/390 was to support DB2 Connect as
middleware between our Sun/AIX/NT webserver thingies (mostly BEA on Solaris)
and our OS/390 DB2 backend dbases.  Hope was that we would spin this success
story to possibly drive some server consolidation of
low-CPU/high-availability stuff like DNS/DHCP from stand-alone boxes to
L/390 and then perhaps some SAMBA fileservers, etc.
*    Current implementation (only very recently full production) is SUSE 7.0
*    We had just about settled on SUSE as "the way" based upon a number of
reasons, highest being their early support of 390 and the critical mass we
believed they had achieved among 390 sites.
*    We have a solid S/390 support staff very skilled in MVS-OS/390.  More
limited support experience with VM but enough of us with skills, be they
rusty (VM/SP) to pick up this new fangled z/VM stuff with some education and
manuals.  We also have a pretty good (more than VM) base of support staff
(me included, even though I'm now management) that have been running various
iterations of Linux for years in one way or another for at least hobbyist
reasons.
*    Even though I felt that the combo of our Linux/VM/390 experience was
probably enough to be self-supporting, combined with the excellent on-line
Linux resources available which of course includes this mailing list, I was
still leaning heavily towards purchasing a support contract if for no other
reason than to have a nice fuzzy safety net.
*    Then SUSE came out with their latest release, with their policy of only
bundling the distro with support.  I have a hard time with being forced to
pay out 5 figures for software I'm not even certain works in my environment.
*    So now I'm probably going to ask the folks that work on this to obtain
and eval the Red Hat distro ("Ahh . . . the power of VM!") and see how that
goes.
*    Three months ago if you had asked me what the chances were of me
switching from SUSE to Red Hat I'd have said probably slim and none.  Now
I'd say it's probably quite likely.
Sorry for the length, just wanted to spell out how SUSE's decision has had a
potential business impact and how I agree with most of the other posts about
it being essential that customers, existing or potential, be able to at
least evaulate distro's.  I was quite suprised at one of the responses that
questioned why the poster even needed Linux on 390 and that they could have
easily eval'd the same distro on Intel.  How does one eval an OSA/Express
GBe card and associated support on Intel, for example?  It's quite important
to conduct evals in an environment as close as possible, if not identical,
to the one you wish to eventually deploy and it's almost always a "try
BEFORE you buy" proposition.

Anyway, that's my $0.02 on the matter.

Reply via email to