Re: Why are company's not certifying Debian?
On Sat, 30 Oct 2004, Rishi wrote:
> > the server you buy, is herein "blessed/certified to work with debian"
> > - see the incompatibility list as what was previously
> > posted for what is known NOT to work
> joking apart. :-) .. really what I meant was once I tried to install
> Debian Woody on an HP server and it was unable to do so ... meaning...
> it did not have the SCSI controller driver to install the OS on the
> HDD. However, with Red Hat, I was able to download an RPM which
> contained the SCSI drivers for it load at startup.
now we're getting somewhere
scsi ... not all installers know how to handle scsi drive properly
- its the distro installer's kernel problem, that it has to
install the scsi driver for the scsi controller you have
( and usually the problem of the person doing the installing )
everything reasonable works ... you just have to fiddle with this and
tweek that and fumble/grumble about foo widgets to get it all working
in all hardware issues: ( onboard or pci cards )
- how fast can the cpu be, what speed FSB, how much L2 cache
- how much memory, what kind, what speed
- you need to know which IDE chipset it uses
- you need to know which NIC chipset it uses
- you need to know which SVGA chipset it uses
- you need to know which sound chipset it uses
( unless you dont care about sound )
don't buy ANY hardware untill you can answer those 6 basic questions
or that the vendor says that it will guarantee that xxx-OS
will work on foo hardware combination
> In the same way right now I was hoping to buy this server and use the
> built in features like Hardware RAID and Mobile / pager notification
> when disk fails.
i know people that ran a $40K/month pager bill because of being paged
- why the !@#$% do you want to be paged ...
- the hardware should NEVER die within reason ...
( it's okay for it to die of old age at 3 yrs or 5 yrs later...)
with 2 disk ... what's the point of hw raid ???
- the machine has to come down ... if you are using ide disks
- the machine should stay down, until the new replacement disk
is sync'd ... and that can be a whole day ...for a full 200GB
- if you cannot be down for more than 5 minutes... you should have 2
complete independent systems ( properly configured and tested for
high-availability ... ) which has NOTHING to do with raid or mirror'ing
- if you cannot afford the extra hardware and extra time
to configure, than that extra hw costs should be
realistically weighed agaist how much $$$ is lost due to
the machine failure