[TriLUG] Linux on SATA?
jbrigman at nc.rr.com
Sun May 29 21:16:26 EDT 2005
Thanks for your rapid reply, I appreciate hearing about your
experiences! Many folks in my quandry are doubting that the sata_nv
driver (nVidia SATA driver) actually works, and it may be the case that
I've tried the SATA drives both in the RAID and non-RAID configuration,
same results. I've seen some people who are using other SATA controllers
solve the problem by putting the SATA drives in a RAID configuration,
but that did not work for me. (And Jason, I agree 100% about using
kernel-RAID rather than "fakeraid" is it's often referred to.)
Chris: What I mean by switching to PATA is that I can abandon the SATA
drives and go back to some ATA133 drives because the motherboard has
both ATA133 IDE interfaces (the usual two) and dual SATA headers that
happen to be RAID-able. Have you gotten SATA drives to work with any
It does not surprise me that you might have several systems running RHEL
on SATA without problems: the diversity of SATA controller designs is,
well, shocking. I thought the world would have learned from the terrible
MegaIDE stuff back a few years ago in the ServerWorks chipset, but
apparently chaos doesn't matter as long as there's a windows driver
supplied by the mfr. :-(
On Sun, 2005-05-29 at 17:04 -0700, Chris Bullock wrote:
> Can you please clarify, that you are using SATA drives with the SATA
> cables, you refer to switching back to PATA and everything is well, just
> want to make sure you are not using PATA drives for a SATA RAID card. We
> have several systems running RHEL on SATA without any problems. Sorry I
> couldn't answer the question though.
> Good luck,
> --- James Brigman <jbrigman at nc.rr.com> wrote:
> > I'm having a problem getting Linux to install on a system with built-in
> > SATA RAID. The mbd. is that low-cost "Chaintech 7NJL6". It's an nVidia
> > nForce 2 board with the built-in MCPS controller system. The board has
> > The particular problem is that sata_nv driver won't "see" the drives on
> > the SATA interfaces, even though they are A-OK in the BIOS.
> > bug, however, that's manifest in the kernels used by SuSE 9.2 and
> > CentOS4. http://linuxmafia.com/faq/Hardware/sata.html
> > There's been some progress on this bug, found in Kernel 2.6.8.x as
> > http://bugme.osdl.org/show_bug.cgi?id=3352
> > The particular bug is with the ATA_FLAG_SATA_RESET: removing the
> > variable from the header of the driver is reported to solve the problem.
> > My questions to the group are:
> > 1) Are there any magic incantations that could be typed at the
> > bootloader prompt such that bugs in the sata_nv driver could be
> > sidestepped? (SuSE 9.2/CentOS 4 x86) Or, similarly, anyone know anything
> > about getting a fixed sata_nv driver, dropping it on a floppy, and
> > booting with an amended driver?
> > 2) Has anyone taste-tested SuSE 9.3 on SATA yet, an particularly, nVidia
> > SATA? Can anyone out there report success with SuSE 9.3 on built-in SATA
> > controller hardware?
> > 3) Is anyone getting ANY distro to work with nVidia-based mbds with
> > SATA? Or with the Chaintech 7NJL6 working under ANY distro? This board
> > has been on sale at TigerDirect for some time, and costs $40 or less,
> > after rebate. So I figure someone in the local area has, at least, seen
> > this problem too.
> > All experienced feedback welcome. Tangential comments not related to the
> > solution to /dev/null, please.
More information about the TriLUG