Description of problem: RHEL4U3 x86 will not install on IBM x460. No driver for SAS controller found. Version-Release number of selected component (if applicable): Default boot kernel in RHEL4U3 x86 How reproducible: Every time Steps to Reproduce: 1. Attempt to install RHEL4U3 x86 on IBM x460 scaled across 2 chassis 2. Watch as it finds no storage controller. 3. Pull out hair, call IBM to no avail, enjoy as they blame Red Hat for the issue 4. Fall back to RHEL4U2 & required driver disk from IBM's site: http://www-1.ibm.com/support/docview.wss?rs=0&dc=D400&q1=88726RU&uid=psg1MIGR-62897&loc=en_US&cs=utf-8&cc=us&lang=en Actual results: No install. I can install x86_64 with no problem, but it does not scale across multiple chassis as x86 does, and this is an 8-way scaled config in question Expected results: Drivers are _supposed_ to be in the base OS Additional info: None really. If there's some "secret sauce" I'm missing I'd love to hear it - but there's no data posted publicly that I can find that leads me to believe it's not broken. As an FYI - ALL firmware at latest rev.
indeed. we haven't included that driver due to stability concerns...as soon as the driver is ready we will include it. thanks.
Don, A bit of history might clear things up. There are two drivers for the Adaptec SAS 9410SAS controller: adp94xx and aic94xx. The adp94xx is a standalone driver that Adaptec developed for years. It is what you currently can download from IBM website. The aic94xx is a driver that IBM, Adaptec and Red Hat have been working on that has been integrated in mainline kernel. The reason for this work is that much of the adp94xx code functionality was spread in the mainline kernel and used by other drivers. No sense of putting in the adp94xx when much of the code is replicated in the kernel. So to remove some of the duplicate code that other drivers were using (LSI SAS for example), Adaptec decided that create a new driver - aic94xx. And since 2.6 is different from 2.5 and 2.4 (which the adp94xx driver can also support), some of the code logic changed as well. With that history in mind, Jason is correct in regards to the mainline driver (aic94xx) stability concerns. During U3 cycle the aic94xx driver was tested for inclusion and died in a flaming death. This was when the aic94xx just started being integrated in mainline. Since then Adaptec/IBM has been working on the mainline driver to make it work as best, or even better than the standalone driver (adp94xx). I am surprised to hear that the IBM link does not have the U3 driver, it definitly should. Let me contact the IBM person responsible for it and see if there has been a new tarball posted that somehow didn't show in "public" pages. I will post my findings on this bug.
Thank you for the detail. This will make it much easier to answer my customers concern.
The driver is being certified inside IBM. Not sure what that exactly means (as in, is this Red Hat certification, or some other ones). Will get more details.
The adp94xx driver is included as part of the beta of RHEL 4.4 and barring any unforeseen circumstances will be part of the final product. The beta is available for testing from RHN. Red Hat does currently certify storage solutions.
Closing bug and changing state to Next Release.