Test results from I2O installation... failure.
I rebuilt my own tree with rawhide + modified kudzu + grub + lvm2. Booted x86_64 boot.iso.
<jeremy> warren_laptop: I haven't moved them yet... still staring at anaconda. but going cross-eyed, so think I'll move them and just finish looking at anaconda in the morning <jeremy> you'll get told that there are no hard drives in the loader, but then if you continue, you should see them in disk druid
Anaconda says there are no hard drives in the loader, and when I go to Disk Druid it says this fatal message, "An error has ocurred - no valid devices were found on which to create new file systems. Please check your hardware for the cause of this problem."
In ALT-F2, kudzu-probe displays this among the devices:
Scanning for SCSI Controllers Device Distributed tech|SmartRAID V Controller: Desc: Distributed Tech|SmartRAID V Controller Driver: dpt_i2o Device: None
kudzu still knows about the old 2.4 kernel dpt_i2o driver, while in this kernel we are using i2o_block. I suspect kudzu or hwdata needs changing, and kudzu needs more changing to work properly here. I am working more on this again in several hours after I fix a few xorg-x11 upgrade issues.
Warren Togami wtogami@redhat.com
On Thu, 2004-05-06 at 21:30 -1000, Warren Togami wrote:
kudzu still knows about the old 2.4 kernel dpt_i2o driver, while in this kernel we are using i2o_block. I suspect kudzu or hwdata needs changing, and kudzu needs more changing to work properly here. I am working more on this again in several hours after I fix a few xorg-x11 upgrade issues.
Will it work for update ? As in up2date -u kernel, and the install script to change /etc/modprobe.conf.
Is there any doc for the i2o block driver? I was being misguided and I put i2o_scsi instead of dpt_i2o in modprobe.conf and it failed on me. Should it work with i2o_block ?
Dumitru Ciobarcianu wrote:
On Thu, 2004-05-06 at 21:30 -1000, Warren Togami wrote:
kudzu still knows about the old 2.4 kernel dpt_i2o driver, while in this kernel we are using i2o_block. I suspect kudzu or hwdata needs changing, and kudzu needs more changing to work properly here. I am working more on this again in several hours after I fix a few xorg-x11 upgrade issues.
Will it work for update ? As in up2date -u kernel, and the install script to change /etc/modprobe.conf.
Other changes in your config will be needed because your block devices change from /dev/sd* to /dev/i2o/hd*.
Is there any doc for the i2o block driver? I was being misguided and I put i2o_scsi instead of dpt_i2o in modprobe.conf and it failed on me. Should it work with i2o_block ?
i2o_scsi is almost never used because it is for accessing disks directly. i2o_block is used for accessing the logical block devices of your I2O RAID controller. If you have disabled RAID arrays, then these block devices become the individual disks.
Note that only 351 and later kernels are patched with Markus Lidel's newer fixes. All previous kernels have severe problems if your RAID controller has more than a single logical array.
http://i2o.shadowconnect.com/ One more patch for i2o_proc available at Markus' I2O homepage, along with other information.
Warren Togami wtogami@redhat.com
Dumitru Ciobarcianu wrote:
On Thu, 2004-05-06 at 22:53 -1000, Warren Togami wrote:
Other changes in your config will be needed because your block devices change from /dev/sd* to /dev/i2o/hd*.
Shouldn't LABEL= take care of that ?
If the kernel can't see the device because the driver is not loaded, LABEL= is not enough. For this reason you would need to manually run mkinitrd to preload the right modules.
Warren