FC4 rawhide 20050518 fails I2O grub install. This seems to have been broken throughout most of FC4's development cycle. ALT-F5 says this after installation is complete: grub> root(hd0,0) Filesystem type is ext2fs, partition type 0x83 grub> install /grub/stage1 d (hd0) /grub/stage2 p (hd0,0)/grub/grub.conf Error 22: No such partition My test install today was with Auto-Partitioning and LVM, but this happens with non-LVM installs too. I will attempt to get more data with a non-LVM simple install on a single partition tomorrow. Markus Lidel has more test information.
Note: I'm not convinced this is a blocker since the hardware is not at all common
http://videl.ics.hawaii.edu/~warren/temp/anacdump.txt Anaconda crash with dump when clicking on Next on the boot loader screen. This happened with Rawhide 20050519 but not 20050518. pjones said to attach it here anyway.
pjones found that the bug in #2 was already fixed in tomorrow's rawhide. He suggested "cp anaconda/iw/network_gui.py RHupdates" but it didn't seem to work. When was the last time this was tested? I managed to workaround the problem by manually copying network_gui.py into /tmp/updates/
Oh, msw said RHupdates doesn't work for FTP.
Created attachment 114601 [details] /proc/partitions
Created attachment 114602 [details] anaconda.log
Created attachment 114603 [details] anaconda-ks.cfg
Created attachment 114604 [details] install.log
Created attachment 114605 [details] install.log.syslog
Created attachment 114606 [details] grub.conf
Created attachment 114607 [details] device.map
Apache screwed up the device.map, it actually looks like this: # this device map was generated by anaconda (fd0) /dev/fd0 (hd0) /dev/i2o/hda
/sbin/grub-install /dev/i2o/hda /dev/root: Not found or not a block device. grub-install fails when done from chroot after installation is complete. But it worked after reboot into rescue mode. Testing a reinstall now without modifying the partition table.
Failed after reinstall (without modifying the partition table)
Hopefully fixed in grub-0.95-13 , building for rawhide now.
I suspect this might need to be backported to RHEL4. I will test. * Thu May 19 2005 Peter Jones <pjones> 0.95-13 - don't treat i2o like a cciss device, since its partition names aren't done that way. (#158158) diff -ur grub-0.94/lib/device.c grub-0.94.new/lib/device.c --- grub-0.94/lib/device.c 2004-05-07 04:50:36.375238696 +0200 +++ grub-0.94.new/lib/device.c 2004-05-07 04:48:57.611253104 +0200 @@ -419,6 +419,12 @@ { sprintf (name, "/dev/rd/c%dd%d", controller, drive); } + +static void +get_i2o_disk_name (char *name, int unit) +{ + sprintf (name, "/dev/i2o/hd%c", unit + 'a'); +} #endif /* Check if DEVICE can be read. If an error occurs, return zero, @@ -789,6 +795,26 @@ } } } + + /* I2O disks. */ + for (i = 0; i < 8; i++) + { + char name[16]; + + get_i2o_disk_name (name, i); + if (check_device (name)) + { + (*map)[num_hd + 0x80] = strdup (name); + assert ((*map)[num_hd + 0x80]); + + /* If the device map file is opened, write the map. */ + if (fp) + fprintf (fp, "(hd%d)\t%s\n", num_hd, name); + + num_hd++; + } + } + #endif /* __linux__ */ /* OK, close the device map file if opened. */