First of all, I apologise for any shortcomings in this report; I know
little about Linux or Unix.
RH Linux 7.0 hard disk installation apparently proceeds as normal at first.
* The "Loading mnt/runtime ramdisk..." announcement appears and disappears.
* The "Loading mnt/runtime/usr ramdisk..." announcement appears and
* In expert mode, the "special drivers" dialog appears and disappears when
Running anaconda -- may take some time to load...
appears at the bottom of the screen. A few seconds later, it is
install exited abnormally -- received signal 9
sending termination signals...done
sending kill signals...done
you may safely reboot your system
I have "successfully" reproduced this bug every time I attempt to install
on my PC. Using or omitting any of these boot options:
has no visible effect on it. Nor does changing the language or keyboard
Neither /tmp/syslog nor <the output to the alt-F3 console> show any
obvious relevant error messages. Here is the /tmp/syslog produced in a
typical install attempt:
<4>Linux version 2.2.16-22BOOT (firstname.lastname@example.org) (gcc version
egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)) #1 Tue Aug 22 15:35:29
<4>Detected 119753 kHz processor.
<4>Console: colour VGA+ 80x25
<4>Calibrating delay loop... 238.39 BogoMIPS
<4>Memory: 13572k/16384k available (1016k kernel code, 416k reserved, 548k
data, 60k init, 0k bigmem)
<4>Dentry hash table entries: 262144 (order 9, 2048k)
<4>Buffer cache hash table entries: 16384 (order 4, 64k)
<4>Page cache hash table entries: 4096 (order 2, 16k)
<4>CPU: Intel Pentium 75 - 200 stepping 0c
<6>Checking 386/387 coupling... OK, FPU using exception 16 error reporting.
<6>Checking 'hlt' instruction... OK.
<6>Checking for popad bug... OK.
<6>Intel Pentium with F0 0F bug - workaround enabled.
<4>POSIX conformance testing by UNIFIX
<4>PCI: PCI BIOS revision 2.10 entry at 0xfd8a1
<4>PCI: Using configuration type 1
<4>PCI: Probing PCI hardware
<6>Linux NET4.0 for Linux 2.2
<6>Based upon Swansea University Computer Society NET3.039
<6>NET4: Unix domain sockets 1.0 for Linux NET4.0.
<6>NET4: Linux TCP/IP 1.0 for NET4.0
<6>IP Protocols: ICMP, UDP, TCP
<4>TCP: Hash tables configured (ehash 16384 bhash 16384)
<4>Starting kswapd v 1.5
<6>Detected PS/2 Mouse Port.
<6>Serial driver version 4.27 with no serial options enabled
<6>ttyS00 at 0x03f8 (irq = 4) is a 16550A
<6>ttyS01 at 0x02f8 (irq = 3) is a 16550A
<6>ttyS02 at 0x03e8 (irq = 4) is a 16550A
<4>pty: 256 Unix98 ptys configured
<4>RAM disk driver initialized: 16 RAM disks of 4096K size
<6>loop: registered device at major 7
<4>PIIX3: IDE controller on PCI bus 00 dev 39
<4>PIIX3: not 100% native mode: will probe irqs later
<4> ide0: BM-DMA at 0xffa0-0xffa7, BIOS settings: hda:pio, hdb:pio
<4> ide1: BM-DMA at 0xffa8-0xffaf, BIOS settings: hdc:pio, hdd:pio
<4>hda: Seagate Technology 1275MB - ST31276A, ATA DISK drive
<4>hdb: QUANTUM FIREBALLlct15 15, ATA DISK drive
<4>hdc: TOSHIBA CD-ROM XM-5602B, ATAPI CDROM drive
<4>ide2: ports already in use, skipping probe
<4>ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
<4>ide1 at 0x170-0x177,0x376 on irq 15
<6>hda: Seagate Technology 1275MB - ST31276A, 1221MB w/0kB Cache,
<6>hdb: QUANTUM FIREBALLlct15 15, 14324MB w/418kB Cache, CHS=1826/255/63
<4>hdc: ATAPI 8X CD-ROM drive, 256kB Cache
<6>Uniform CD-ROM driver Revision: 3.10
<6>Floppy drive(s): fd0 is 1.44M
<6>FDC 0 is a National Semiconductor PC87306
<6>md driver 0.90.0 MAX_MD_DEVS=256, MAX_REAL=12
<6>raid5: measuring checksumming speed
<4> 8regs : 122.301 MB/sec
<4> 32regs : 89.916 MB/sec
<4>using fastest function: 8regs (122.301 MB/sec)
<4>scsi : 0 hosts.
<4>scsi : detected total.
<4>md.c: sizeof(mdp_super_t) = 4096
<4> hda: hda1 hda2 < hda5 > hda4
<4> hdb: hdb1 hdb2 hdb3 hdb4 < hdb5 >
<5>RAMDISK: Compressed image found at block 0
<4>VFS: Mounted root (ext2 filesystem).
<4>EXT2-fs warning: maximal mount count reached, running e2fsck is
<6>raid0 personality registered
<6>raid1 personality registered
<6>raid5 personality registered
I scribbled down the last few messages from the alt-F3 console from the
same install attempt:
* partition /dev/hdb3 selected
* mounting device hdb3 as ext2
* created inode
* done loading 2969600 bytes
* created inode
* done loading 4194304 bytes
* got url hd://hdb3:ext2/rh7/i386/en
* found url image hd://hdb3:ext2/rh7/i386/en
* getting ready to spawn shell now
* going to insmod raid0.o (path is NULL)
* going to insmod raid1.o (path is NULL)
* going to insmod raid5.o (path is NULL)
* going to insmod ext3.o (path is NULL)
* looking for USB mouse...
The /proc/cmdline from that attempt was:
initrd=initrd.img devfs=nomount BOOT_IMAGE=vmlinuz mem=16M expert text
All the information about my computer in the /tmp/syslog above appears to
be correct. Additionally, I can tell you that it is an AST Advantage! 7302
(with a new HD and a replaced modem) and the BIOS version is 1.00.05CW0Q -
unfortunately I don't have a name for the motherboard or chipset.
Possible problems I have considered include:
1) Corrupt files? The md5sums of the copies of these files
in /dev/hda3 appear to be correct (or at least, they are the same
as those on ftp.ibiblio.org). I have created new boot floppies from
boot.img a number of times.
2) Bad hardware? My computer is successfully running both Windows 95 (OEM
release 1) and Red Hat 6.0, which was successfully (re)installed (from CD,
without a boot floppy) a few days ago. All the RAM is good. Both hard
drives have been operating successfully under both OSes. The partitions on
both were successfully detected by the 7.0 installer, and I was able to
mount, unmount and use partitions on both drives from the alt-F2
console's /bin/sh prompt without problems. OTOH, the IDE controller with
the hard drives is sometimes very slow to respond to the BIOS at boot, and
I've had two corrupt ext2 superblocks recently (why I don't know).
3) A software bug? I've tried running /usr/bin/anaconda from the alt-F2
console's /bin/sh while the "special drivers" dialog was still up in the
main console. The result was (copied by hand):
error 2 reading header: cpio: Bad magic
cpio failed on (null): (internal)
(Entering /usr/bin/anaconda for a second time produces another
segfault, with a much longer error message. If you try /usr/bin/anaconda
and then proceed with the install, the same longer error message appears
in the main console, and the install aborts on signal 11, not 9.)
4) Not enough RAM? I know that 16MB is pitiful these days, but don't tell
me that you can't even *install* without more. If you can't, the installer
should at least make that clear. There certainly shouldn't be any shortage
of swap; /dev/hdb2 is a 128Mb Linux swap partition.
Please don't hesitate to email me if there is anything else I can add or
I made a small typo in the original description - the files with correct
md5sums are in /dev/hdb3 (not /dev/hda3), along with all the rest of the
I agree the installer ought to do saner things but I suspect you are out of ram
with 15.5Mb. Try a text mode install.
Bug left open since the 16Mb warning is a very valid point
Unfortunately, a text-mode installation seems to be impossible: text is one of
the boot: options I have tried without seeing any change (see above). If this
means that RHL 7.0 really can't be installed without >16MB RAM, then I would
suggest that the online documentation needs changing: there doesn't seem to be
any minimum physical RAM figure given for RHL 7.0 on RH's website. The nearest
I could find was one for RHL 6.2, which was 16Mb. Also, Chapter 5 of the 7.0
Reference Guide says that "If your computer has 16Mb of RAM or less, you must
create a swap partition.", implying that 16Mb RAM is enough.
The text installer requires at least 20 megs to run. It is a bug it did not
inform you that 16 megs was not going to be enough.
I can confirm that diagnosis. RHL now installs successfully after an upgrade to
If you don't mind, what does /proc/meminfo report if you go back to the original
memory configuration that was causing problems?
Here is a typical /proc/meminfo from an installation attempt using the old
(16MB) memory configuration, taken when the expert mode "special drivers" dialog
total: used: free: shared: buffers: cached:
Mem: 14688256 14311424 376832 446464 9027584 1290240
Swap: 0 0 0
MemTotal: 14344 kB
MemFree: 368 kB
MemShared: 436 kB
Buffers: 8816 kB
Cached: 1260 kB
BigTotal: 0 kB
BigFree: 0 kB
SwapTotal: 0 kB
SwapFree: 0 kB
The /proc/cmdline from that attempt was
expert initrd=initrd.img devfs=nomount BOOT_IMAGE=vmlinuz
Added a really nasty "your system will now reboot" message for unsupported