On Wed, May 11, 2016 at 07:20:16PM +0000, Albin Otterhäll wrote:
I'm trying to install Debian on a Lenovo Thinkpad T430. This is not the
first time I install Debian on this machine, no problems earlier. The
problem is that the installer crash at the end of the "[s]elect and install
packages" step. All the packages have been downloaded and most of them are
installed; the installer crash about 85 percent into the installation. I've
tried to redo the step and the whole installation, nothing helps. I've
verified the ISO with GPG and SHA512sums and done an integrity check of the
USB disk.
This is the last part of the error log (/var/log/syslog):
Your problem is the disk, the cable, or the USB port.
kernel: [...] Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
kernel: [...] sd 0:0:0:0: [sda] CDB:
kernel: [...] Write(10): 2a 00 00 17 c0 00 00 00 08 00
kernel: [...] EXT4-fs error (device dm-1): ext4_journal_check_start:56:
Detected aborted journal
This. Which led to the kernel remounting it read-only:
'/usr/bin/scrollkeeper-rebuilddb.distrib': Read-only file system
and this will bite you later:
lilo-installer: LILO not usable on EFI PCs without BIOS compatibility;
use grub-efi
The answers I got on IRC suggests that it was something wrong with my SSD.
So I tested it with all the tools I know.
First I run an extended self-test with S.M.A.R.T. tools. Everything comes
up OK. No fails. I when run badblocks:
# badblocks -vsw /dev/sda
Checking for bad blocks in read-write mode
From block 0 to 234431063
[---]
Pass completed, 0 bad blocks found. (0/0/0 errors)
It's suggested to me to test by write random data to a file in a similar
environment to "emulate" the debian-installer. This is what I did:
1. Create new GPT partition table
2. Create new partition labeled for LVM with recommended start and end
(+1MiB to last block)
3. Encrypt the partition with LUKS.
4. Create physical volume, volume group and a logical volume spanning the
whole partition.
4. Create EXT4 file system on logical volume.
6. Create file 'foo' and write to it with:
# dd bs=4M if=/dev/urandom of=foo status=progress
236059623424 bytes (236 GB, 220 GiB) copied, 13192.2 s, 17.9 MB/s
dd: error writing 'foo': No space left on device
56284+0 records in
56283+0 records out
236072116224 bytes (236 GB, 220 GiB) copied, 13192.8 s, 17.9 MB/s
Have you any suggestions on what I can do?
Return the disk, get a new one. This one is defective.
-dsr-