Raid with a very, very old server

circus78 · 05-21-2019, 03:10 AM

Hi,
on a very old Slackware, I got this situation:

Code:

# cat /proc/mdstat 
Personalities : [linear] [raid0] [raid1] [raid5] 
read_ahead 1024 sectors
md0 : active raid1 hdc1[1]
      79492288 blocks [2/1] [_U]
      
unused devices: <none>

I checked with smartctl, hda is actually broken:

Code:

# smartctl -a /dev/hda
smartctl version 5.1-7 Copyright (C) 2002 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF INFORMATION SECTION ===
Device Model:     Maxtor 6Y080L0                          
Serial Number:    Y2BJZQZC            
Firmware Version: YAR41BW0
ATA Version is:   7
ATA Standard is:  Unrecognized. Minor revision code: 0x1e
Local Time is:    Tue May 21 09:32:17 2019 CEST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

..
..
Self-test execution status:      ( 118) The previous self-test completed having
                                        the read element of the test failed.

so I need to replace it.

First of all, how can I be sure that server can boot with just /dev/hdc (LILO is used in this case)?

Should I first install LILO in /dev/hdc before? How?
Thank you!

business_kid · 05-21-2019, 01:35 PM

/dev/hdc, boy, that brings bac (bad) memories.

I'd check for parts availability first. We're talking IDE disks here, right? It's a long time since that stuff was made.
2nd, I'd do a full backup before you do anything

circus78 · 05-22-2019, 05:52 AM

Hi business_kid, yes, it is IDE.
I do have disk replacement.
My question was: how can I be sure that system will boot with hdc only?
I wondering if I can make a bootable floppy disk, just in case.
Thank you

syg00 · 05-22-2019, 06:23 AM

Quote:

Originally Posted by circus78

My question was: how can I be sure that system will boot with hdc only?
I wondering if I can make a bootable floppy disk, just in case.

Whoa - haven't been able to do that in decades ...

The short answer is you can't be sure without actually booting. Even a dump of the MBR can only show if it has boot code (I presume lilo has an eyecatcher, I haven't used it since early this century), not if it will actually boot.
A boot medium (floppy, CD, whatever) will allow you to chroot into the system if it fails. Better to just update the lilo.conf and run the lilo command before you shutdown.

business_kid · 05-23-2019, 06:38 AM

Here's a ready made boot floppy. It saved my a** more often than I care to remember back in the bad old days.

Tomsrtbt linux

That's about the only way now to make a boot floppy. IIRC, it's kernel 2.0 or maybe 2.2, & busybox. It hasn't been updated in decades. It will see ext4 as ext2 :-/. You need an excellent floppy; it reformats it to 1.82MB, and writes. Best to do the minimum with disks, but if you run mke2fs on them in tomsrtbt, you can update to ext3/4 with tune2fs. The guy was fond of lua, and may have some compiler for that or a very early cc.

syg00 · 05-23-2019, 07:35 AM

A boot floppy doesn't need to invoke a kernel ...
Some years ago I had a laptop that refused to boot after a normal (Ubuntu) update. The installed system was fine, but the BIOS wouldn't see the disk. The only way I was able to get it to boot was to create a grub legacy boot disk and interrupt the boot then point it to the on-disk system.
Serious PITA but worked. Until I got pissed off and tossed it.

Maybe lilo can do similar, I don't know.