Can you not simply stop the systemd FSCK from running (I am asking the
obvious question and well as learning when the correct answer comes along.)
Tim H
On 12/9/24 14:31, Ralph Corderoy wrote:
Hi,
I've been asked to help with an old server which is using LVM and
a /dev/mapper RAID 1, mirror, with two disks. One of the disks has
hardware problems which keeps causing bus resets, delaying access to the
good drive and eventually causing lock ups.
Unplugging SATA power and data to that drive makes the kernel's dmesg
buffer much quieter and happier. The machine swiftly boots from the
remaining drive but then one of systemd's targets is to fsck the RAID
mirror and that fails as a drive is missing. systemd promptly stops
many of the things it had started and drops me into rescue mode.
I wish to boot fully with the degraded, one-disk, RAID ‘array’. I don't
want to put a new drive in and repair the array. I'd have thought
there'd be a simple command to confirm the degradation was expected.
There's ‘dmraid -x’ but I've found no good documentation, despite Red
Hat seeming to be the author, and suspect it deletes the whole array
rather than removing a disk from it.
Below is some output collected before the fault occurred which may be
useful.
An alternative is to alter the systemd configuration so the fsck always
passes or is no longer a dependency. Views on the practicality of that
are also welcome.
# dmraid -r # format, RAID level, sectors used, data offset.
/dev/sdc: pdc, "pdc_ghjgcbej", mirror, ok, 3906249984 sectors, data@ 0
/dev/sdb: pdc, "pdc_ghjgcbej", mirror, ok, 3906249984 sectors, data@ 0
#
# dmsetup table
pdc_ghjgcbej5: 0 3905746944 linear 252:1 501760
xserver-root: 0 3885056000 linear 252:3 384
xserver-swap_1: 0 20684800 linear 252:3 3885056384
pdc_ghjgcbej: 0 3906249984 mirror core 2 131072 nosync 2 8:16 0 8:32 0 1
handle_errors
pdc_ghjgcbej2: 0 2 linear 252:1 501758
pdc_ghjgcbej1: 0 497664 linear 252:1 2048
#
# dmsetup status
pdc_ghjgcbej5: 0 3905746944 linear
xserver-root: 0 3885056000 linear
xserver-swap_1: 0 20684800 linear
pdc_ghjgcbej: 0 3906249984 mirror 2 8:16 8:32 29803/29803 1 RA 1 core
pdc_ghjgcbej2: 0 2 linear
pdc_ghjgcbej1: 0 497664 linear
#
/dev/mapper:
crw-rw---- 1 root root 10, 59 2018-03-11 09:27 control
brw-rw---- 1 root disk 251, 0 2018-03-11 09:27 pdc_ghjgcbej
/boot brw-rw---- 1 root disk 251, 1 2018-03-11 09:27 pdc_ghjgcbej1
lvm2 brw-rw---- 1 root disk 251, 2 2018-03-11 09:27 pdc_ghjgcbej5
/ brw-rw---- 1 root disk 251, 3 2018-03-11 09:29 xserver-root
swap brw-rw---- 1 root disk 251, 4 2018-03-11 09:27 xserver-swap_1
parted:
Model: Linux device-mapper (mirror) (dm)
Disk /dev/mapper/pdc_ghjgcbej: 2000GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Number Start End Size Type File system Flags
1 1049kB 256MB 255MB primary ext2 boot
2 257MB 2000GB 2000GB extended
5 257MB 2000GB 2000GB logical lvm
lsblk:
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
...
sdb 8:16 0 1.8T 0 disk
└─pdc_ghjgcbej 252:0 0 1.8T 0 dmraid
├─pdc_ghjgcbej1 252:1 0 243M 0 part /boot
├─pdc_ghjgcbej2 252:2 0 1K 0 part
└─pdc_ghjgcbej5 252:3 0 1.8T 0 part
├─xserver-root 252:4 0 1.8T 0 lvm /
└─xserver-swap_1 252:5 0 9.9G 0 lvm
sdc 8:32 0 1.8T 0 disk
└─pdc_ghjgcbej 252:0 0 1.8T 0 dmraid
├─pdc_ghjgcbej1 252:1 0 243M 0 part /boot
├─pdc_ghjgcbej2 252:2 0 1K 0 part
└─pdc_ghjgcbej5 252:3 0 1.8T 0 part
├─xserver-root 252:4 0 1.8T 0 lvm /
└─xserver-swap_1 252:5 0 9.9G 0 lvm
--
Next meeting: Online, Jitsi, Tuesday, 2025-01-07 20:00
Check to whom you are replying
Meetings, mailing list, IRC, ... https://dorset.lug.org.uk
New thread, don't hijack: mailto:dorset@mailman.lug.org.uk
