| View previous topic :: View next topic |
| Author |
Message |
jpsollie
Guru
Joined: 17 Aug 2013
Posts: 324
|
 Posted: Sat Aug 19, 2023 4:45 am Post subject: [SOLVED] mdadm & acpi S3: what are the options |
 |
|
I'd like to set a gentoo-based NAS device in acpi S3 during the night (and send a magic packet when I need it again) to save power.
Unfortunately, there's an issue with one of the drives when resuming:
| Code: |
[ 140.017840] md/raid:md1: Disk failure on sdan, disabling device.
[ 140.017843] md/raid:md1: Operation continuing on 9 devices.
[ 140.017879] md1_raid6: attempt to access beyond end of device
sdaq: rw=395265, sector=8, nr_sectors = 1 limit=0
[ 140.017892] md: super_written gets error=-5
[ 140.017894] md/raid:md1: Disk failure on sdaq, disabling device.
[ 140.017896] md/raid:md1: Operation continuing on 9 devices.
[ 140.017919] md1_raid6: attempt to access beyond end of device
sdap: rw=395265, sector=8, nr_sectors = 1 limit=0
[ 140.017922] md: super_written gets error=-5
[ 140.017923] md/raid:md1: Disk failure on sdap, disabling device.
[ 140.017925] md/raid:md1: Operation continuing on 8 devices.
|
it looks like md subsystem is trying to access the disks before they are actually ready. It kicks out the device, whereas it should simply wait a bit and it will be ok.
Anybody who knows some tricks here to "guide" the kernel when resuming from S3?
_________________
The power of Gentoo optimization (not overclocked): [img]https://www.passmark.com/baselines/V10/images/503714802842.png[/img]
Last edited by jpsollie on Thu Aug 24, 2023 10:25 am; edited 1 time in total |
|
| Back to top |
|
 |
jpsollie
Guru
Joined: 17 Aug 2013
Posts: 324
|
| Posted: Thu Aug 24, 2023 10:24 am Post subject: |
|
|
I kinda found a solution myself. Just posting here in case anybody would like to see:
elogind has a way to specify what needs to be done before and after suspend.
So I made sure all arrays are properly unmounted before and afterwards remounted:
| Code: |
linuxserver # cat /lib64/elogind/system-sleep/hddtimings.sh
#!/bin/bash
case $1/$2 in
pre/*)
# the HP disk doesn't like "sync" before when it's still in "sleep" state, so power it on first
sg_start /dev/disk/by-id/ata-MB4000GDUPB_55S6K0ZAF17A
#umount filesystems attached to raid volumes
/etc/init.d/nfs stop
umount /data/md0
umount /data/md1
/etc/init.d/mdraid stop
;;
post/*)
# first, we rescan _EVERYTHING_ related to disk drives on the system: all controllers, all disks, ... everything needs to be rescanned:
for i in /sys/class/scsi_host/host*/scan; do echo "- - -" > $i; done
for i in /sys/block/sd*/device/rescan; do echo 1 > $i; done
#rebuild the raid arrays
/etc/init.d/mdraid start
mount /data/md1
mount /data/m0
/etc/init.d/nfs start
;;
esac
|
but: do not forget to deactivate the mdraid rules in udev! as long as udev starts assembling them on resume, this won't help a lot.
you'll unfortunately need to specify everything in mdadm.conf from now on, but no pain no gain
_________________
The power of Gentoo optimization (not overclocked): [img]https://www.passmark.com/baselines/V10/images/503714802842.png[/img] |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Kernel & Hardware
