| View previous topic :: View next topic |
| Author |
Message |
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
 Posted: Wed Apr 09, 2014 7:43 pm Post subject: dmraid, no automount during boot with newer kernel |
 |
|
Hi there.
I have a raid 1 array of two (USB) drives set up and running for quite some time. When I use a kernel newer than gentoo-3.7.10 the raid array does not automount during boot.
Kernel 3.7.10:
I have /dev/md127. In /etc/fstab a line assigns /dev/md127 to /mnt/raid. The raid automounts during boot.
Kernel 3.10.x and 3.12.12:
I have /dev/md1. I changed /etc/fstab to automount /dev/md1 to /mnt/raid. I can manually mount /dev/md1 and the files are showing in /mnt/raid. But the array is not mounted during boot.
I already tried to manually add an entry for /dev/md1 and it's UUID to /etc/conf.d/mdadm with no success.
What is going wrong here? What more info do you need? Any help appreciated. |
|
| Back to top |
|
 |
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
| Posted: Fri Apr 11, 2014 10:48 pm Post subject: |
|
|
Nobody?
I have just made a backup of the files on the raid. I'm going to rebuild the raid from scratch and hope that it automounts again.  |
|
| Back to top |
|
|
Goverp
Veteran
Joined: 07 Mar 2007
Posts: 1999
|
| Posted: Sat Apr 12, 2014 11:44 am Post subject: |
|
|
FWIW, kernel automounting has been deprecated for some time. You're supposed to use an initramfs. AFAIK automounting is still expected to work, and people (not me) still use it successfully.
_________________
Greybeard |
|
| Back to top |
|
|
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
| Posted: Sat Apr 12, 2014 1:34 pm Post subject: |
|
|
| Goverp wrote: | | FWIW, kernel automounting has been deprecated for some time. You're supposed to use an initramfs. AFAIK automounting is still expected to work, and people (not me) still use it successfully. |
Thank you.
The raid does not contain the root file system or anything that is needed during boot. There's only user data on it. The device is supposed to mount via an fstab entry. I'd prefer not to use an initramfs. Maybe I'll just write a little script to mount it later in the boot process... |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54220
Location: 56N 3W
|
| Posted: Sat Apr 12, 2014 2:44 pm Post subject: |
|
|
hoacker,
Is mdraid in the default runlevel?
Is your mdraid.conf correct ?
For the kernel to start your raid during boot, you need all the USB support built in.
As your raid is only user data, an initrd is not required.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
| Posted: Sat Apr 12, 2014 5:46 pm Post subject: |
|
|
Hi NeddySeagoon.
Mdadm is in default runlevel. I had no DEVICE or ARRAY entry in mdraid.conf until yesterday, automount worked anyway with older kernel.
Two identical USB drives are connected to the machine. Mdadm says the follwing:
| Code: | # mdadm --detail --scan
ARRAY /dev/md/1 metadata=1.2 name=steward:1 UUID=716584c8:65b663cb:efc4803d:0c6517f7
|
With kernel 3.7.10 the device /dev/md127 is detected and mounted (xfs). There is an entry in /etc/fstab for automount of /dev/md127 to /mnt/raid.
Here's snippet from /dev/log/messages with 3.7.10, the two TekStor are the raid devices:
| Code: | Apr 12 19:05:25 localhost kernel: md: raid0 personality registered for level 0
Apr 12 19:05:25 localhost kernel: md: raid1 personality registered for level 1
Apr 12 19:05:25 localhost kernel: md: raid10 personality registered for level 10
Apr 12 19:05:25 localhost kernel: md: raid6 personality registered for level 6
Apr 12 19:05:25 localhost kernel: md: raid5 personality registered for level 5
Apr 12 19:05:25 localhost kernel: md: raid4 personality registered for level 4
Apr 12 19:05:25 localhost kernel: device-mapper: ioctl: 4.23.0-ioctl (2012-07-25) initialised: dm-devel@redhat.com
[..]
Apr 12 19:05:25 localhost kernel: md: Waiting for all devices to be available before autodetect
Apr 12 19:05:25 localhost kernel: md: If you don't use raid, use raid=noautodetect
Apr 12 19:05:25 localhost kernel: md: Autodetecting RAID arrays.
Apr 12 19:05:25 localhost kernel: md: Scanned 0 and added 0 devices.
Apr 12 19:05:25 localhost kernel: md: autorun ...
Apr 12 19:05:25 localhost kernel: md: ... autorun DONE.
[..]
Apr 12 19:05:25 localhost kernel: usb 2-3: new high-speed USB device number 4 using ehci_hcd
Apr 12 19:05:25 localhost kernel: scsi8 : usb-storage 2-3:1.0
Apr 12 19:05:25 localhost kernel: scsi 3:0:0:0: Direct-Access TrekStor DS maxi Xpress 0001 PQ: 0 ANSI: 6
Apr 12 19:05:25 localhost kernel: sd 3:0:0:0: Attached scsi generic sg2 type 0
Apr 12 19:05:25 localhost kernel: scsi 4:0:0:0: Direct-Access TrekStor DS maxi Xpress 0001 PQ: 0 ANSI: 6
Apr 12 19:05:25 localhost kernel: sd 4:0:0:0: Attached scsi generic sg3 type 0
Apr 12 19:05:25 localhost kernel: md: bind<sdb1>
Apr 12 19:05:25 localhost kernel: sdc: sdc1
Apr 12 19:05:25 localhost kernel: sd 4:0:0:0: [sdc] No Caching mode page present
Apr 12 19:05:25 localhost kernel: sd 4:0:0:0: [sdc] Assuming drive cache: write through
Apr 12 19:05:25 localhost kernel: sd 4:0:0:0: [sdc] Attached SCSI disk
Apr 12 19:05:25 localhost kernel: md: bind<sdc1>
Apr 12 19:05:25 localhost kernel: md/raid1:md127: active with 2 out of 2 mirrors
Apr 12 19:05:25 localhost kernel: md127: detected capacity change from 0 to 2000395101184
Apr 12 19:05:25 localhost kernel: md127: unknown partition table
Apr 12 19:05:25 localhost kernel: XFS (md127): Mounting Filesystem
Apr 12 19:05:25 localhost kernel: XFS (md127): Ending clean mount |
With kernel 3.10 or 3.12 the device /dev/md127 does not show, it's /dev/md1 instead. I changed fstab accordingly. After boot there's no automounted raid array, but manual mount works fine.
I've added the following lines to mdraid.conf:
| Code: | DEVICE /dev/sd[bcdefghi]1
ARRAY /dev/md1 uuid=716584c8:65b663cb:efc4803d:0c6517f7
|
The snippet from /var/log/messages with kernel 3.12:
| Code: | Apr 12 18:47:08 localhost kernel: md: raid0 personality registered for level 0
Apr 12 18:47:08 localhost kernel: md: raid1 personality registered for level 1
Apr 12 18:47:08 localhost kernel: md: raid10 personality registered for level 10
Apr 12 18:47:08 localhost kernel: md: raid6 personality registered for level 6
Apr 12 18:47:08 localhost kernel: md: raid5 personality registered for level 5
Apr 12 18:47:08 localhost kernel: md: raid4 personality registered for level 4
Apr 12 18:47:08 localhost kernel: device-mapper: ioctl: 4.26.0-ioctl (2013-08-15) initialised: dm-devel@redhat.com
Apr 12 18:47:08 localhost kernel: device-mapper: raid: Loading target version 1.5.2
Apr 12 18:47:08 localhost kernel: usb 1-2: new high-speed USB device number 2 using ehci-pci
[..]
Apr 12 18:47:08 localhost kernel: usb 1-3: new high-speed USB device number 3 using ehci-pci
[..]
Apr 12 18:47:08 localhost kernel: usb 1-4: new high-speed USB device number 4 using ehci-pci
[..]
Apr 12 18:47:08 localhost kernel: usb-storage 2-1:1.0: USB Mass Storage device detected
Apr 12 18:47:08 localhost kernel: scsi6 : usb-storage 2-1:1.0
Apr 12 18:47:08 localhost kernel: usb 2-2: new high-speed USB device number 3 using ehci-pci
Apr 12 18:47:08 localhost kernel: ata3: SATA link down (SStatus 0 SControl 300)
Apr 12 18:47:08 localhost kernel: md: Waiting for all devices to be available before autodetect
Apr 12 18:47:08 localhost kernel: md: If you don't use raid, use raid=noautodetect
Apr 12 18:47:08 localhost kernel: md: Autodetecting RAID arrays.
Apr 12 18:47:08 localhost kernel: md: Scanned 0 and added 0 devices.
Apr 12 18:47:08 localhost kernel: md: autorun ...
Apr 12 18:47:08 localhost kernel: md: ... autorun DONE.
Apr 12 18:47:08 localhost kernel: usb-storage 2-2:1.0: USB Mass Storage device detected
Apr 12 18:47:08 localhost kernel: scsi7 : usb-storage 2-2:1.0
Apr 12 18:47:08 localhost kernel: usb 2-3: new high-speed USB device number 4 using ehci-pci
Apr 12 18:47:08 localhost kernel: usb-storage 2-3:1.0: USB Mass Storage device detected
Apr 12 18:47:08 localhost kernel: scsi8 : usb-storage 2-3:1.0
Apr 12 18:47:08 localhost kernel: scsi 3:0:0:0: Direct-Access TrekStor DS maxi Xpress 0001 PQ: 0 ANSI: 6
Apr 12 18:47:08 localhost kernel: sd 3:0:0:0: Attached scsi generic sg2 type 0
Apr 12 18:47:08 localhost kernel: scsi 4:0:0:0: Direct-Access TrekStor DS maxi Xpress 0001 PQ: 0 ANSI: 6
Apr 12 18:47:08 localhost kernel: sd 4:0:0:0: Attached scsi generic sg3 type 0
Apr 12 18:47:08 localhost kernel: scsi 5:0:0:0: Direct-Access Generic- Multi-Card 1.00 PQ: 0 ANSI: 0 CCS
Apr 12 18:47:08 localhost kernel: sd 5:0:0:0: Attached scsi generic sg4 type 0
Apr 12 18:47:08 localhost kernel: sd 5:0:0:0: [sdd] Attached SCSI removable disk
Apr 12 18:47:08 localhost kernel: Switched to clocksource tsc
Apr 12 18:47:08 localhost kernel: usb 7-2: new low-speed USB device number 2 using uhci_hcd
Apr 12 18:47:08 localhost kernel: scsi 6:0:0:0: Direct-Access SAMSUNG HD501LJ PQ: 0 ANSI: 2 CCS
Apr 12 18:47:08 localhost kernel: sd 6:0:0:0: Attached scsi generic sg5 type 0 |
|
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54220
Location: 56N 3W
|
| Posted: Sat Apr 12, 2014 6:37 pm Post subject: |
|
|
hoacker,
With the kernel will not auto assemble your raid. That has only works with metadata=0.9.
Indeed , | kernel-3.7.10: | Apr 12 19:05:25 localhost kernel: md: If you don't use raid, use raid=noautodetect
Apr 12 19:05:25 localhost kernel: md: Autodetecting RAID arrays.
Apr 12 19:05:25 localhost kernel: md: Scanned 0 and added 0 devices.
Apr 12 19:05:25 localhost kernel: md: autorun ...
Apr 12 19:05:25 localhost kernel: md: ... autorun DONE. |
shows your USB raid not being assembled by the kernel.
That snippet is the same in 3.12.
It follows that its not a kernel raid auto assembly issue.
When you use 3.7.10 something prods mdadm to assemble your raid.
That manual mounting works with the later kernel suggests that the raid is assembled, even though there is nothing in dmesg.
Reboot a problem kerenel. Do not do the manual mount. Post the content of /proc/mdstat
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
| Posted: Sat Apr 12, 2014 6:46 pm Post subject: |
|
|
| NeddySeagoon wrote: | | Do not do the manual mount. Post the content of /proc/mdstat |
Here we go:
| Quote: | # cat /proc/mdstat
Personalities : [raid0] [raid1] [raid10] [raid6] [raid5] [raid4]
md1 : active raid1 sdc1[1] sdb1[2]
1953510841 blocks super 1.2 [2/2] [UU]
unused devices: <none> |
|
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54220
Location: 56N 3W
|
| Posted: Sat Apr 12, 2014 9:13 pm Post subject: |
|
|
hoacker,
That shows /dev/md1 correctly assembled.
What does produce?
I'm thinking of moving mdadm to the boot runlevel, or making everything needed for usb-storage as built into the kernel, so usb-storage works before you can load modules.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
| Posted: Sat Apr 12, 2014 9:27 pm Post subject: |
|
|
Moving mdadm to boot runlevel makes no difference, tried both. I think I have compile relevant USB options into the kernel. If uploaded .config to pastebin if you would like to have a look. http://pastebin.com/7KTdRbws
| Code: | # rc-update
acpid | default
apache2 | default
bootmisc | boot
clamsmtpd | default
consolefont | boot
cpufrequtils | default
cupsd | default
devfs | sysinit
dhcpd | default
dmesg | sysinit
dovecot | default
fsck | boot
hostname | boot
hwclock | boot
irqbalance | default
keymaps | boot
killprocs | shutdown
local | default nonetwork
localmount | boot
loopback | boot
mdadm | boot
modules | boot
mount-ro | shutdown
mtab | boot
mysql | default
named | default
net.lo | boot
netmount | default
nfs | default
ntpd | default
postfix | default
postgresql-9.1 | default
procfs | boot
root | boot
rsyncd | default
samba | default
saslauthd | default
savecache | shutdown
spamd | default
sshd | default
svnserve | default
swap | boot
swapfiles | boot
sysctl | boot
sysfs | sysinit
syslog-ng | default
termencoding | boot
timekeeper | default
tmpfiles.dev | sysinit
tmpfiles.setup | boot
udev | sysinit
udev-mount | sysinit
urandom | boot
usettings | default
vixie-cron | default |
|
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54220
Location: 56N 3W
|
| Posted: Sat Apr 12, 2014 9:39 pm Post subject: |
|
|
hoacker,
That looks good.
Put mdadm back into the default runlevel.
Post your /etc/fstab. You could try the bg option on your /devmd1. I'm not sure if it would be valid, or if its only good for nfs.
See man mount
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
| Posted: Sat Apr 12, 2014 9:52 pm Post subject: |
|
|
| Code: | | XFS (md1): unknown mount option [bg] |
I can't find a similar option for xfs.
The fstab entry:
| Code: | | /dev/md1 /mnt/raid xfs auto 0 0 |
|
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54220
Location: 56N 3W
|
| Posted: Sat Apr 12, 2014 10:02 pm Post subject: |
|
|
hoacker,
I think what's happening is that your raid in not assembled until after localmount has run.
As its not there, it does not get mounted.
This is a bit of a hack ...
See /etc/local.d/README
Make a file called say mountraid.start with one line
| Code: | mount /dev/md1 /mnt/raid
chmod +x mountraid.start |
I don't think it needs #!/bin/bash at the top.
You could also edit /etc/init.d/localmount so it runs after mdadm
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
| Posted: Sat Apr 12, 2014 10:08 pm Post subject: |
|
|
| NeddySeagoon wrote: | This is a bit of a hack ...
See /etc/local.d/README |
Yes, was thinking about that, too. I hoped that we could figure out the problem and solve it in a different way. local.d will solve my problem, I consider this case closed.
Thank you very much for your support, NeddySeagoon. |
|
| Back to top |
|
|
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
| Posted: Sat Apr 12, 2014 10:22 pm Post subject: |
|
|
| NeddySeagoon wrote: | | You could also edit /etc/init.d/localmount so it runs after mdadm |
I chose that solution. Works!
| Code: | - after lvm modules
+ after lvm modules mdadm |
|
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54220
Location: 56N 3W
|
| Posted: Sat Apr 12, 2014 11:34 pm Post subject: |
|
|
hoacker,
Take care. One day etc-update will revert that change, if you let it.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
hoacker
Guru
Joined: 04 Aug 2007
Posts: 505
Location: Bürstadt, Germany
|
| Posted: Sat Apr 12, 2014 11:40 pm Post subject: |
|
|
| NeddySeagoon wrote: | | Take care. One day etc-update will revert that change, if you let it. |
I know, dispatch-conf is my friend...  |
|
| Back to top |
|
|
|
Kernel & Hardware
