[SOLVED] Long boot delay: mangled fstab diskdrake mounts

[spiffx]

Hi All,

first off, thanks to everyone for creating and maintaining Mageia: I’ve been a seriously happy user of Mandrake / Mandriva / Mageia for many years (since Mandrake 7.1 in 2000 in fact!)

I’ve recently (finally) upgraded my PC from an AMD Phenom II X4 955 to a Ryzen 5 8500G. The migration itself went without a hitch, which I’m really pleased about, but unfortunately, I’ve managed to break something with the filesystem mounting on boot:

There is now a delay during boot, before the login screen comes up, of around 120 seconds

I believe this is due to me removing one of the SSDs, and unwisely using gparted to update fstab rather than diskdrake. The SSD was an old mageia 8 boot, rather than the newer mageia 9 I use daily now.

Fstab currently looks like this, which I think correctly reflects the current setup:

Code: Select all

# Entry for /dev/sdd1 :
UUID=a33638f8-af29-431f-af27-a48f4eb02f76 / ext4 relatime,acl 1 1
# Entry for /dev/sdd1 :
UUID=28bf9240-7fcc-4ed1-8e67-a721b36fd821 /data ext4 defaults 1 2
# Entry for /dev/sda1 :
UUID=62754ee6-85d0-4ae1-9123-e1bbbe42cfa4 /platforms ext4 defaults 1 2
none /proc proc defaults 0 0
# Entry for /dev/sdc1 :
UUID=a87e6687-0d70-4768-96fd-93684ead24dd /ryzenboot ext4 defaults 1 2
# Entry for /dev/sde1 :
UUID=a3967b5d-f215-4055-8699-6f6c29e17735 /vbox1tbssd ext4 defaults 1 2
# Entry for /dev/sda5 :
#UUID=73be4774-84df-4e85-8ea4-5f000e1af15e swap swap defaults 0 0
# Entry for /dev/sdd5 :
UUID=ddca4939-b312-4f87-8fe9-9000e320f042 swap swap defaults 0 0

diskdrake shows a slightly different arrangement:

Code: Select all

sda1: /platforms UUID: 62754ee6-85d0-4ae1-9123-e1bbbe42cfa4
sdb1: / UUID: a33638f8-af29-431f-af27-a48f4eb02f76
sdb5: swap UUID: ddca4939-b312-4f87-8fe9-9000e320f042
sdc1: /ryzenboot UUID: a87e6687-0d70-4768-96fd-93684ead24dd
sdd1: /data UUID: 28bf9240-7fcc-4ed1-8e67-a721b36fd821
sde1: /vbox1bssd UUID: a3967b5d-f215-4055-8699-6f6c29e17735

I’ve tried setting the following in /etc/systemd/system.conf

Code: Select all

DefaultTimeoutStartSec=10s
DefaultTimeoutStopSec=10s
DefaultDeviceTimeoutSec=10s

but that’s had no discernible effect on the delay


systemctl and systemd-analyse produce the following:

Code: Select all

systemctl list-unit-files -t mount
UNIT FILE                     STATE     PRESET 
-.mount                       generated -       
data.mount                    generated -       
dev-hugepages.mount           static    -       
dev-mqueue.mount              static    -       
platforms.mount               generated -       
proc-fs-nfsd.mount            static    -       
proc-sys-fs-binfmt_misc.mount disabled  disabled
ryzenboot.mount               generated -       
sys-fs-fuse-connections.mount static    -       
sys-kernel-config.mount       static    -       
sys-kernel-debug.mount        static    -       
sys-kernel-tracing.mount      static    -       
tmp.mount                     static    -       
var-lib-machines.mount        static    -       
var-lib-nfs-rpc_pipefs.mount  static    -       
vbox1tbssd.mount              generated -       

16 unit files listed.

Code: Select all

systemctl list-units -t mount --all
  UNIT                                                              LOAD      ACTIVE   SUB     DESCRIPTION                                         
  -.mount                                                           loaded    active   mounted Root Mount
● boot.mount                                                        not-found inactive dead    boot.mount
  data.mount                                                        loaded    active   mounted /data
  dev-hugepages.mount                                               loaded    active   mounted Huge Pages File System
  dev-mqueue.mount                                                  loaded    active   mounted POSIX Message Queue File System
● home.mount                                                        not-found inactive dead    home.mount
  platforms.mount                                                   loaded    active   mounted /platforms
  proc-sys-fs-binfmt_misc.mount                                     loaded    active   mounted Arbitrary Executable File Formats File System
  run-credentials-systemd\x2dsysctl.service.mount                   loaded    active   mounted /run/credentials/systemd-sysctl.service
  run-credentials-systemd\x2dtmpfiles\x2dsetup.service.mount        loaded    active   mounted /run/credentials/systemd-tmpfiles-setup.service
  run-credentials-systemd\x2dtmpfiles\x2dsetup\x2ddev.service.mount loaded    active   mounted /run/credentials/systemd-tmpfiles-setup-dev.service
  run-credentials-systemd\x2dvconsole\x2dsetup.service.mount        loaded    active   mounted /run/credentials/systemd-vconsole-setup.service
  run-user-1000-doc.mount                                           loaded    active   mounted /run/user/1000/doc
  run-user-1000-gvfs.mount                                          loaded    active   mounted /run/user/1000/gvfs
  run-user-1000.mount                                               loaded    active   mounted /run/user/1000
  ryzenboot.mount                                                   loaded    active   mounted /ryzenboot
  sys-fs-fuse-connections.mount                                     loaded    active   mounted FUSE Control File System
  sys-kernel-config.mount                                           loaded    active   mounted Kernel Configuration File System
  sys-kernel-debug.mount                                            loaded    active   mounted Kernel Debug File System
  sys-kernel-tracing.mount                                          loaded    active   mounted Kernel Trace File System
● sysroot.mount                                                     not-found inactive dead    sysroot.mount
  tmp.mount                                                         loaded    active   mounted Temporary Directory /tmp
  var-lib-machines.mount                                            loaded    inactive dead    Virtual Machine and Container Storage (Compatibility)
● var.mount                                                         not-found inactive dead    var.mount
  vbox1tbssd.mount                                                  loaded    active   mounted /vbox1tbssd

LOAD   = Reflects whether the unit definition was properly loaded.
ACTIVE = The high-level unit activation state, i.e. generalization of SUB.
SUB    = The low-level unit activation state, values depend on unit type.
25 loaded units listed.

Code: Select all

systemctl list-units -t mount --all
  UNIT                                                              LOAD      ACTIVE   SUB     DESCRIPTION                                         
  -.mount                                                           loaded    active   mounted Root Mount
● boot.mount                                                        not-found inactive dead    boot.mount
  data.mount                                                        loaded    active   mounted /data
  dev-hugepages.mount                                               loaded    active   mounted Huge Pages File System
  dev-mqueue.mount                                                  loaded    active   mounted POSIX Message Queue File System
● home.mount                                                        not-found inactive dead    home.mount
  platforms.mount                                                   loaded    active   mounted /platforms
  proc-sys-fs-binfmt_misc.mount                                     loaded    active   mounted Arbitrary Executable File Formats File System
  run-credentials-systemd\x2dsysctl.service.mount                   loaded    active   mounted /run/credentials/systemd-sysctl.service
  run-credentials-systemd\x2dtmpfiles\x2dsetup.service.mount        loaded    active   mounted /run/credentials/systemd-tmpfiles-setup.service
  run-credentials-systemd\x2dtmpfiles\x2dsetup\x2ddev.service.mount loaded    active   mounted /run/credentials/systemd-tmpfiles-setup-dev.service
  run-credentials-systemd\x2dvconsole\x2dsetup.service.mount        loaded    active   mounted /run/credentials/systemd-vconsole-setup.service
  run-user-1000-doc.mount                                           loaded    active   mounted /run/user/1000/doc
  run-user-1000-gvfs.mount                                          loaded    active   mounted /run/user/1000/gvfs
  run-user-1000.mount                                               loaded    active   mounted /run/user/1000
  ryzenboot.mount                                                   loaded    active   mounted /ryzenboot
  sys-fs-fuse-connections.mount                                     loaded    active   mounted FUSE Control File System
  sys-kernel-config.mount                                           loaded    active   mounted Kernel Configuration File System
  sys-kernel-debug.mount                                            loaded    active   mounted Kernel Debug File System
  sys-kernel-tracing.mount                                          loaded    active   mounted Kernel Trace File System
● sysroot.mount                                                     not-found inactive dead    sysroot.mount
  tmp.mount                                                         loaded    active   mounted Temporary Directory /tmp
  var-lib-machines.mount                                            loaded    inactive dead    Virtual Machine and Container Storage (Compatibility)
● var.mount                                                         not-found inactive dead    var.mount
  vbox1tbssd.mount                                                  loaded    active   mounted /vbox1tbssd

LOAD   = Reflects whether the unit definition was properly loaded.
ACTIVE = The high-level unit activation state, i.e. generalization of SUB.
SUB    = The low-level unit activation state, values depend on unit type.
25 loaded units listed.

Code: Select all

systemctl list-units -t mount --all
  UNIT                                                              LOAD      ACTIVE   SUB     DESCRIPTION                                         
  -.mount                                                           loaded    active   mounted Root Mount
● boot.mount                                                        not-found inactive dead    boot.mount
  data.mount                                                        loaded    active   mounted /data
  dev-hugepages.mount                                               loaded    active   mounted Huge Pages File System
  dev-mqueue.mount                                                  loaded    active   mounted POSIX Message Queue File System
● home.mount                                                        not-found inactive dead    home.mount
  platforms.mount                                                   loaded    active   mounted /platforms
  proc-sys-fs-binfmt_misc.mount                                     loaded    active   mounted Arbitrary Executable File Formats File System
  run-credentials-systemd\x2dsysctl.service.mount                   loaded    active   mounted /run/credentials/systemd-sysctl.service
  run-credentials-systemd\x2dtmpfiles\x2dsetup.service.mount        loaded    active   mounted /run/credentials/systemd-tmpfiles-setup.service
  run-credentials-systemd\x2dtmpfiles\x2dsetup\x2ddev.service.mount loaded    active   mounted /run/credentials/systemd-tmpfiles-setup-dev.service
  run-credentials-systemd\x2dvconsole\x2dsetup.service.mount        loaded    active   mounted /run/credentials/systemd-vconsole-setup.service
  run-user-1000-doc.mount                                           loaded    active   mounted /run/user/1000/doc
  run-user-1000-gvfs.mount                                          loaded    active   mounted /run/user/1000/gvfs
  run-user-1000.mount                                               loaded    active   mounted /run/user/1000
  ryzenboot.mount                                                   loaded    active   mounted /ryzenboot
  sys-fs-fuse-connections.mount                                     loaded    active   mounted FUSE Control File System
  sys-kernel-config.mount                                           loaded    active   mounted Kernel Configuration File System
  sys-kernel-debug.mount                                            loaded    active   mounted Kernel Debug File System
  sys-kernel-tracing.mount                                          loaded    active   mounted Kernel Trace File System
● sysroot.mount                                                     not-found inactive dead    sysroot.mount
  tmp.mount                                                         loaded    active   mounted Temporary Directory /tmp
  var-lib-machines.mount                                            loaded    inactive dead    Virtual Machine and Container Storage (Compatibility)
● var.mount                                                         not-found inactive dead    var.mount
  vbox1tbssd.mount                                                  loaded    active   mounted /vbox1tbssd

LOAD   = Reflects whether the unit definition was properly loaded.
ACTIVE = The high-level unit activation state, i.e. generalization of SUB.
SUB    = The low-level unit activation state, values depend on unit type.
25 loaded units listed.

Code: Select all

[root@localhost systemd]# systemd-analyze
Startup finished in 2min 16.834s (kernel) + 12.049s (userspace) = 2min 28.883s
graphical.target reached after 11.724s in userspace.

Code: Select all

[root@localhost systemd]# systemd-analyze critical-chain
The time when unit became active or started is printed after the "@" character.
The time the unit took to start is printed after the "+" character.

graphical.target @11.724s
└─sddm.service @11.724s
  └─plymouth-quit.service @11.699s +23ms
    └─systemd-user-sessions.service @11.644s +2ms
      └─network.target @11.594s
        └─network-up.service @6.045s +5.549s
          └─basic.target @5.997s
            └─mandriva-everytime.service @5.236s +760ms
              └─dkms-autorebuild.service @4.915s +287ms
                └─local-fs.target @4.915s
                  └─run-credentials-systemd\x2dtmpfiles\x2dsetup.service.mount @4.927s
                    └─local-fs-pre.target @4.463s
                      └─systemd-tmpfiles-setup-dev.service @390ms +138ms
                        └─kmod-static-nodes.service @322ms +10ms
                          └─systemd-journald.socket @318ms
                            └─-.mount @281ms
                              └─-.slice @281ms

boot.mount, home.mount, sysroot.mount and var.mount look suspicious I think? They may well be left over from the old mageia8 boot drive I removed. I have no idea how to get fstab, Diskdrake and systemd to talk to each other to resolve the discrepancy any help / pointers would be much appreciated

Spiff

Please use code tags for output. ~Germ

[doktor5000]

Please provide the actual output as root of

Code: Select all

lsblk -f

Also the UUID of your swap partition has changed, comparing your fstab and diskdrake output.
This needs to be updated in the initrd /dracut droplet as well as otherwise this can also lead to long timeouts or even boot failing.

I'd probably comment out all partitions that are not your Mageia installation, and then go on from there.

[spiffx]

Hi,

thanks for your reply.

below is the result of the lsblk command:

Code: Select all

lsblk -f
NAME   FSTYPE FSVER LABEL UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
sda                                                                           
`-sda1 ext4   1.0         62754ee6-85d0-4ae1-9123-e1bbbe42cfa4   91.1G    75% /platforms
sdb                                                                           
|-sdb1 ext4   1.0         a33638f8-af29-431f-af27-a48f4eb02f76   43.4G    42% /
|-sdb2                                                                       
`-sdb5 swap   1           ddca4939-b312-4f87-8fe9-9000e320f042                [SWAP]
sdc                                                                           
`-sdc1 ext4   1.0         a87e6687-0d70-4768-96fd-93684ead24dd   78.7G    66% /ryzenboot
sdd                                                                           
`-sdd1 ext4   1.0         28bf9240-7fcc-4ed1-8e67-a721b36fd821  836.5G    49% /data
sde                                                                           
`-sde1 ext4   1.0         a3967b5d-f215-4055-8699-6f6c29e17735  142.3G    80% /vbox1tbssd
sr0                         



fstab is a bit confusing i think, but my understanding that the following was already commented out:

Code: Select all

# Entry for /dev/sda5 :
#UUID=73be4774-84df-4e85-8ea4-5f000e1af15e swap swap defaults 0 0



[doktor5000]

fstab is a bit confusing i think, but my understanding that the following was already commented out:

Code: Select all

# Entry for /dev/sda5 :
#UUID=73be4774-84df-4e85-8ea4-5f000e1af15e swap swap defaults 0 0



fstab alone does not matter. swap UUID is by default also contained in /etc/dracut.conf.d/51-mageia-resume.conf where is also needs to be updated manually, afterwards you need to run

Code: Select all

dracut -f --regenerate-all

Apart from that it would be helpful if you could run

Code: Select all

journalctl -ab > /tmp/journal.log

as root after a fresh boot, and then later attach /tmp/journal.log here as you didn't provide any logs excerpts related to the timeouts.

[spiffx]

(apologies for the delay in responding)

i've commented out the line in /etc/dracut.conf.d/51-mageia-resume.conf

Code: Select all

#add_device+=" UUID=73be4774-84df-4e85-8ea4-5f000e1af15e "


and then run

Code: Select all

dracut -f --regenerate-all


unfortunately, the long boot delay persists.

Attached is the result of

Code: Select all

journalctl -ab > /tmp/journal.log


[doktor5000]

There is now a delay during boot, before the login screen comes up, of around 120 seconds

unfortunately, the long boot delay persists.

Your boot takes 30 seconds according to the logs, not sure when that 120 second delay is supposed to happen. Log starts at 17:15:22 and ends at 17:15:57
Although:

Oct 31 17:15:34 localhost systemd[1]: Startup finished in 2min 16.977s (kernel) + 12.069s (userspace) = 2min 29.046s.

Is the 120 second delay after the graphical bootloader ends where you select Mageia to boot ?

Also do you use legacy boot, not UEFI?

[spiffx]

Hi,

below is a link to a video which shows the boot sequence, up to graphical plasma login (apologies for the low / poor quality)

https://u.pcloud.link/publink/show?code=XZcMSV5Z18v5UJAhSmkIDW9pwh3sDFAznd67

there is a pause from approximately T=0m15s to T=2m30s in the boot sequence video where there is no disk activity at all, just the mageia cauldron animation. i.e. a delay of around 2m15s where i get the impression nothing is happening

i'm not using UEFI (I only use windows in VMs)

many thanks

Spiff

[doktor5000]

That's what I mean with "120 second delay after the graphical bootloader ends" - this is actually before the kernel loads, so you can't really debug that.
I've seen this in the past on one pretty old box IIRC, this was fixed by a BIOS update and/or by a full reinstallation. Reinstalling bootloader did not help.
You could try booting with systemd debug options enabled, and later posting a full log from such a boot here.

Code: Select all

systemd.log_level=debug systemd.log_target=kmsg log_buf_len=1M printk.devkmsg=on

Could please post your grub configuration and check if rootdelay or rootwait is set anywhere ?

[Germ]

I fixed the same problem on a old Acer laptop by doing a BIOS update.

[spiffx]

hurrah!

mashing ctrl-alt-f1..f12 during the stalled animation eventually brought up

Code: Select all

dracut warning: cancelling resume operation. device not found

looking at the grub2 config, one of the appended kernel parameters was

Code: Select all

resume=UUID=73be4774-84df-4e85-8ea4-5f000e1af15e

which is the ID of the originally removed boot drive swap partition. i've replaced this with the ID of the correct swap partition: boot time is now down to a far more reasonable 25 seconds in total.

all good!

thanks so much for all your help, it's very much appreciated

[Germ]

Glad you got it sorted.

Could you mark your topic as SOLVED? Thanks.

[doktor5000]

looking at the grub2 config, one of the appended kernel parameters was

Code: Select all

resume=UUID=73be4774-84df-4e85-8ea4-5f000e1af15e

which is the ID of the originally removed boot drive swap partition. i've replaced this with the ID of the correct swap partition: boot time is now down to a far more reasonable 25 seconds in total.

That is exactly what I told you beforehand as per your fstab entries the UUID of your swap partition has changed.

Please don't forget to mark the thread as solved, by editing the subject of the first post and prefix it with [SOLVED], thanks in advance.