[SOLVED] Mageia takes so long to boot

[vecciora]

Everything is fine except it took so long to get through "checking for new hardware" part at boot. The boot process stuck for some time at that part.

It happens every time I boot.

Sorry if the information is lacking because it is the only thing I understand.

[benmc]

Hi

Just some questions.....

Is this Mageia 1 / 2 / 3 ?

How long is a long time, 30sec, 5 minutes, long enough to go and make scones ?

Have you had a previous version of Mageia that was much faster ?

Is this a clean / fresh install that has always been slow ?

Was this a fast booting system that has degraded over time or after an update ?

Have you been removing or changing system components, HDD , video card, etc ?

[vecciora]

Mageia 3.

About 30s maybe.

Yes.

Not fresh install. The boot became slow after installing other Linux OS. I suspect it become like this because the swap partition is being reformatted, but I really am not sure.

No. It's always an issue whenever I install other Linux OS after Mageia.

No.

[vecciora]

I think it is related to swap partition being reformatted because Mageia boot became slow if the swap partition is reformatted.

Here is my /var/log/boot.log: http://pastebin.com/NQmWAdP5

[vecciora]

Just installed another Mageia in another partition and the boot is fine. But, it's not solved anything because the boot always became like this everytime I install another Linux OS.

[junk_no_spam]

Mageia 3.

About 30s maybe.

I would like to get that. Snippet from journalctl | grep "Startup finished"
Startup finished in 3.915s (kernel) + 38.628s (userspace) = 42.543s.

Not fresh install. The boot became slow after installing other Linux OS. I suspect it become like this because the swap partition is being reformatted, but I really am not sure.

You might double check swap UUID in /etc/fstab with results from blkid. Personally, I use labels to avoid that problem.
$ blkid | grep swap
/dev/sdb1: LABEL="swap" <snip,snip>

$ grep swap /etc/fstab
LABEL=swap swap swap defaults 0 0

Commands to get swap partition labeled:
swapoff -a ; mkswap -L swap /dev/sdxx ; swapon LABEL=swap
You have to solve for sdxx
After the above, you can change fstab.

As to troubleshooting start up, I would look through journalctl results after fixing any service failures, if any.
Since I added my user id to the adm group, it is easy to do a
"systemctl --failed" and "journalctl | grep -i fail"
to get a quick view of any problems.

Another method is "systemd-analyze blame" and/or
systemd-analyze plot > plot.svg
firefox plot.svg

[vecciora]

Here is the result of journalctl | grep "Startup finished": http://pastebin.com/YntmVh3E

Looks like there is a jump in Jul 22 from about 30s to 3 minutes.

Here is the systemctl --failed:

Code: Select all

UNIT               LOAD   ACTIVE SUB    JOB DESCRIPTION
mdmonitor.service  loaded failed failed     Software RAID monitoring and management
shorewall6.service loaded failed failed     Shorewall IPv6 firewall

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.
JOB    = Pending job for the unit.

2 loaded units listed. Pass --all to see loaded but inactive units, too.
To show all installed unit files use 'systemctl list-unit-files'.


And journalctl | grep -i fail: http://pastebin.com/4Upx8Rts

As for swap, there is no problem.

[vecciora]

shorewall6.service failed because of "no IP zone defined".

As for mdmonitor.service, it says "mdadm: No mail address or alert command - not monitoring.".

I've just disabled mdmonitor.service to start at boot. I will try to reboot now.

[junk_no_spam]

shorewall6.service failed because of "no IP zone defined".

As for mdmonitor.service, it says "mdadm: No mail address or alert command - not monitoring.".

I've just disabled mdmonitor.service to start at boot. I will try to reboot now.

I disabled both services on my installs.

[vecciora]

Nothing changed. Still stuck at "checking for new hardware" part.

[filip]

Have you tried to rebuild initrd with "dracut -f" as the root user already?

[vecciora]

Have you tried to rebuild initrd with "dracut -f" as the root user already?

No. I'll try it.

[vecciora]

Still stuck at "checking for new hardware" part.

I've done the dracut thing.

[doktor5000]

Please do a fresh boot, and when it has finished, please post there the output as root of

Code: Select all

journalctl -b -a --no-pager

[vecciora]

Oh, no. I've just removed the problematic Mageia.

I'll reproduce this problem later by installing another Linux OS with swap reformatted.

Thanks.

[vecciora]

Please do a fresh boot, and when it has finished, please post there the output as root of

Code: Select all

journalctl -b -a --no-pager

I've reproduce the problem by installing other Linux OS with swap reformatted (if not reformatted it's just fine).

Here is the output: http://pastebin.com/L9gKb106

[vecciora]

Code: Select all

Sep 18 19:20:01 localhost.localdomain systemd[1]: Startup finished in 2s 486ms 264us (kernel) + 3s 868ms 385us (initrd) + 1min 8s 260ms 158us (userspace) = 1min 14s 614ms 807us.
Sep 18 19:28:06 localhost.localdomain systemd[1]: Startup finished in 1s 741ms 701us (kernel) + 7s 481ms 865us (initrd) + 1min 887ms 764us (userspace) = 1min 10s 111ms 330us.
Sep 18 19:32:19 localhost.localdomain systemd[1]: Startup finished in 2s 462ms 30us (kernel) + 7s 596ms 185us (initrd) + 1min 5s 17ms 421us (userspace) = 1min 15s 75ms 636us.
Sep 18 21:53:04 localhost.localdomain systemd[1]: Startup finished in 2s 452ms 913us (kernel) + 7s 639ms 433us (initrd) + 55s 286ms 194us (userspace) = 1min 5s 378ms 540us.
Sep 18 23:05:04 localhost.localdomain systemd[1]: Startup finished in 2s 446ms 192us (kernel) + 7s 539ms 35us (initrd) + 56s 360ms 387us (userspace) = 1min 6s 345ms 614us.
Sep 18 23:25:09 localhost.localdomain systemd[1]: Startup finished in 2s 494ms 362us (kernel) + 3s 999ms 401us (initrd) + 58s 934ms 671us (userspace) = 1min 5s 428ms 434us.
Sep 18 19:00:29 localhost systemd[1]: Startup finished in 3s 112ms 193us (kernel) + 6s 660ms 484us (initrd) + 1min 53s 32ms 19us (userspace) = 2min 2s 804ms 696us.
Sep 19 04:55:13 localhost.localdomain systemd[1]: Startup finished in 2s 449ms 385us (kernel) + 3s 987ms 431us (initrd) + 58s 391ms 514us (userspace) = 1min 4s 828ms 330us.
Sep 19 18:20:04 localhost.localdomain systemd[1]: Startup finished in 2s 443ms 922us (kernel) + 24s 695ms 993us (initrd) + 58s 835ms 501us (userspace) = 1min 25s 975ms 416us.
Sep 20 09:36:26 localhost.localdomain systemd[1]: Startup finished in 1s 757ms 375us (kernel) + 24s 988ms 426us (initrd) + 1min 11s 900ms 985us (userspace) = 1min 38s 646ms 786us.
Sep 20 09:43:16 localhost.localdomain systemd[1]: Startup finished in 2s 481ms 919us (kernel) + 24s 977ms 969us (initrd) + 35s 609ms 595us (userspace) = 1min 3s 69ms 483us.
Sep 20 09:47:08 localhost.localdomain systemd[1]: Startup finished in 2s 469ms 807us (kernel) + 24s 824ms 153us (initrd) + 33s 641ms 490us (userspace) = 1min 935ms 450us.
Sep 21 00:50:32 localhost.localdomain systemd[1]: Startup finished in 2s 455ms 41us (kernel) + 24s 640ms 850us (initrd) + 33s 387ms 471us (userspace) = 1min 483ms 362us.
Sep 21 00:54:12 localhost.localdomain systemd[1]: Startup finished in 2s 464ms 786us (kernel) + 24s 745ms 678us (initrd) + 32s 415ms 852us (userspace) = 59s 626ms 316us.
Sep 21 06:48:53 localhost.localdomain systemd[1]: Startup finished in 2s 464ms 373us (kernel) + 24s 478ms 118us (initrd) + 34s 69ms 776us (userspace) = 1min 1s 12ms 267us.
Sep 21 07:23:19 localhost.localdomain systemd[1]: Startup finished in 2s 492ms 894us (kernel) + 24s 610ms 95us (initrd) + 32s 627ms 238us (userspace) = 59s 730ms 227us.
Sep 21 21:27:12 localhost.localdomain systemd[1]: Startup finished in 2s 472ms 202us (kernel) + 24s 978ms 754us (initrd) + 3min 1s 155ms 222us (userspace) = 3min 28s 606ms 178us.
Sep 21 21:31:25 localhost.localdomain systemd[1]: Startup finished in 2s 486ms 128us (kernel) + 24s 838ms 591us (initrd) + 3min 1s 23ms 546us (userspace) = 3min 28s 348ms 265us.
Sep 21 21:44:32 localhost.localdomain systemd[1]: Startup finished in 2s 463ms 217us (kernel) + 25s 27ms 248us (initrd) + 3min 3s 825ms 833us (userspace) = 3min 31s 316ms 298us.
Sep 21 22:07:18 localhost.localdomain systemd[1]: Startup finished in 2s 471ms 82us (kernel) + 25s 81ms 323us (initrd) + 32s 73ms 888us (userspace) = 59s 626ms 293us.
Sep 21 22:12:53 localhost.localdomain systemd[1]: Startup finished in 2s 478ms 676us (kernel) + 24s 743ms 778us (initrd) + 31s 848ms 11us (userspace) = 59s 70ms 465us.


The boot is back to normal.

I'm confused, but happy.

Just fixed the swap entry in fstab and ran dracut -f.

Maybe it is solved. I don't know why this doesn't work on previous Mageia installation (the one which I mistakenly removed).

[doktor5000]

Probably the problem comes when reformatting any drive, as the UUID will be changed, which is used to identify partitions.
If the UUID changes, there will be a huge delay when trying to mount the partition.

So it's not surprising this happens when you use a common swap device which gets reformatted each time you install another linux.
To circumvent this, use label or device name to reference such partitions in /etc/fstab.

[vecciora]

But, why I must 'dracut -f'?

[doktor5000]

Because it rebuilds the initial ramdisk (initrd) used at boottime.