Slow boot and old kernels

I have a problem with my OpenSUSE 13.1 install. It’s not a big problem but somewhat irritating. Booting goes fine until “Loading initial ramdisk” shows up and then it’s a 20s pause doing nothing. It then starts normally. What can I do to speed it up?

My second problem is maybe related to the first: I can not uninstall the old kernel! I have two kernels installed (13.11.10 and 13.11.6) and if I try to remove 13.11.6 in Yast it complains about dependencies and refuses to delete it. if I keep booth versions installed everything is alright. Well, except for the slow boot.

:\

You could uninstall plymouth to get a smaller initrd. It should then load faster.
But you would loose the boot splash in that case of course.

Maybe a BIOS update might help as well. Or try a different Boot loader (YaST->System->Boot Loader).

My second problem is maybe related to the first: I can not uninstall the old kernel! I have two kernels installed (13.11.10 and 13.11.6) and if I try to remove 13.11.6 in Yast it complains about dependencies and refuses to delete it. if I keep booth versions installed everything is alright. Well, except for the slow boot.

The number of kernels that are installed should NOT have any influence at the boot time, since only one kernel/initrd is loaded anyway.

Regarding the dependency conflict:
I suppose you have some kernel modules for the 3.11.6 kernel installed. You have to remove those as well when you uninstall the kernel.
What exact conflicts do you get?

On 2014-05-17 17:26, fhansson wrote:
>
> I have a problem with my OpenSUSE 13.1 install. It’s not a big problem
> but somewhat irritating. Booting goes fine until “Loading initial
> ramdisk” shows up and then it’s a 20s pause doing nothing. It then
> starts normally. What can I do to speed it up?

Is that a text, verbose, mode boot, where no text lines are printed for
a while, or is it a boot with a graphical display with somekind of
progress bar that appears to do nothing?

>
> My second problem is maybe related to the first: I can not uninstall the
> old kernel! I have two kernels installed (13.11.10 and 13.11.6) and if I
> try to remove 13.11.6 in Yast it complains about dependencies and
> refuses to delete it. if I keep booth versions installed everything is
> alright. Well, except for the slow boot.

Should have no effect on boot speed.


Cheers / Saludos,

Carlos E. R.

(from 13.1 x86_64 “Bottle” (Minas Tirith))

OK, I tried GRUB but the result is almost the same. It stops at “initrd vmlinuz… something” for a while and then continues. I don’t know if that is normal since my computer is not very fast, it’s an Acer Aspire One Netbook with Intel SSD. I tried Windows 8 and it started in 30 sec to desktop, OpenSUSE is more like 1 min 30 sec which is a bit too long.

I suppose you have some kernel modules for the 3.11.6 kernel installed. You have to remove those as well when you uninstall the kernel.
What exact conflicts do you get?

If I mark it for delete in Yast it complains about a dependency with the “microcode_ctl” package. I removed the package and it seems to work without it but I can not uninstall the old kernel no matter what I do.

How do I remove the modules? What is the name of the package?

Thanks!

Well, the boot loader has to load the kernel and the initrd. At this point there is no hard disk driver loaded yet (there’s not even the kernel loaded of course), so it has to use generic routines, possibly calling the BIOS.
That’s slower of course, and might depend on the specific BIOS.

But there’s no way around that. A different boot loader might do things differently, so maybe try LILO or ELILO.

Other than that, the only way to speed this up is to make the initrd smaller, by uninstalling plymouth f.e.

Windows on the other hand does all of this completely different. That might explain why it’s faster in your case.

If I mark it for delete in Yast it complains about a dependency with the “microcode_ctl” package. I removed the package and it seems to work without it but I can not uninstall the old kernel no matter what I do.

What do you mean with “I can not uninstall the old kernel no matter what I do”? What happens when you try to uninstall it?

Regarding the microcode_ctl package: this does not exist any more in openSUSE 13.1. So it must be a left-over from an older version.
And no, you don’t need it.

How do I remove the modules? What is the name of the package?

There is no single package. The name depends on the module of course.
F.e. there’s the package “virtualbox-host-kmp-xxx” for the virtualbox kernel modules.

YaST should tell you what requires the kernel package.

But again, the second kernel has absolutely no impact on your boot speed.

On 18/05/2014 18:26, wolfi323 wrote:

> That’s slower of course, and might depend on the specific BIOS.

Not that noticeably slow… at most, 50 MB have to be read. Even if the
BIOS is 10 times slower, it would be a question of seconds.

In theory, at least…

> But there’s no way around that. A different boot loader might do things
> differently, so maybe try LILO or ELILO.
>
> Other than that, the only way to speed this up is to make the initrd
> smaller, by uninstalling plymouth f.e.

What if there is a problem in the disk itself?

I would try an fsck of the partition used for booting, preferably run
from a Linux live cd, and then I would run the short and long SMART
tests, with “smartctl” (the --help option and the manual page says how).


Cheers,
Carlos E. R.
(W7 - minas-morgul)

Well, IIRC at least 12.3 beta took quite long to load the kernel/initrd in VirtualBox.
30s to 1 minute I think. And I think this had to do with the way grub(2) loads the data.

A similar thing might happen in the OP’s case.

A bad hard disk could be the reason as well of course. But I guess he would have bigger problems then…

It seems that Intel BIOS, CPU and SSD does not co-op very well with GRUB then. I can live with that.

What do you mean with “I can not uninstall the old kernel no matter what I do”? What happens when you try to uninstall it?

If I mark the older version of kernel-default package for Delete in Yast it will install the old kernel as kernel-base package instead. If I mark that for Delete it will install the old version of kernel-default again and I am back at square 1. No matter what I choose to do it will install it somehow anyway. :\

Regarding the microcode_ctl package: this does not exist any more in openSUSE 13.1. So it must be a left-over from an older version.
And no, you don’t need it.

True, I did an upgrade from 12.3 to 13.1 a couple of months ago.

There is no single package. The name depends on the module of course.
F.e. there’s the package “virtualbox-host-kmp-xxx” for the virtualbox kernel modules.

YaST should tell you what requires the kernel package.

But again, the second kernel has absolutely no impact on your boot speed.

OK, but I think I can live with two kernels as long as it does not impact anything else.

Thanks!

Regarding the disk it is an Intel X25-M G2 and the CPU is Intel Atom N270 with Intel 945GSE chipset. OpenSUSE itself is upgraded from 12.3 and installed on an extended partition. It’s also resized and moved with GParted at one time. i don’t know if that made it slower but I can not remember it was this slow with 12.3 version. Maybe a fresh install of 13.1 would speed things up?

Ah, ok.
Something requires the kernel package, so YaST wants to install the kernel-base package instead when you mark it for removal to satisfy the dependencies.
Right-click on the kernel-default-base package and select “Taboo - Never install”. Then mark the older kernel-default package for removal and you should get a conflict resolution dialog that tells you what needs that kernel and should allow you to uninstall that as well.

Why are you using kernel-default, btw?
Normally kernel-desktop is installed by default, which is better fine-tuned for desktop (as in desktop environment, not as in desktop vs. laptop/netbook) usage.

OK, but I think I can live with two kernels as long as it does not impact anything else.

Well, at least two kernels are kept by default since 12.3.
The advantage is that you can still boot the older kernel if you experience some problem after a kernel update.

On 2014-05-19 07:56, fhansson wrote:
>
> robin_listas;2644007 Wrote:
>> On 18/05/2014 18:26, wolfi323 wrote:
>>
>> What if there is a problem in the disk itself?
>>
>> I would try an fsck of the partition used for booting, preferably run
>> from a Linux live cd, and then I would run the short and long SMART
>> tests, with “smartctl” (the --help option and the manual page says how).
>>
>
> Regarding the disk it is an Intel X25-M G2 and the CPU is Intel Atom
> N270 with Intel 945GSE chipset. OpenSUSE itself is upgraded from 12.3
> and installed on an extended partition. It’s also resized and moved with
> GParted at one time. i don’t know if that made it slower but I can not
> remember it was this slow with 12.3 version. Maybe a fresh install of
> 13.1 would speed things up?

Run the checks I told you first.

Is your disk perchance one of those with 4 KiB internal sectors, with
512 byte emulation? The output of “fdisk -l” will tell you if you don’t
know. Or is it 4k/4k?


Cheers / Saludos,

Carlos E. R.

(from 13.1 x86_64 “Bottle” (Minas Tirith))

This is the output from smartctl:

=== START OF INFORMATION SECTION ===
Model Family: Intel X18-M/X25-M/X25-V G2 SSDs
Device Model: INTEL SSDSA2M080G2GC
Serial Number: CVPO944500B6080BGN
LU WWN Device Id: 5 001517 9590d59c3
Firmware Version: 2CV102M3
User Capacity: 80,026,361,856 bytes [80.0 GB]
Sector Size: 512 bytes logical/physical
Rotation Rate: Solid State Device
Device is: In smartctl database [for details use: -P show]
ATA Version is: ATA/ATAPI-7 T13/1532D revision 1
SATA Version is: SATA 2.6, 3.0 Gb/s
Local Time is: Mon May 19 17:01:35 2014 UTC
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

|ID#
|
|ATTRIBUTE_NAME
|FLAG|VALUE|WORST
|THRESH
|TYPE|UPDATED|WHEN_FAILED|RAW_VALUE|
|3|Spin_Up_Time
|0x0020|100|100
|0|Old_age|Offline|-|0|
|4|Start_Stop_Count
|0x0030|100|100
|0|Old_age|Offline|-|0|
|5|Reallocated_Sector_Ct|0x0032|100|100|0|Old_age|Always|-|13|
|9|Power_On_Hours|0x0032|100|100|0|Old_age|Always|-|17292|
|12|Power_Cycle_Count|0x0032|100|100|0|Old_age|Always|-|1680|
|192|Unsafe_Shutdown_Count|0x0032|100|100|0|Old_age|Always|-|258|
|225|Host_Writes_32MiB|0x0030|100|100|0|Old_age|Offline|-|832764|
|226|Workld_Media_Wear_Indic|0x0032|100|100|0|Old_age|Always|-|2534700|
|227|Workld_Host_Reads_Perc|0x0032|100|100|0|Old_age|Always|-|0|
|228|Workload_Minutes|0x0032|100|100|0|Old_age|Always|-|3163330797|
|232|Available_Reservd_Space|0x0033|100|100|10|Pre-fail|Always|-|0|
|233|Media_Wearout_Indicator|0x0032|91|91|0|Old_age|Always|-|0|
|184|End-to-End_Error|0x0033|100|100|90|Pre-fail|Always|-|0|

Signs of failure? :open_mouth:

Don’t know but I have apache and some server apps installed for testing/education so I might require it. Anyway I installed desktop kernel and it has not complained yet.

On 2014-05-19 18:16, fhansson wrote:

> This is the output from smartctl:

Argh. Almost unreadable, I forgot to tell you to use code tags. The ‘#’
button - for command outputs and listings
See how

> Sector Size: 512 bytes logical/physical
> Rotation Rate: Solid State Device

Oh. It is a flash device. I don’t know how to interpret them.


> ID# 	ATTRIBUTE_NAME 	FLAG 	VALUE 	WORST 	THRESH	TYPE 	UPDATED 	WHEN_FAILED 	RAW_VALUE

> 5 	Reallocated_Sector_Ct 	0x0032 	100 	100 	0 	Old_age 	Always 	- 	13

On magnetic media, that one would be bad news.

> 9 	Power_On_Hours  0x0032 	100 	100 	0 	Old_age 	Always 	- 	17292

It is an old unit.

> 192 	Unsafe_Shutdown_Count 	0x0032 	100 	100 	0 	Old_age 	Always 	- 	258

No idea about that one, but I don't like it.



Cheers / Saludos,

Carlos E. R.

(from 13.1 x86_64 “Bottle” (Minas Tirith))

It’s an old drive but I think it is OK. Media_wearout_indicator is 91% so it should be alright I guess.

I tested uninstalling Plymouth but without any success. When I re-installed it with /usr/lib/plymouth/plymouth-update-initrd I got the following output:


Kernel image:   /boot/vmlinuz-3.11.10-7-desktop
Initrd image:   /boot/initrd-3.11.10-7-desktop
KMS drivers:     i915
Root device:    /dev/disk/by-id/ata-INTEL_SSDSA2M080G2GC_CVPO944500B6080BGN-part6 (/dev/sda6) (mounted on / as ext4)
Resume device:  /dev/disk/by-id/ata-INTEL_SSDSA2M080G2GC_CVPO944500B6080BGN-part5 (/dev/sda5)
Microcode: Adding Intel microcode 06-1c-02
Kernel Modules: thermal_sys thermal processor fan scsi_dh scsi_dh_hp_sw scsi_dh_emc scsi_dh_alua scsi_dh_rdac video button i2c-algo-bit drm drm_kms_helper i915 xhci-hcd hid-logitech-dj hid-holtek-kbd hid-lenovo-tpkbd hid-ortek hid-roccat hid-roccat-common hid-roccat-arvo hid-roccat-isku hid-samsung ohci-pci libcrc32c xor zlib_deflate raid6_pq btrfs crc32c-intel 
Features:       acpi intel_microcode kms plymouth block usb btrfs resume.userspace resume.kernel
lddlibc4: cannot read header from `/usr/sbin/fsck.btrfs': No such file or directory
lddlibc4: cannot read header from `/usr/sbin/fsck.btrfs': No such file or directory

I am not using btrfs so why should it be checked?

I meant you should completely uninstall the plymouth package.

I am not using btrfs so why should it be checked?

It doesn’t say it should be checked.
But it tries to install btrfs support into the initrd and fails, because /usr/sbin/fsck.btrfs is not found.

Do you have btrfsprogs installed? Remove it if you don’t use btrfs anyway.

What happens when you run “sudo /sbin/mkinitrd”?

On 2014-05-19 20:16, fhansson wrote:

> It’s an old drive but I think it is OK. Media_wearout_indicator is 91%
> so it should be alright I guess.

You have bad sectors remapped already. On a magnetic media disk, that’s
an indicator to consider replace unit, maybe as soon as possible.

Maybe you should run the short and long tests - if flash media do them.
Your smartctl output does not say.


Cheers / Saludos,

Carlos E. R.

(from 13.1 x86_64 “Bottle” (Minas Tirith))