Ubuntu /boot full – Server Fault

On an old server that still runs Ubuntu 16.04 the /boot partition is full, not with old kernels but with kernels that are all newer than the one that is running.

I’m running 4.4.0-176-generic, and the listing is like this (same with initrd.img* files):

-rw------- 1 root root 7220256 Dec  1 19:58 vmlinuz-4.4.0-198-generic
-rw------- 1 root root 7217792 Nov 27 14:03 vmlinuz-4.4.0-197-generic
-rw------- 1 root root 7220608 Nov 10 19:34 vmlinuz-4.4.0-194-generic
-rw------- 1 root root 7221056 Oct 14  2020 vmlinuz-4.4.0-193-generic
-rw------- 1 root root 7220448 Aug 31  2020 vmlinuz-4.4.0-190-generic
-rw------- 1 root root 7217792 Aug 14  2020 vmlinuz-4.4.0-189-generic
-rw------- 1 root root 7219360 Jul 22  2020 vmlinuz-4.4.0-187-generic
-rw------- 1 root root 7209656 Mar 16  2020 vmlinuz-4.4.0-177-generic
-rw------- 1 root root 7209560 Feb 28  2020 vmlinuz-4.4.0-176-generic

How can I safely clean things up?

I currently cannot install anything with apt-get, I get errors such as:

root@dev-server1:/boot# apt-get  install  postfix
Reading package lists... Done
Building dependency tree
Reading state information... Done
You might want to run 'apt-get -f install' to correct these:
The following packages have unmet dependencies:
 linux-image-generic : Depends: linux-image-4.4.0-200-generic but it is not going to be installed or
                                linux-image-unsigned-4.4.0-200-generic but it is not going to be installed
                       Recommends: thermald but it is not going to be installed
 linux-modules-extra-4.4.0-200-generic : Depends: linux-image-4.4.0-200-generic but it is not going to be installed or
                                                  linux-image-unsigned-4.4.0-200-generic but it is not going to be installed
 postfix : Depends: ssl-cert but it is not going to be installed
E: Unmet dependencies. Try 'apt-get -f install' with no packages (or specify a solution).

…which looks strange as it is referring to a kernel that is not even there.

Thanks,

George

boot – Realme C2 phone bricked, bootloader locked, can’t access fastboot or recovery mode, not getting connected to PC

My Realme C2 phone has been bricked, it had bootloader locked, and I can’t go to fastboot or recovery mode, it is not getting connected to PC.

After booting the phone, it shows this screen and not doing anything next.

The current image(boot/recovery) have been destroyed and can not boot.
Please flash the correct image or contact customer service to fix it.

I want to flash stock ROM but not able to do it because it’s not connecting to the PC.

how can i boot Oppo a12 into fastboot mode

I am not able to boot my oppo a12 into fast mode. Because i want to root my oppo a12 withot going fastbbot mode it is not possible .so please help me

if USB cable is not plugged, it will restart after 10 seconds while fast boot in redmi note 4

fastboot automatically reboots

"This solution not works
https://forum.xda-developers.com/t/fastboot-reboot-after-20-seconds.3673456/"

if USB cable is not plugged, it will restart after 10 seconds

linux – Metro boot locked me out of windows due to loop

Here is what happened.

I managed to successfully dual boot manjaro and windows 10, from windows, using a program called easyBCD. It was all working normally, but there was 1 little issue I found, which was that you had to go through the windows metro boot to select manjaro. and then go through manjaro’s dual boot process, which i found a bit tedious.

so…

I made the completely destructive choice of changing easybcd’s options so that it would skip the boot process and automatically boot linux*. Therefore, every time i choose to boot windows 10 from manjaro’s grub system, it just automatically takes me back to the grub bootloader. (atleast i think thats the case, windows doesnt load past the message “processing the preset menu…”)

Is there any way I could perhaps edit the windows 10 boot settings from linux to fix this?

*I ended up starting the issue by changing these freebcd settings:

  • skip the boot menu on.
  • set default to manjaro.

rom flashing – Emdoor EM TS809 Dead Boot – KCOLO point wanted to revive

Flashing with a bad firmware to this tablet became completely unresponsive and does not switch on. M/B marks report that is created from Emdoor and its model is EM TS809.

https://www.globalsources.com/Voice-activated-speaker/MT8167-EM-TS809_R70-1173578814p.htm#1173578814

I tried to revive it, by shorting GND and KCOLO joints and reflashing with SP Tools. Unfortunately, joints do not have descriptions so I tried every joint without voltage. None of them worked. Does anyone know if this M/B has KCOLO or just COLO/KOLO joint?

enter image description here

enter image description here

twrp – Device stuck in boot animation after reverting to LineageOS 17.1

I tried to update my OnePlus One to LineageOS 18.1, following my usual procedure:

  • do a full backup (boot, system, data excl. storage, cache) with TWRP
  • then flash the new image, along with the zip files Magisk and microG

As I then noticed that Android 11 breaks some of the apps I use, I decided to roll back the update by simply booting into recovery, restoring the backup and rebooting. However, this failed with the system being stuck in the boot animation for significantly above 10 minutes.

I then tried the following:

  • boot into TWRP
  • flash LineageOS 17.1 from the previous image I had installed (plus Magisk and microG)
  • restore my backup
  • wipe caches

Still no joy, I am still stuck in the boot animation.

Suspecting the upgrade had converted some app data in /sdcard/Android/data, I moved all app dirs starting with com.android or org.lineageos off the device. (The rest should be user app data. I do not have Google Play services installed.)

That still was not successful, I am still stuck in the startup sequence.

I have never (knowingly) encrypted any data on the device.

What gives? How can I find out where the device gets stuck, so I can hunt the error down? Wiping all user data is not an option.

applications – Disable autostart on boot and restrict background execution

Without getting into the details whether or not one should stop autostarting apps, and what could be the consequences as discussed in a plenty of other answers, here are my simple solutions which may work without any third party apps, at least on Android Pie.

ROOT SOLUTION:

Apps use BroadcastRceivers to listen for broadcast ACTION_BOOT_COMPLETED so that they can run on boot. Broadcast receiver is an app component that can be disabled with root privileges.

Using Package Manager list all broadcast receivers listening for BOOT_COMPLETED:

~# pm query-receivers --components -a android.intent.action.BOOT_COMPLETED

It will give a list of broadcast receivers in format package_name/component_name. See dumpsys activity broadcast-stats and dumpsys activity broadcasts for more details.

Now to disable a component:

~# pm disable <package/component>

There are apps like Autostarts (com.elsdoerfer.android.autostarts) and SD Maid (eu.thedarken.sdm) which can do same for you. File /data/system/users/0/package-restrictions.xml can also be edited directly to disable apps or their components, but it’s not recommended.
It’s possible to disable multiple broadcast receivers of an app, and a single receiver can also possibly listen to multiple types of broadcast events.

NON-ROOT SOLUTION: (non-universal)

In order to receive android.intent.action.BOOT_COMPLETED, apps need android.permission.RECEIVE_BOOT_COMPLETED which is a normal permission and hence can’t be revoked by user.

However there is a hidden permission management framework, named AppOps that provides a slightly more fine-grained control of (permission-like) operations. OP_BOOT_COMPLETED is one of those but it’s not a part of AOSP, only added by some custom ROMs like LineageOS. If you are on one of such ROMs, you can control the autostart behavior through adb shell:

~$ appops set <package> BOOT_COMPLETED deny

Now the app won’t be allowed to receive BOOT_COMPLETED broadcast. There are apps like App Ops (rikka.appops) which can do same for you. Some custom ROMs have built-in front-ends to AppOps with different names like Privacy Guard, AutoStart Manager etc.

Please note that AppOps:

is not generally intended for third party application developers; most features are only available to system applications

So its usage without root may be disallowed or get harder in next Android releases.


Both of the above methods can stop apps from starting on boot only. An app can listen for some other broadcast events too and it can keep on restarting if killed, or run in background continuously (as a service) if it’s designed to be so. See dumpsys activity services for more details.

A slightly different approach would be to stop apps from running in background by using OP_RUN_IN_BACKGROUND (introduced in Nougat) and/or RUN_ANY_IN_BACKGROUND (introduced in Pie) which are part of AOSP:

~$ appops set <package> RUN_IN_BACKGROUND deny

They don’t have an equivalent manifest permission, but there is an experimental permission with same name.

boot – How to rename/delete .kext file in Big Sur

I have one of those MacBook Pro’s (2015) with the hardware fault which causes frequent spontaneous shutdowns.

The problem can be solved by renaming or deleting AppleThunderboltNHI.kext which I have done after every MacOS Catalina update for the past couple of years, basically everything except step 7 & 8 below.

In Big Sur, the change doesn’t “stick,” the AppleThunderboltNHI.kext file is restored automatically after restarting in normal mode, so I have followed the procedure which works according to many on the discussion referenced below.

Everything goes okay unless I do step 8, at which point the MacOS will not boot, I get a message like “…was restarted because of a problem…”

The only thing that seems to fix that, is to reinstall Big Sur from recovery mode, which does work. But then the AppleThunderboltNHI.kext file is restored.

I suspect that something in step 8 is causing this procedure to fail somehow but I am not at all familiar with using “snapshots” or the “bless” utility. So I’m hoping someone might be able to help me solve it.

Note: The following instructions are not mine, and I am not recommending them as a solution.

https://developer.apple.com/forums/thread/666567?answerId=646500022#646500022

Posted by sf98723

Follow this procedure :

  1. Reboot in rescue mode (reboot while “Cmd + R”)

  2. csrutil disable

  3. csrutil authenticated-root disable

  4. reboot in rescue mode

  5. mount -uw /Volumes/(MacOS )

  6. delete (or rename, or move elsewhere) the AppleThunderboltNHI.kext directory (I’ve moved all thunderbolt kext directories since I have no
    needs about this interface, but I think it works only by disabling
    AppleThunderboltNHI)

  7. REBUILD the extensions cache ! (new has-to-do in Big Sur…) : kmutil install -u –force –volume-root /Volumes/(MacOS )

  8. DON’T FORGET to create another system snapshot to take these modifications under account at next reboot : bless –folder
    /Volumes/(MacOS )/System/Library/CoreServices
    –bootefi –create-snapshot

  9. csrutil enable (mays be you have to reboot in rescue mode.. not tested)

  10. Finally, reboot in normal mode

HPE Proliant Gen10 Stack Fault Exception during PXE boot with UEFI

I am trying to pxe boot a HPE Proliant Gen10 using UEFI BIOS with RHEL (have tried both 7 and 8) and am getting a stack fault exception after the server receives the first boot file (shimx64.efi):

HPE ProLiant System BIOS U32 v2.42 (01/23/2021)
(C) Copyright 1982-2021 Hewlett Packard Enterprise Development LP
Early system initialization, please wait…

iLO 5 IPv4:
iLO 5 IPv6:

iLO 5 IPv4: (redacted)
iLO 5 IPv6: (redacted)

2%: System Chipset Initialization
4%: UPI Link Initialization – Start
5%: UPI Link Initialization – Complete
7%: Early Processor Initialization
9%: Memory Initialization – Start
10%: 11%: 13%: 14%: 15%: 21%: Memory Initialization – Complete
25%: SMBIOS Table Initialization
28%: USB Controller Initialization
31%: Chipset Root Ports Initialization
35%: HPE SmartMemory Initialization
38%: System Security Initialization
42%: iLO Embedded Health Initialization
45%: ACPI Table Initialization
49%: BIOS Configuration Initialization
L 55%: Early PCI Initialization – Start
62%: Early PCI Initialization – Complete
73%: Switching console output to Primary Video. Please wait…

(=3h

(=3h
(01;01

(=3h

(=3h

(=3h

(=3h

(C) Copyright 1982-2021 Hewlett Packard Enterprise Development LP
HPE ProLiant DL360 Gen10
System ROM Version: U32 v2.42 (01/23/2021)
Serial Number: (redacted)

Installed System Memory: 384 GB, Available System Memory: 384 GB

2 Processor(s) detected, 40 total cores enabled, Hyperthreading is enabled
Proc 1: Intel(R) Xeon(R) Gold 6248 CPU @ 2.50GHz
Proc 2: Intel(R) Xeon(R) Gold 6248 CPU @ 2.50GHz
UPI Speed: 10.4 GT/s

Workload Profile: I/O Throughput
Power Regulator Mode: Dynamic Power Savings
Advanced Memory Protection Mode: Fast Fault Tolerant Memory (ADDDC)
Boot Mode: UEFI
HPE SmartMemory authenticated in all populated DIMM slots.

For access via BIOS Serial Console:
Press ‘ESC+9’ for System Utilities
Press ‘ESC+0’ for Intelligent Provisioning
Press ‘ESC+!’ for One-Time Boot Menu
Press ‘ESC+@’ for Network Boot

Starting required devices. Please wait, this may take a few moments….

Booting Embedded FlexibleLOM 1 Port 1 : HPE Eth 10/25Gb 2p 640FLR-SFP28 Adptr – NIC (PXE IPv4)

Start PXE over IPv4.
Station IP address is (redacted)

Server IP address is (redacted)
NBP filename is bootloader/pxe.boot.host/image
NBP filesize is 1877432 Bytes
Downloading NBP file…

NBP file downloaded successfully.

X64 Exception Type 0x0C – Stack Fault Exception

RCX=0000000000000001 DX=0000000036E13971 R8=00000000A0D29060 R9=000000007B2314F0
RSP=000000007B2315F8 BP=AFAFAFAFAFAFAFAF AX=0000000000000000 BX=000000003932D000
R10=000000007B231558 11=9A02503F3EB2EC9E 12=0000000000000000 13=0000000000000040
R14=00000000982470C0 15=0000000000000001 SI=000000009ED79018 DI=0000000000000E40
CR2=0000000000000000 CR3=00000000A347B000 CR0=80000013 CR4=00000668 CR8=00000000
CS=00000038 DS=00000030 SS=00000030 ES=00000030 RFLAGS=00210286
MSR: 0x1D9 = 00004801, 0x345=000033C5, 0x1C9=00000004

LBRs From To From To
01h 0000000036E0DE86->000000009A1D2174 0000000036E0E4C4->0000000036E0DE23
03h 0000000036E0E977->0000000036E0E4C4 000000007B0FB7AB->0000000036E0E948
05h 000000007B0FB82C->000000007B0FB85E 000000007B0FB738->000000007B0FB80C
07h 000000007B0FB73D->000000007B0FB71B 000000007B0FB863->000000007B0FB73A
09h 000000007B0FB868->000000007B0FB860 000000007B0FB844->000000007B0FB864
0Bh 000000007B0FB738->000000007B0FB80C 000000007B0FB73D->000000007B0FB71B
0Dh 000000007B0FB863->000000007B0FB73A 000000007B0FB868->000000007B0FB860
0Fh 000000007B0FB844->000000007B0FB864 000000009A1D217F->000000009FBAD940

CALL ImageBase ImageName+Offset
00h 0000000000000000 No Image Information
01h 0000000000000000 No Image Information

CALL ImageBase ImageName+Offset

STACK 00h 04h 08h 0Ch 10h 14h 18h 1Ch
RSP+00h 9EFC7818 9A08AB98 9EFBF56D 9EFC7818 00000000 9ED79018 00000000 4CF5B200
RSP+20h 9EFC1B9B 4CF5B200 733AA782 7B231710 982470C0 365CA000 00000040 7B109960
RSP+40h 733A904F 3932D000 365CA000 7331C340 733A787F 734C8EE5 7340ADE0 734C8F55
RSP+60h 7331E4A8 734C84A0 73450911 7331C048 7331E4A0 7B231800 7331C320 9EFBAC48
RSP+80h 7331E520 7B231840 00000001 7331E420 7B231870 734500AB 7331C048 7331E500
RSP+A0h 9EFBAC48 73450A83 7331C288 7331E580 734500AB 7331C288 7331E640 00000000
RSP+C0h 73450EC5 7331E5E0 7344CF6E 73324300 7331E640 7331E5E0 73438071 73325580
RSP+E0h 7331E640 73324300 7331EA52 7331EA40 00000000 734382B2 73325580 00000000