/etc/kdump/pre.d/* Not getting Executed

/etc/kdump/pre.d/* Not getting Executed

[arshad hussain]

Hello,

I am trying to run a custom script before the dump (ie
/etc/kdump/pre.d/* ) which I am unable to do.
I can say this, as the script is supposed to create a folder under
/var/crash and a file which it is not. It also does not generate the
vmcore file.

I am failing to understand what obvious step I am missing. Can someone
please help out.

While the "echo c > /proc/sysrq-trigger" successfully reboots the
output is not what is expected.

Here is my setup for kdump/kexec.

Thanks!

-------Expected output-------
01. Under /var/crash we should be having a folder as "DBGMSG-<date>".
02. Under /var/crash we should be having a file as "test512".
03. vmcore file
04. dmesg file

-------current output-------
Nothing under /var/crash

-------Below are the info and conf script details-------
# uname -a (Kernel version)
Linux rocky8 4.18.0-240.1.1.el8.x86_64 #1 SMP Fri Feb 19 20:41:01 UTC
2021 x86_64 x86_64 x86_64 GNU/Linux

# rpm -qa | grep -i kexec (kexec-tool version)
kexec-tools-2.0.20-46.el8.x86_64

# cat /proc/cmdline
BOOT_IMAGE=(hd0,msdos1)/boot/vmlinuz-4.18.0-240.1.1.el8.x86_64
root=UUID=<UUID...> ro crashkernel=128M resume=UUID=<UUID...>


------/etc/kdump.conf-------
path /var/crash
core_collector makedumpfile -l --message-level 7 -d 31
extra_bins /var/crash/kdumppre.sh /bin/bash /usr/bin/cat /usr/bin/date
kdump_pre /var/crash/kdumppre.sh
force_rebuild 1

# tree /etc/kdump/
/etc/kdump/
??? post.d
??? pre.d
    ??? 50-collect.sh

-------/etc/kdump/pre.d/50-collect.sh-------
#!/bin/bash

# base location
LOC=/sysroot/var/crash
# dbgmsg identifier string
DBGMSG="DBGMSG"
# log file
LFILE=log.out
# Granularity in seconds
LOCFOLDER=${LOC}/$DBGMSG-$(date +%Y-%m-%d:%H:%M:%S)

# Create folder.
mkdir -p $LOCFOLDER

if [ -f "/proc/spl/kstat/zfs/dbgmsg" ]; then
    echo "/proc/spl/kstat/zfs/dbgmsg found" >> $LOCFOLDER/$LFILE
    cat /proc/spl/kstat/zfs/dbgmsg >> $LOCFOLDER/$LFILE
else
    echo "/proc/spl/kstat/zfs/dbgmsg not found" >> $LOCFOLDER/$LFILE
fi

exit 0

-------/var/crash/kdumppre.sh-------
#!/bin/bash
echo "Test512.......$(date)" >> /sysroot/var/carsh/test512
exit 0


(Lines starting with [Debug] is the debug statement put for debugging
to verify if we are picking up correct key=value pairs)
# kdumpctl restart
kdump: DEFAULT_DUMP_MODE=kdump
kdump: kexec: unloaded kdump kernel
kdump: Stopping kdump: [OK]
[DEBUG]found config... path : /var/crash
[DEBUG]found config... core_collector : makedumpfile -l --message-level 7 -d 31
[DEBUG]found config... extra_bins : /var/crash/kdumppre.sh /bin/bash
/usr/bin/cat /usr/bin/date
[DEBUG]found config... kdump_pre : /var/crash/kdumppre.sh
[DEBUG]found config... force_rebuild : 1
kdump: restore default initrd:
/boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64.img
[DEBUG]files=/etc/kdump.conf
/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64  /var/crash/kdumppre.sh
/var/crash/kdumppre.sh /bin/bash /usr/bin/cat /usr/bin/date
/usr/sbin/makedumpfile
[DEBUG]hook files=/etc/kdump.conf
/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64  /var/crash/kdumppre.sh
/var/crash/kdumppre.sh /bin/bash /usr/bin/cat /usr/bin/date
/usr/sbin/makedumpfile /etc/fstab /etc/kdump/post.d/ /etc/kdump/pre.d/
  /etc/kdump/pre.d/50-collect.sh :: /etc/kdump/post.d/
/etc/kdump/pre.d/   /etc/kdump/pre.d/50-collect.sh
kdump: Detected change(s) in the following file(s):  /etc/kdump/pre.d/
kdump: Force rebuild /boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
kdump: Rebuilding /boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
kdump: rebuild kdump initrd: /sbin/mkdumprd -f
/boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
4.18.0-240.1.1.el8_lustre.x86_64
kdump: /sbin/kexec -s -d -p
--command-line=BOOT_IMAGE=(hd0,msdos1)/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64
ro resume=UUID=d5fad714-147f-419c-a8ea-9e7984b0ef2f irqpoll nr_cpus=1
reset_devices cgroup_disable=memory mce=off numa=off
udev.children-max=2 panic=10 rootflags=nofail acpi_no_memhotplug
transparent_hugepage=never nokaslr novmcoredd hest_disable
disable_cpu_apicid=0
--initrd=/boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64
kdump: kexec: loaded kdump kernel
kdump: Starting kdump: [OK]

Re: /etc/kdump/pre.d/* Not getting Executed

[Coiby Xu]

Hi,

Have you resolved this issue? If not, feel free to open a bug on
https://bugzilla.redhat.com since /etc/kdump/pre.d/ is a feature
provided by Fedora/RHEL's kexec-tools.


On Wed, May 11, 2022 at 01:42:40PM +0530, arshad hussain wrote:
>Hello,
>
>I am trying to run a custom script before the dump (ie
>/etc/kdump/pre.d/* ) which I am unable to do.
>I can say this, as the script is supposed to create a folder under
>/var/crash and a file which it is not. It also does not generate the
>vmcore file.
>
>I am failing to understand what obvious step I am missing. Can someone
>please help out.
>
>While the "echo c > /proc/sysrq-trigger" successfully reboots the
>output is not what is expected.
>
>Here is my setup for kdump/kexec.
>
>Thanks!
>
>-------Expected output-------
>01. Under /var/crash we should be having a folder as "DBGMSG-<date>".
>02. Under /var/crash we should be having a file as "test512".
>03. vmcore file
>04. dmesg file
>
>-------current output-------
>Nothing under /var/crash
>
>-------Below are the info and conf script details-------
># uname -a (Kernel version)
>Linux rocky8 4.18.0-240.1.1.el8.x86_64 #1 SMP Fri Feb 19 20:41:01 UTC
>2021 x86_64 x86_64 x86_64 GNU/Linux
>
># rpm -qa | grep -i kexec (kexec-tool version)
>kexec-tools-2.0.20-46.el8.x86_64
>
># cat /proc/cmdline
>BOOT_IMAGE=(hd0,msdos1)/boot/vmlinuz-4.18.0-240.1.1.el8.x86_64
>root=UUID=<UUID...> ro crashkernel=128M resume=UUID=<UUID...>
>
>
>------/etc/kdump.conf-------
>path /var/crash
>core_collector makedumpfile -l --message-level 7 -d 31
>extra_bins /var/crash/kdumppre.sh /bin/bash /usr/bin/cat /usr/bin/date
>kdump_pre /var/crash/kdumppre.sh
>force_rebuild 1
>
># tree /etc/kdump/
>/etc/kdump/
>├── post.d
>└── pre.d
>    └── 50-collect.sh
>
>-------/etc/kdump/pre.d/50-collect.sh-------
>#!/bin/bash
>
># base location
>LOC=/sysroot/var/crash
># dbgmsg identifier string
>DBGMSG="DBGMSG"
># log file
>LFILE=log.out
># Granularity in seconds
>LOCFOLDER=${LOC}/$DBGMSG-$(date +%Y-%m-%d:%H:%M:%S)
>
># Create folder.
>mkdir -p $LOCFOLDER
>
>if [ -f "/proc/spl/kstat/zfs/dbgmsg" ]; then
>    echo "/proc/spl/kstat/zfs/dbgmsg found" >> $LOCFOLDER/$LFILE
>    cat /proc/spl/kstat/zfs/dbgmsg >> $LOCFOLDER/$LFILE
>else
>    echo "/proc/spl/kstat/zfs/dbgmsg not found" >> $LOCFOLDER/$LFILE
>fi
>
>exit 0
>
>-------/var/crash/kdumppre.sh-------
>#!/bin/bash
>echo "Test512.......$(date)" >> /sysroot/var/carsh/test512
>exit 0
>
>
>(Lines starting with [Debug] is the debug statement put for debugging
>to verify if we are picking up correct key=value pairs)
># kdumpctl restart
>kdump: DEFAULT_DUMP_MODE=kdump
>kdump: kexec: unloaded kdump kernel
>kdump: Stopping kdump: [OK]
>[DEBUG]found config... path : /var/crash
>[DEBUG]found config... core_collector : makedumpfile -l --message-level 7 -d 31
>[DEBUG]found config... extra_bins : /var/crash/kdumppre.sh /bin/bash
>/usr/bin/cat /usr/bin/date
>[DEBUG]found config... kdump_pre : /var/crash/kdumppre.sh
>[DEBUG]found config... force_rebuild : 1
>kdump: restore default initrd:
>/boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64.img
>[DEBUG]files=/etc/kdump.conf
>/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64  /var/crash/kdumppre.sh
>/var/crash/kdumppre.sh /bin/bash /usr/bin/cat /usr/bin/date
>/usr/sbin/makedumpfile
>[DEBUG]hook files=/etc/kdump.conf
>/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64  /var/crash/kdumppre.sh
>/var/crash/kdumppre.sh /bin/bash /usr/bin/cat /usr/bin/date
>/usr/sbin/makedumpfile /etc/fstab /etc/kdump/post.d/ /etc/kdump/pre.d/
>  /etc/kdump/pre.d/50-collect.sh :: /etc/kdump/post.d/
>/etc/kdump/pre.d/   /etc/kdump/pre.d/50-collect.sh
>kdump: Detected change(s) in the following file(s):  /etc/kdump/pre.d/
>kdump: Force rebuild /boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
>kdump: Rebuilding /boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
>kdump: rebuild kdump initrd: /sbin/mkdumprd -f
>/boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
>4.18.0-240.1.1.el8_lustre.x86_64
>kdump: /sbin/kexec -s -d -p
>--command-line=BOOT_IMAGE=(hd0,msdos1)/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64
>ro resume=UUID=d5fad714-147f-419c-a8ea-9e7984b0ef2f irqpoll nr_cpus=1
>reset_devices cgroup_disable=memory mce=off numa=off
>udev.children-max=2 panic=10 rootflags=nofail acpi_no_memhotplug
>transparent_hugepage=never nokaslr novmcoredd hest_disable
>disable_cpu_apicid=0
>--initrd=/boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
>/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64
>kdump: kexec: loaded kdump kernel
>kdump: Starting kdump: [OK]
>
>_______________________________________________
>kexec mailing list
>kexec@lists.infradead.org
>http://lists.infradead.org/mailman/listinfo/kexec

-- 
Best regards,
Coiby


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

Re: /etc/kdump/pre.d/* Not getting Executed

[arshad hussain]

Hi Coiby Xu,

I have not yet resolved this so far. I could get it working once and
thereafter I could not reproduce it. So I do not have a deterministic
step to reproduce this. Still WIP from my end. I have already posted
the question to kxec folkes:
http://lists.infradead.org/pipermail/kexec/2022-May/024919.html. Will
also post the bug to redhat.

Thanks
Arshad

On Tue, Jun 7, 2022 at 8:05 AM Coiby Xu <coxu@redhat.com> wrote:
>
> Hi,
>
> Have you resolved this issue? If not, feel free to open a bug on
> https://bugzilla.redhat.com since /etc/kdump/pre.d/ is a feature
> provided by Fedora/RHEL's kexec-tools.
>
>
> On Wed, May 11, 2022 at 01:42:40PM +0530, arshad hussain wrote:
> >Hello,
> >
> >I am trying to run a custom script before the dump (ie
> >/etc/kdump/pre.d/* ) which I am unable to do.
> >I can say this, as the script is supposed to create a folder under
> >/var/crash and a file which it is not. It also does not generate the
> >vmcore file.
> >
> >I am failing to understand what obvious step I am missing. Can someone
> >please help out.
> >
> >While the "echo c > /proc/sysrq-trigger" successfully reboots the
> >output is not what is expected.
> >
> >Here is my setup for kdump/kexec.
> >
> >Thanks!
> >
> >-------Expected output-------
> >01. Under /var/crash we should be having a folder as "DBGMSG-<date>".
> >02. Under /var/crash we should be having a file as "test512".
> >03. vmcore file
> >04. dmesg file
> >
> >-------current output-------
> >Nothing under /var/crash
> >
> >-------Below are the info and conf script details-------
> ># uname -a (Kernel version)
> >Linux rocky8 4.18.0-240.1.1.el8.x86_64 #1 SMP Fri Feb 19 20:41:01 UTC
> >2021 x86_64 x86_64 x86_64 GNU/Linux
> >
> ># rpm -qa | grep -i kexec (kexec-tool version)
> >kexec-tools-2.0.20-46.el8.x86_64
> >
> ># cat /proc/cmdline
> >BOOT_IMAGE=(hd0,msdos1)/boot/vmlinuz-4.18.0-240.1.1.el8.x86_64
> >root=UUID=<UUID...> ro crashkernel=128M resume=UUID=<UUID...>
> >
> >
> >------/etc/kdump.conf-------
> >path /var/crash
> >core_collector makedumpfile -l --message-level 7 -d 31
> >extra_bins /var/crash/kdumppre.sh /bin/bash /usr/bin/cat /usr/bin/date
> >kdump_pre /var/crash/kdumppre.sh
> >force_rebuild 1
> >
> ># tree /etc/kdump/
> >/etc/kdump/
> >├── post.d
> >└── pre.d
> >    └── 50-collect.sh
> >
> >-------/etc/kdump/pre.d/50-collect.sh-------
> >#!/bin/bash
> >
> ># base location
> >LOC=/sysroot/var/crash
> ># dbgmsg identifier string
> >DBGMSG="DBGMSG"
> ># log file
> >LFILE=log.out
> ># Granularity in seconds
> >LOCFOLDER=${LOC}/$DBGMSG-$(date +%Y-%m-%d:%H:%M:%S)
> >
> ># Create folder.
> >mkdir -p $LOCFOLDER
> >
> >if [ -f "/proc/spl/kstat/zfs/dbgmsg" ]; then
> >    echo "/proc/spl/kstat/zfs/dbgmsg found" >> $LOCFOLDER/$LFILE
> >    cat /proc/spl/kstat/zfs/dbgmsg >> $LOCFOLDER/$LFILE
> >else
> >    echo "/proc/spl/kstat/zfs/dbgmsg not found" >> $LOCFOLDER/$LFILE
> >fi
> >
> >exit 0
> >
> >-------/var/crash/kdumppre.sh-------
> >#!/bin/bash
> >echo "Test512.......$(date)" >> /sysroot/var/carsh/test512
> >exit 0
> >
> >
> >(Lines starting with [Debug] is the debug statement put for debugging
> >to verify if we are picking up correct key=value pairs)
> ># kdumpctl restart
> >kdump: DEFAULT_DUMP_MODE=kdump
> >kdump: kexec: unloaded kdump kernel
> >kdump: Stopping kdump: [OK]
> >[DEBUG]found config... path : /var/crash
> >[DEBUG]found config... core_collector : makedumpfile -l --message-level 7 -d 31
> >[DEBUG]found config... extra_bins : /var/crash/kdumppre.sh /bin/bash
> >/usr/bin/cat /usr/bin/date
> >[DEBUG]found config... kdump_pre : /var/crash/kdumppre.sh
> >[DEBUG]found config... force_rebuild : 1
> >kdump: restore default initrd:
> >/boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64.img
> >[DEBUG]files=/etc/kdump.conf
> >/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64  /var/crash/kdumppre.sh
> >/var/crash/kdumppre.sh /bin/bash /usr/bin/cat /usr/bin/date
> >/usr/sbin/makedumpfile
> >[DEBUG]hook files=/etc/kdump.conf
> >/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64  /var/crash/kdumppre.sh
> >/var/crash/kdumppre.sh /bin/bash /usr/bin/cat /usr/bin/date
> >/usr/sbin/makedumpfile /etc/fstab /etc/kdump/post.d/ /etc/kdump/pre.d/
> >  /etc/kdump/pre.d/50-collect.sh :: /etc/kdump/post.d/
> >/etc/kdump/pre.d/   /etc/kdump/pre.d/50-collect.sh
> >kdump: Detected change(s) in the following file(s):  /etc/kdump/pre.d/
> >kdump: Force rebuild /boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
> >kdump: Rebuilding /boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
> >kdump: rebuild kdump initrd: /sbin/mkdumprd -f
> >/boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
> >4.18.0-240.1.1.el8_lustre.x86_64
> >kdump: /sbin/kexec -s -d -p
> >--command-line=BOOT_IMAGE=(hd0,msdos1)/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64
> >ro resume=UUID=d5fad714-147f-419c-a8ea-9e7984b0ef2f irqpoll nr_cpus=1
> >reset_devices cgroup_disable=memory mce=off numa=off
> >udev.children-max=2 panic=10 rootflags=nofail acpi_no_memhotplug
> >transparent_hugepage=never nokaslr novmcoredd hest_disable
> >disable_cpu_apicid=0
> >--initrd=/boot/initramfs-4.18.0-240.1.1.el8_lustre.x86_64kdump.img
> >/boot/vmlinuz-4.18.0-240.1.1.el8_lustre.x86_64
> >kdump: kexec: loaded kdump kernel
> >kdump: Starting kdump: [OK]
> >
> >_______________________________________________
> >kexec mailing list
> >kexec@lists.infradead.org
> >http://lists.infradead.org/mailman/listinfo/kexec
>
> --
> Best regards,
> Coiby
>

_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec