Created attachment 741 [details] 默认没有crashkernel值 Description of problem: arm/x86镜像默认没有crashkernel值,无法启动kdump服务,报错为: anolis systemd[1]: kdump.service - Crash recovery kernel arming was skipped because of an unmet condition check (ConditionKernelCommandLine=crashkernel) 无法触发crash 以arm dvd-iso为例: # uname -r 5.10.134-14.an23.aarch64 # cat /etc/anolis-release Anolis OS release 23 # cat /etc/os-release NAME="Anolis OS" VERSION="23" ID="anolis" VERSION_ID="23" PLATFORM_ID="platform:an23" PRETTY_NAME="Anolis OS 23" ANSI_COLOR="0;31" HOME_URL="https://openanolis.cn/" BUG_REPORT_URL="https://bugzilla.openanolis.cn/" Version-Release number of selected component (if applicable): Steps to Reproduce: 1.yum install crash kexec-tools -y 2.systemctl start kdump && systemctl status kdump 3.cat /proc/cmdline Actual results: kdump服务无法启动,没有crashkernel Expected results: /proc/cmdline存在crashkernel值,kdump服务正常启动,执行echo c >/proc/sysrq-trigger正常触发crash Additional info:
问题镜像为anolis23-dvd-iso和anolis23-boot-iso镜像,包含x86和arm64
是不是系统配置的内存太小了？系统内存小于 8G 的时候，会出现 kdump 启动失败的现象。 我本地没有复现出来，启动服务是可以的。 [root@iZbp1gmbng4di4cw992erkZ ~]# systemctl start kdump [root@iZbp1gmbng4di4cw992erkZ ~]# systemctl status kdump ● kdump.service - Crash recovery kernel arming Loaded: loaded (/usr/lib/systemd/system/kdump.service; enabled; preset: enabled) Active: active (exited) since Sat 2023-06-10 14:59:36 CST; 3min 32s ago Main PID: 873 (code=exited, status=0/SUCCESS) CPU: 17.807s Jun 10 22:59:26 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Install squash loader *** Jun 10 22:59:26 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Stripping files *** Jun 10 14:59:28 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Stripping files done *** Jun 10 14:59:28 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Squashing the files inside the initramfs > Jun 10 14:59:35 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Squashing the files inside the initramfs > Jun 10 14:59:35 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Creating image file '/boot/initramfs-5.10> Jun 10 14:59:35 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Creating initramfs image file '/boot/init> Jun 10 14:59:36 iZbp1gmbng4di4cw992erkZ kdumpctl[889]: kdump: kexec: loaded kdump kernel Jun 10 14:59:36 iZbp1gmbng4di4cw992erkZ kdumpctl[889]: kdump: Starting kdump: [OK] Jun 10 14:59:36 iZbp1gmbng4di4cw992erkZ systemd[1]: Finished kdump.service - Crash recovery kernel > [root@iZbp1gmbng4di4cw992erkZ ~]# cat /etc/os-release NAME="Anolis OS" VERSION="23" ID="anolis" VERSION_ID="23" PLATFORM_ID="platform:an23" PRETTY_NAME="Anolis OS 23" ANSI_COLOR="0;31" HOME_URL="https://openanolis.cn/" BUG_REPORT_URL="https://bugzilla.openanolis.cn/"
(In reply to happy_orange from comment #2) > 是不是系统配置的内存太小了？系统内存小于 8G 的时候，会出现 kdump 启动失败的现象。 > > 我本地没有复现出来，启动服务是可以的。 > [root@iZbp1gmbng4di4cw992erkZ ~]# systemctl start kdump > [root@iZbp1gmbng4di4cw992erkZ ~]# systemctl status kdump > ● kdump.service - Crash recovery kernel arming > Loaded: loaded (/usr/lib/systemd/system/kdump.service; enabled; preset: > enabled) > Active: active (exited) since Sat 2023-06-10 14:59:36 CST; 3min 32s ago > Main PID: 873 (code=exited, status=0/SUCCESS) > CPU: 17.807s > > Jun 10 22:59:26 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Install squash > loader *** > Jun 10 22:59:26 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Stripping files *** > Jun 10 14:59:28 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Stripping files > done *** > Jun 10 14:59:28 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Squashing the > files inside the initramfs > > Jun 10 14:59:35 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Squashing the > files inside the initramfs > > Jun 10 14:59:35 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Creating image > file '/boot/initramfs-5.10> > Jun 10 14:59:35 iZbp1gmbng4di4cw992erkZ dracut[1225]: *** Creating initramfs > image file '/boot/init> > Jun 10 14:59:36 iZbp1gmbng4di4cw992erkZ kdumpctl[889]: kdump: kexec: loaded > kdump kernel > Jun 10 14:59:36 iZbp1gmbng4di4cw992erkZ kdumpctl[889]: kdump: Starting > kdump: [OK] > Jun 10 14:59:36 iZbp1gmbng4di4cw992erkZ systemd[1]: Finished kdump.service - > Crash recovery kernel > > [root@iZbp1gmbng4di4cw992erkZ ~]# cat /etc/os-release > NAME="Anolis OS" > VERSION="23" > ID="anolis" > VERSION_ID="23" > PLATFORM_ID="platform:an23" > PRETTY_NAME="Anolis OS 23" > ANSI_COLOR="0;31" > HOME_URL="https://openanolis.cn/" > BUG_REPORT_URL="https://bugzilla.openanolis.cn/" 额,我想说的是镜像里默认是没有crashkernel的,不是值太小的问题,需要手动在/etc/default/grub里添加crashkernel="xxx",没有这个crashkernel值,kdump服务是起不来的
Created attachment 782 [details] kernel-core 包 post 脚本执行 该问题仅存在 iso 启动场景里，vhd 启动没有问题。 经过初步定位，应该是 kernel-core 的安装脚本没有执行成功，手动执行脚本之后，是有 crashkernel 值的。
在 kernel-core 中尝试了增加 grubby 的依赖，并进行 iso 安装验证，经过验证之后发现该问题没有被修复。
anolis23 GA RC2版本测试依旧存在该问题
尝试使用 beta 版本，也是没有该变量，应该是没有预留空间。
可以先进行手动设置： （1）grep "GRUB_CMDLINE_LINUX" /etc/default/grub && sed -i 's/GRUB_CMDLINE_LINUX="/GRUB_CMDLINE_LINUX="crashkernel=0M-2G:0M,2G-8G:192M,8G-128G:256M,128G-:384M /g' /etc/default/grub; （2）/usr/sbin/grub2-mkconfig -o /boot/grub2/grub.cfg （3）reboot 重启 （4）重新启动 kdump 服务。
有 workaround 方法，先降级。
An23.1 RC1物理机镜像有相同问题 workaround方法可行，未阻塞物理机checklist [root@anolis ~]# cat /proc/cmdline BOOT_IMAGE=(hd1,gpt2)/vmlinuz-5.10.134-16.2_rc1.an23.x86_64 root=/dev/mapper/ao_anolis-root ro resume=/dev/mapper/ao_anolis-swap rd.lvm.lv=ao_anolis/root rd.lvm.lv=ao_anolis/swap rhgb quiet kdump服务： [root@anolis ~]# systemctl status kdump.service ○ kdump.service - Crash recovery kernel arming Loaded: loaded (/usr/lib/systemd/system/kdump.service; enabled; preset: enabled) Active: inactive (dead) Condition: start condition failed at Tue 2024-01-30 18:52:08 CST; 939ms ago └─ ConditionKernelCommandLine=crashkernel was not met Jan 05 17:17:32 anolis systemd[1]: kdump.service - Crash recovery kernel arming was skipped because of an unmet condition check (ConditionKernelCommandLine=cr> Jan 30 18:52:08 anolis systemd[1]: kdump.service - Crash recovery kernel arming was skipped because of an unmet condition check (ConditionKernelCommandLine=cr> ESCOC abled) was skipped because of an unmet condition check (ConditionKernelCommandLine=crashkernel). was skipped because of an unmet condition check (ConditionKernelCommandLine=crashkernel). [root@anolis ~]# cat /etc/os-release NAME="Anolis OS" VERSION="23" ID="anolis" VERSION_ID="23" PLATFORM_ID="platform:an23" PRETTY_NAME="Anolis OS 23" ANSI_COLOR="0;31" HOME_URL="https://openanolis.cn/" BUG_REPORT_URL="https://bugzilla.openanolis.cn/" [root@anolis ~]# uname -a Linux anolis 5.10.134-16.2_rc1.an23.x86_64 #1 SMP Tue Jan 2 10:09:28 CST 2024 x86_64 GNU/Linux
在an23.1 GA版本arm环境仍存在此问题 [root@anolis ~]# cat /etc/os-release NAME="Anolis OS" VERSION="23" ID="anolis" VERSION_ID="23" PLATFORM_ID="platform:an23" PRETTY_NAME="Anolis OS 23" ANSI_COLOR="0;31" HOME_URL="https://openanolis.cn/" BUG_REPORT_URL="https://bugzilla.openanolis.cn/" [root@anolis ~]# uname -r 6.6.25-2_rc1.an23.aarch64 [root@anolis ~]# cat /proc/cmdline BOOT_IMAGE=(hd0,gpt2)/vmlinuz-6.6.25-2_rc1.an23.aarch64 root=UUID=cb8e7cfe-0e15-478c-aad7-5903e90e574e ro selinux=0 [root@anolis ~]# systemctl start kdump.service [root@anolis ~]# systemctl status kdump.service ○ kdump.service - Crash recovery kernel arming Loaded: loaded (/usr/lib/systemd/system/kdump.service; enabled; preset: enabled) Active: inactive (dead) Condition: start condition unmet at Fri 2024-04-26 16:23:31 CST; 4s ago └─ ConditionKernelCommandLine=crashkernel was not met CPU: 0 4月 25 17:40:04 anolis systemd[1]: kdump.service - Crash recovery kernel arming was skipped because of an unmet condition check (ConditionKernelCom> 4月 26 16:23:31 anolis systemd[1]: kdump.service - Crash recovery kernel arming was skipped because of an unmet condition check (ConditionKernelCom>