记一次CentOS 7.6内核升级(5.9.6)及故障总结
导入源
# rpm -import https://www.elrepo.org/RPM-GPG-KEY-elrepo.org
# rpm -Uvh http://www.elrepo.org/elrepo-release-7.0-2.el7.elrepo.noarch.rpm
# 查看kernrl版本
# yum --disablerepo="*" --enablerepo="elrepo-kernel" list available
Loaded plugins: fastestmirror, langpacks
Loading mirror speeds from cached hostfile
* elrepo-kernel: syd.mirror.rackspace.com
Available Packages
elrepo-release.noarch 7.0-5.el7.elrepo elrepo-kernel
kernel-lt.x86_64 4.4.241-1.el7.elrepo elrepo-kernel
kernel-lt-devel.x86_64 4.4.241-1.el7.elrepo elrepo-kernel
kernel-lt-doc.noarch 4.4.241-1.el7.elrepo elrepo-kernel
kernel-lt-headers.x86_64 4.4.241-1.el7.elrepo elrepo-kernel
kernel-lt-tools.x86_64 4.4.241-1.el7.elrepo elrepo-kernel
kernel-lt-tools-libs.x86_64 4.4.241-1.el7.elrepo elrepo-kernel
kernel-lt-tools-libs-devel.x86_64 4.4.241-1.el7.elrepo elrepo-kernel
kernel-ml-doc.noarch 5.9.6-1.el7.elrepo elrepo-kernel
kernel-ml-headers.x86_64 5.9.6-1.el7.elrepo elrepo-kernel
kernel-ml-tools.x86_64 5.9.6-1.el7.elrepo elrepo-kernel
kernel-ml-tools-libs.x86_64 5.9.6-1.el7.elrepo elrepo-kernel
kernel-ml-tools-libs-devel.x86_64 5.9.6-1.el7.elrepo elrepo-kernel
perf.x86_64 5.9.6-1.el7.elrepo elrepo-kernel
python-perf.x86_64 5.9.6-1.el7.elrepo elrepo-kernel安装内核
版本性质：主分支ml(mainline)，稳定版(stable)，长期维护版lt(longterm)
本文安装的是主分支ml的5.9.6
# yum -y --enablerepo=elrepo-kernel install kernel-ml.x86_64 kernel-ml-devel.x86_64
# 查看安装好的kernel,已经安装成功
[root@qd01-stop-k8s-node004 ~]# rpm -qa|grep kernel
kernel-headers-3.10.0-957.el7.x86_64
kernel-tools-libs-3.10.0-957.el7.x86_64
kernel-ml-devel-5.9.6-1.el7.elrepo.x86_64
kernel-3.10.0-957.el7.x86_64
abrt-addon-kerneloops-2.1.11-52.el7.centos.x86_64
kernel-devel-3.10.0-957.el7.x86_64
kernel-tools-3.10.0-957.el7.x86_64
kernel-ml-5.9.6-1.el7.elrepo.x86_64
libreport-plugin-kerneloops-2.1.11-42.el7.centos.x86_64查看并修改grub的启动顺序
主要修改了GRUB_DEFAULT=0
cat /etc/default/grub
GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR="$(sed 's, release .*$,,g' /etc/system-release)"
GRUB_DEFAULT=0
GRUB_DISABLE_SUBMENU=true
GRUB_TERMINAL_OUTPUT="console"
GRUB_CMDLINE_LINUX="ipv6.disable=1 crashkernel=auto biosdevname=0 net.ifnames=0 rhgb quiet"
GRUB_DISABLE_RECOVERY="true"生成grub重新创建内核配置并重启
# grub2-mkconfig -o /boot/grub2/grub.cfg
# reboot重启确认内核版本
如果重启后如果发现系统一直没起来。
~恭喜你~，进入排错环节
这里通过管理网卡，查看报错如下图
CentOS 7.6 从内核 3.10.0 升级到 5.x后，重启不了，原因是 Linux 内核 4.4.x 版本之后，dracut 命令生成的 initramfs 默认不会自动加载 mpt3sas 驱动模块。因为没有对应的驱动程序，所以识别不了系统盘，从而导致内核 vmlinuz 无法加载它的 rootfs
问题解决
1、查询initramfs的驱动
# lsinitrd -k 3.10.0-957.el7.x86_64 | grep mpt[23]sas
drwxr-xr-x 2 root root 0 Nov 2 20:22 usr/lib/modules/3.10.0-957.el7.x86_64/kernel/drivers/scsi/mpt3sas
-rw-r--r-- 1 root root 92972 Nov 9 2018 usr/lib/modules/3.10.0-957.el7.x86_64/kernel/drivers/scsi/mpt3sas/mpt2sas.ko.xz
# ls /usr/lib/modules/5.9.6-1.el7.elrepo.x86_64/kernel/drivers/scsi/mpt3sas/mpt3sas.ko
/usr/lib/modules/5.9.6-1.el7.elrepo.x86_64/kernel/drivers/scsi/mpt3sas/mpt3sas.ko
可以看到在3.10内核的时候是mpt2sas驱动,可以在5.x内核中看到,新版的内核已经把mpt2sas升级为mpt3sas2、解决方法一
a、修改/etc/dracut.conf文件
sed -i '/#add_drivers+=""/aadd_drivers+="mpt3sas"' /etc/dracut.conf
b、重新生成initramfs
# cd /boot
# dracut -f /boot/initramfs-5.9.6-1.el7.elrepo.x86_64.img 5.9.6-1.el7.elrepo.x86_64
c、检查
# lsinitrd -k 5.9.6-1.el7.elrepo.x86_64 | grep mpt[23]sas
drwxr-xr-x 2 root root 0 Nov 9 10:27 usr/lib/modules/5.9.6-1.el7.elrepo.x86_64/kernel/drivers/scsi/mpt3sas
-rwxr--r-- 1 root root 481320 Nov 9 10:27 usr/lib/modules/5.9.6-1.el7.elrepo.x86_64/kernel/drivers/scsi/mpt3sas/mpt3sas.ko
d、重启
# reboot3、解决方法二
强制加载驱动
# dracut --force --add-drivers mpt3sas --kver=5.9.6
# reboot
