原本我是想KVM支持NAT，在重新编译kernel，当然为了NAT，我得去重新menuconfig，不知为啥手贱，估计将“Networking options”里面的“Unix domain sockets”选项选为了编译为module，于是就遇到了如下的kernel panic。
FATAL: Module scsi_wait_scan not found.
udevadm[2067]: error getting socket: Address family not supported by protocol
udevadm[2073]: error getting socket: Address family not supported by protocol
dracut Warning: No root device "block:/dev/disk/by-uuid/e006b79d-9eb6-4f18-a11c-517448eb074f" found
dracut Warning: Boot has failed. To debug this issue add "rdshell" to the kernel command line.
dracut Warning: Signal caught!
dracut Warning: Boot has failed. To debug this issue add "rdshell" to the kernel command line.
Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000100
Pid: 1, comm: init Not tainted 3.5.0+ #1
Call Trace:
[] ? panic+0xbe/0x1c6
[] ? do_exit+0x366/0x7c0
[] ? do_group_exit+0x74/0x9e
[] ? sys_exit_group+0x12/0x19
[] ? system_call_fastpath+0x16/0x1b
当我遇到这个panic的现象后，其实我也进行了.config文件的对比，我其实也看到了“CONFIG_UNIX”这一项的差异，由于还有其他的不同点，我就没把这个差异放在心上，没重视它。后来才发现，“CONFIG_UNIX”是为了配置“Unix domain sockets”的，“CONFIG_UNIX”一般都是应该配置为y（编译进kernel），如果只有在少数（如嵌入式）系统中才配置为m（作为module，名为”unix”的module）。即是没有连接到任何网络，有很多常用的程序都使用“unix domain socket”进行通信的，如X-Window、syslog等程序，估计还有我本例遇到的udevadm（udevd）等程序，都是依赖于“CONFIG_UNIX=y”这个配置的。
另外，看了一下配置kernel时，对CONIFG_UNIX选项的说明，如下，看了就都明白了吧。
CONFIG_UNIX: x
x x
x If you say Y here, you will include support for Unix domain sockets; x
x sockets are the standard Unix mechanism for establishing and x
x accessing network connections. Many commonly used programs such as x
x the X Window system and syslog use these sockets even if your x
x machine is not connected to any network. Unless you are working on x
x an embedded system or something similar, you therefore definitely x
x want to say Y here. x
x x
x To compile this driver as a module, choose M here: the module will be x
x called unix. Note that several important services won't work x
x correctly if you say M here and then neglect to load the module. x
x x
x Say Y unless you know what you are doing. x
x x
x Symbol: UNIX [=y] x
x Type : tristate x
x Prompt: Unix domain sockets x
x Defined at net/unix/Kconfig:5 x
x Depends on: NET [=y] x
x Location: x
x -> Networking support (NET [=y]) x
x -> Networking options x
x
了解更多关于Unix Domain Socket的内容，可以参考如下链接：
http://beej.us/guide/bgipc/output/html/multipage/unixsock.html
【后记1】2012.08
关于“FATAL: Module scsi_wait_scan not found.”这个错误，应该在kernel config中配置“CONFIG_SCSI_SCAN_ASYNC=y”，如果有“CONFIG_SCSI_WAIT_SCAN”则也应该配置“CONFIG_SCSI_WAIT_SCAN=m”。加这个后记是因为，我在8月份遇到一次kernel panic是和“FATAL: Module scsi_wait_scan not found.”相关的。
【后记2】2012.10.30
我在启动KVM guest（guest中使用Linux 3.6 kernel）再次遇到一直打印“FATAL: Module scsi_wait_scan not found.”的错误信息。后来查了一下，Linux 3.6中已经没有CONFIG_SCSI_WAIT_SCAN配置项可选了，后来发现guest的grub中的root设备写法的问题（这时不能写成/dev/sda1）。
出现问题的grub中kernel行如下：
kernel (hd0,0)/boot/vmlinuz-3.6.0 ro root=/dev/sda1
修改为UUID的方式去找到root设备，就没有“FATAL: Module scsi_wait_scan not found.”错误，可以正常启动了，如下：
kernel /boot/vmlinuz-3.6.0 ro ro root=UUID=7391683c-12c6-405a-b2e0-ee05c678772c
One Comment