Openstack对接Ceph时的错误

本次 Openstack 版本为N版ceph版本为J版在对接volume和glance时非常轻松不过在对接nova的时候坑了一下午。

ERROR nova.compute.manager [instance: b6b53548-2d12-4019-a189-1d006455dfa3] 2017-08-07T14:09:55.209454Z qemu-system-x86_64: -drive file=rbd:vms/b6b53548-2d12-4019-a189-1d006455dfa3_disk:id=cinder:auth_supported=cephx\;none:mon_host=192.168.0.151\:6789,file.password-secret=virtio-disk0-secret0,format=raw,if=none,id=drive-virtio-disk0,cache=none: Unknown protocol 'rbd'

意思是说不支持rbd协议 于是乎查看自己的qemu。我的qemu是编译安装的详情见 http://home.51cto.com/space/6890594

[root@compute01 ~]# qemu-img --help | grep rbd
[root@compute01 ~]#

注此时我已经装了ceph相关的包现在已经有了rbd的命令内核已经也已经有了rbd现在是qemu不支持rbd协议接下来继续证明这个事情

[root@compute01 ~]# rados lspools
glance
volumes
vms
rbd
images
[root@compute01 ~]# rbd ls volumes
volume-4a898a0e-40db-4846-b944-079891598ae4
[root@compute01 ~]#
[root@compute01 ~]# qemu-img create -f raw rbd:volumes/test 5G
qemu-img: rbd:volumes/test: Unknown protocol 'rbd'
[root@compute01 ~]#

可以看到直接让qemu使用ceph都不行报不支持rbd协议此时 我回想自己编译qemu时是默认选项查了好多文档默认是不支持rbd的所以要重新编译此次开启rbd

[root@compute01 qemu-2.9.0]# ./configure --enable-rbd
No C++ compiler available; disabling C++ specific optional code

ERROR: User requested feature rados block device
configure was not able to find it.
Install librbd/ceph devel

报No C++我们来安装包

[root@compute01 qemu-2.9.0]# yum install gcc  gcc-c++
Installed:
gcc-c++.x86_64 0:4.8.5-11.el7

Dependency Installed:
libstdc++-devel.x86_64 0:4.8.5-11.el7

Complete!

再次编译

[root@compute01 qemu-2.9.0]# ./configure --enable-rbd

ERROR: User requested feature rados block device
configure was not able to find it.
Install librbd/ceph devel

报找不到ceph 的librdb我们继续装包

[root@compute01 qemu-2.9.0]# yum install ceph-devel
Loaded plugins: fastestmirror, priorities
Loading mirror speeds from cached hostfile
Resolving Dependencies
--> Running transaction check
---> Package ceph-devel-compat.x86_64 1:10.2.9-0.el7 will be installed
Installed:
ceph-devel-compat.x86_64 1:10.2.9-0.el7

Dependency Installed:
libcephfs1-devel.x86_64 1:10.2.9-0.el7                          libcephfs_jni1.x86_64 1:10.2.9-0.el7
libcephfs_jni1-devel.x86_64 1:10.2.9-0.el7                      librados2-devel.x86_64 1:10.2.9-0.el7
libradosstriper1-devel.x86_64 1:10.2.9-0.el7                    librbd1-devel.x86_64 1:10.2.9-0.el7

Complete!

再次编译

[root@compute01 qemu-2.9.0]# ./configure --enable-rbd
Install prefix    /usr/local
BIOS directory    /usr/local/share/qemu
binary directory  /usr/local/bin
library directory /usr/local/lib
module directory  /usr/local/lib/qemu
libexec directory /usr/local/libexec
include directory /usr/local/include
config directory  /usr/local/etc
local state directory   /usr/local/var
Manual directory  /usr/local/share/man
ELF interp prefix /usr/gnemul/qemu-%M

rbd support       yes ###重点来了
NUMA host support no
tcmalloc support  no
jemalloc support  no
avx2 optimization yes
replication support yes

[root@compute01 qemu-2.9.0]# make && make install

大概等了3个小时终于编译完成！

install -c -m 0755 qemu-tilegx  "/usr/local/bin"
strip "/usr/local/bin/qemu-tilegx"
install -d -m 0755 "/usr/local/bin"
install -c -m 0755 qemu-x86_64  "/usr/local/bin"
strip "/usr/local/bin/qemu-x86_64"
[root@compute01 qemu-2.9.0]#

再次查看是否支持rbd

[root@compute01 ~]# qemu-img -h|grep rbd
Supported formats: blkdebug blkreplay blkverify bochs cloop dmg file host_cdrom host_device luks nbd null-aio null-co parallels qcow qcow2 qed quorum raw rbd replication sheepdog vdi vhdx vmdk vpc vvfat
[root@compute01 ~]#

qemu已经支持rbd了！此次任务完成！
参考文档
http://www.voidcn.com/blog/xiangpingli/article/p-6087550.html
http://docs.ceph.com/docs/jewel/rbd/qemu-rbd/