Ceph搭建过程中遇到的各种问题

  人们在选择存储系统的时候，往往会考虑到它的存储能力、容错冗余机制、负载均衡等等。Ceph是一个非常好的分布式存储系统，可以无限制的扩展你的集群，并且它是一个块存储、对象存储和文件系统统一的分布式文件系统。所以很多公司都会选择ceph作为存储系统。

  那么我们在进行ceph快速搭建的时候往往会遇到很多问题，我这边把我所遇到的问题总结了一下，请大家多多指教！

  安装过程

#ceph-deploy install admin-node node1

报错：

(1)         [ceph_deploy][ERROR ]RuntimeError: Failed to execute command: yum -y install epel-release

解决方法：

进入/etc/yum.repos.d中删除epel.repo和epel-testing.repo

(2)         [ceph_deploy][ERROR ]RuntimeError: NoSectionError: No section: 'ceph'

解决方法：

#yum -y remove ceph-release

(3)         [admin-node][WARNIN] Anotherapp is currently holding the yum lock; waiting for it to exit...

解决方式：

#rm -f /var/run/yum.pid

(4)         RuntimeError: Failed to executecommand: yum -y install ceph ceph-radosgw

解决方式：

# rpm -ivh libunwind-1.1-10.el7.x86_64.rpm

(5)         [ceph_deploy][ERROR ]RuntimeError: Failed to execute command: yum -y install yum-plugin-priorities

解决方式：

进入/etc/yum.repos.d中删除epel.repo和epel-testing.repo

(6)         TTY报错

[node1][DEBUG ] connection detected need for sudo

sudo: sorry, you must have a tty to run sudo

[node1][DEBUG ] connected to host: node1

[ceph_deploy][ERROR ] RuntimeError: remoteconnection got closed, ensure ``requiretty`` is disabled for node1

解决方法：

# visudo

#Defaults    requiretty
Defaults:ceph !requiretty

(7)         ** ERROR: error creating emptyobject store in /var/local/osd0: (13) Permission denied

[admin-node][WARNIN]

[admin-node][ERROR ] RuntimeError: command returned non-zero exitstatus: 1

[ceph_deploy][ERROR ] RuntimeError: Failedto execute command: /usr/sbin/ceph-disk -v activate --mark-init systemd --mount/var/local/osd0

解决方法：

在各个节点上给/var/local/osd1/和/var/local/osd1/添加权限

如下：

chmod 777  /var/local/osd0/

chmod 777  /var/local/osd0/*

chmod 777  /var/local/osd1/

chmod 777  /var/local/osd1/*

(8)         连接超时

[root@admin-node my-cluster]#ceph-deploy  osd activatenode1:/var/local/osd1

(9)         osd节点起不来

[root@admin-node my-cluster]# ceph -s

    clusterbeca06a9-c4c3-443d-9d85-fbc6d620c173

     health HEALTH_WARN

            64 pgs degraded

            64 pgs stuckunclean

            64 pgs undersized

     monmap e1: 1 mons at{admin-node=10.0.3.88:6789/0}

            election epoch 3,quorum 0 admin-node

     osdmap e7: 2 osds: 1 up,1 in

            flags sortbitwise

      pgmap v29: 64 pgs, 1pools, 0 bytes data, 0 objects

            6644 MB used,44544 MB / 51188 MB avail

                  64active+undersized+degraded

[root@admin-node my-cluster]# ceph osd tree

ID WEIGHT  TYPE NAME           UP/DOWN REWEIGHT PRIMARY-AFFINITY

-1 0.09760 root default                                         

-2 0.04880     hostadmin-node                                  

 0 0.04880         osd.0            up 1.00000          1.00000

-3 0.04880     host node1                                       

 10.04880         osd.1          down        0          1.00000

解决办法：

1.      关闭防火墙

2.      检查ssh无密钥登录配置

3.      一般情况下，osd节点启动较慢，等待5分钟左右即可。