CDH报错cloudera-scm-server dead but pid file exists

CDH CM界面出现了无法访问的情现象，查看cloudera-scm-server状态

# service cloudera-scm-server status
cloudera-scm-server dead but pid file exists

提示cloudera-scm-server dead
# service cloudera-scm-server stop
# service cloudera-scm-server status
cloudera-scm-server is stopped

删除cloudera-scm-server.pid
# rm /var/run/cloudera-scm-server.pid


service cloudera-scm-server-db start
不能正常启动cloudera-scm-server-db
waiting for server to start ...............cloud not start server
# service cloudera-scm-server-db status
pg_ctl: no server running

cloudera-scm-server-db没有运行的情况下，如果存在postmaster.pid就删除了
#cd /var/lib/cloudera-scm-server-db/data
# rm postmaster.pid

查看db.log
cat /var/log/cloudera-scm-server/db.log
报错：could not write to log file:No space left on device
cloud not create shared memory segment：Invalid argument。


#cat /var/lib/cloudera-scm-server-db/data/postgresql.conf|grep shared_buffers
shared_buffers=308493

#cat /etc/sysctl.conf|grep kernel.shmmax
kernel.shmmax = 68719476736
kernel.shmmax = 4294967296

注释

vi /etc/sysctl.conf

kernel.shmmax = 4294967296

保留大的设置。

重新启动
service cloudera-scm-server-db start
service cloudera-scm-server-db status
postgresql启动正常了。


# service cloudera-scm-server-db start
DB initialization done.
waiting for server to start.... done
server started

# service cloudera-scm-server start
Starting cloudera-scm-server:                              [  OK  ]

问题解决了，CM可以正常访问了，问题应该是postgresql使用的共享内存超过了系统设置的kernel.shmmax，需要改大些。