最近在Linux上装个GP测试，装的比较简单，单节点，问题是在安装好了GP后第二天使用gpstart启动GP时，启动失败，启动日志

20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[WARNING]:-Failed segment starts                                                = 1   <<<<<<<<
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-   Skipped segment starts (segments are marked down in configuration)   = 0
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-----------------------------------------------------
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-Successfully started 0 of 1 segment instances <<<<<<<<
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-----------------------------------------------------
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[WARNING]:-Segment instance startup failures reported
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[WARNING]:-Failed start 1 of 1 segment instances <<<<<<<<
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[WARNING]:-Review /home/gpadmin/gpAdminLogs/gpstart_20170606.log
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-----------------------------------------------------
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-Commencing parallel segment instance shutdown, please wait...
20170606:23:12:30:018059 gpstart:quickstart:gpadmin-[ERROR]:-gpstart error: Do not have enough valid segments to start the array.

错误显示：

Do not have enough valid segments to start the array.

上网搜索了一波，说是shared_buffers设置的值大了，一般小于1000，或者有的答案是小于250，但是我很奇怪，我从来没有改过这个值，去/gpmaster/gpseg-1/postgresql.conf中查看shared_buffers的值为125Mb。

#------------------------------------------------------------------------------
# RESOURCE USAGE (except WAL)
#------------------------------------------------------------------------------

# - Memory -

shared_buffers = 125MB                  # inserted by initdb
#shared_buffers = 128MB                 # min 128kB or max_connections*16kB
                                        # (change requires restart)
#temp_buffers = 8MB                     # min 800kB
max_prepared_transactions = 250         # can be 0 or more
                                        # (change requires restart)
# Note:  Increasing max_prepared_transactions costs ~600 bytes of shared memory
# per transaction slot, plus lock space (see max_locks_per_transaction).
#work_mem = 32MB                        # min 64kB
#maintenance_work_mem = 64MB            # min 1MB
#max_stack_depth = 2MB                  # min 100kB

所以这个应该不是问题的原因，没办法只能去查看启动日志，

[gpadmin@quickstart gpseg-1]$ cd /home/gpadmin/gpAdminLogs/
[gpadmin@quickstart gpAdminLogs]$ ls
gpconfig_20170518.log                          gpstart_20170515.log
gpinitsystem_20170514.log                      gpstart_20170518.log
gpinitsystem_20170606.log                      gpstart_20170519.log
gprecoverseg_20170518.log                      gpstart_20170523.log
gpsegstart.py_quickstart:gpadmin_20170514.log  gpstart_20170606.log
gpsegstart.py_quickstart:gpadmin_20170606.log  gpstate_20170518.log
gpsegstop.py_quickstart:gpadmin_20170514.log   gpstop_20170514.log
gpsegstop.py_quickstart:gpadmin_20170515.log   gpstop_20170515.log
gpsegstop.py_quickstart:gpadmin_20170518.log   gpstop_20170518.log
gpsegstop.py_quickstart:gpadmin_20170606.log   gpstop_20170606.log
gpstart_20170514.log
[gpadmin@quickstart gpAdminLogs]$ more gpstart_20170606.log 

查看启动日志，发现在执行命令的时候找不着gpServer地址，

20170606:23:12:16:018059 gpstart:quickstart:gpadmin-[INFO]:-   quickstart.cloudera   /home/gpadmin/primary/gpseg0   40000
20170606:23:12:23:018059 gpstart:quickstart:gpadmin-[INFO]:-Commencing parallel segment instance startup, please wait...
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-Process results...
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[ERROR]:-No segment started for content: 0.
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-dumping success segments: []
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-----------------------------------------------------
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-DBID:2  FAILED  host:'quickstart.cloudera' datadir:'/home/gpadmin/primary/gpseg0' with reason:'cmd had rc=255 completed=True halted=False
  stdout=''
  stderr='ssh: Could not resolve hostname gpServer: Name or service not known^M
''
20170606:23:12:29:018059 gpstart:quickstart:gpadmin-[INFO]:-----------------------------------------------------

很奇怪，我昨天在安装的时候在hosts里面添加了这个值，结果去查看hosts中，真没有这个值，也就是说在机器重启之后这个值就没有了，现在还在纠结是为什么，在hosts里面加上gpServer就可以了。

所在在看日志的时候应该要看出问题的关键原因，走弯路是最骚的。