At 2011-01-12 23:22:40,"Bill Moran" <wmo...@potentialtech.com> wrote:

>In response to zab08 <za...@126.com>:
>
>> we have e test db server, we use jdbc to contect.
>> but after a whie, The test db server can not accept connect.
>> 
>> 
>> useps aux | grep postgrescommand :
>> postgres 16904  0.0  0.0  46036  3948 ?        S    17:03   0:00 
>> /usr/local/pgsql/bin/postgres -D /usr/local/pgsql/data
>> postgres 16906  0.0  0.0  46144  2192 ?        Ss   17:03   0:00 postgres: 
>> writer process                              
>> postgres 16907  0.0  0.0  46036  1088 ?        Ss   17:03   0:00 postgres: 
>> wal writer process                          
>> postgres 16908  0.0  0.0  46840  2072 ?        Ss   17:03   0:00 postgres: 
>> autovacuum launcher process                 
>> postgres 16909  0.0  0.0  14248  1012 ?        Ss   17:03   0:00 postgres: 
>> stats collector process                     
>> postgres 16910  0.0  0.0  46972  2992 ?        Ss   17:03   0:00 postgres: 
>> postgres edu 172.29.0.40(40608) idle        
>> postgres 16911  0.0  0.0  46708  2656 ?        Ss   17:03   0:00 postgres: 
>> postgres edu 172.29.0.40(40609) idle        
>> postgres 16912  0.0  0.0  46708  2628 ?        Ss   17:03   0:00 postgres: 
>> postgres edu 172.29.0.40(40610) idle        
>> postgres 16913  0.0  0.0  46708  2628 ?        Ss   17:03   0:00 postgres: 
>> postgres edu 172.29.0.40(40611) idle        
>> postgres 16914  0.0  0.0  46708  2632 ?        Ss   17:03   0:00 postgres: 
>> postgres edu 172.29.0.40(40612) idle        
>> postgres 16915  0.0  0.0  46708  2636 ?        Ss   17:03   0:00 postgres: 
>> postgres edu 172.29.0.40(40613) idle        
>> postgres 16916  0.0  0.0  46708  2632 ?        Ss   17:03   0:00 postgres: 
>> postgres edu 172.29.0.40(40614) idle        
>> postgres 16917  0.0  0.1  47088  4136 ?        Ss   17:03   0:00 postgres: 
>> postgres edu 172.29.0.40(40615) idle        
>> postgres 16918  0.0  0.1  47060  4236 ?        Ss   17:03   0:00 postgres: 
>> postgres edu 172.29.0.40(40616) idle        
>> postgres 16919  0.0  0.1  47052  4136 ?        Ss   17:03   0:02 postgres: 
>> postgres edu 172.29.0.40(40617) idle        
>> postgres 21619  0.0  0.0  46708  2636 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38803) idle        
>> postgres 21620  0.0  0.0  46708  2636 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38804) idle        
>> postgres 21621  0.0  0.0  46708  2640 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38805) idle        
>> postgres 21622  0.0  0.0  46708  2644 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38806) idle        
>> postgres 21623  0.0  0.0  46708  2644 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38807) idle        
>> postgres 21624  0.0  0.0  46708  2636 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38808) idle        
>> postgres 21625  0.0  0.0  46708  2640 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38809) idle        
>> postgres 21626  0.0  0.0  46708  2644 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38810) idle        
>> postgres 21627  0.0  0.0  46708  2640 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38811) idle        
>> postgres 21628  0.0  0.1  47104  4836 ?        Ss   19:43   0:00 postgres: 
>> postgres edu 172.29.0.40(38812) INSERT      
>> ...
>> 
>> 
>>  kill -9 pidcommand to kill all of them in the past? and the test db server 
>> is shut down.
>
>Are you insane?  If your car doesn't start do you throw hand grenades at
>it?  What you're doing is an excellent way to corrupt the data in your
>database.
>
>Use pg_ctl to stop/start/restart the database -- or better, use whatever
>tool is provided by your distribution/operating system.
>
>> how to analyze the reason of these? and how to solve?
>
>Have you ever considered looking in the PostgreSQL log files?  If I had
>to make a guess, I would suspect that you've got the connection limit too
>low for the number of connections required by whatever you're doing and
>a simple bump of the config parameter will fix the problem.  More serious
>issues are possible, such as client programs going idle and never releasing
>the connection, but I wouldn't assume such problems until you have
>evidence.
>
>-- 
>Bill Moran
>http://www.potentialtech.com
>http://people.collaborativefusion.com/~wmoran/
>
>-- 
>Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
>To make changes to your subscription:
>http://www.postgresql.org/mailpref/pgsql-general



client program shut down, but the conetion isnt' close on the test server.
we have 6 clients and , 10 contections each. The test db sever 's max 
conections is 100.  pg_ctl to stop/start/restart the database can't excute 
because of the
 db the shut down. 

Reply via email to