Re: Large delay in pgsql-jdbc SELECT
От | Barry Lind |
---|---|
Тема | Re: Large delay in pgsql-jdbc SELECT |
Дата | |
Msg-id | 3D404978.4050208@xythos.com обсуждение исходный текст |
Ответ на | Large delay in pgsql-jdbc SELECT ("Jackson, Scott M" <Scott.Jackson@pnl.gov>) |
Список | pgsql-jdbc |
Turn on query logging on the server and run the exact same query that jdbc is sending in psql. I suspect you will see the same issue in psql as well. The info you have included below indicates that the server is just taking a long time to run the query. --Barry Jackson, Scott M wrote: >Hello, > >I am hoping somebody might have an idea about this or a suggestion for >troubleshooting... > >I am using jdbc7.1-1.2.jar with jdk1.3.1_01 on an i386 box with RedHat 7.1. > >When running a SELECT query that resulted in 10 rows of 100 columns each, I >could average a 1 second response from my application. >However when I ran a similar SELECT that resulted in 100 rows of 100 columns >each (of exactly the same format), I experienced a very definitive hang of >about 20-25 seconds. > >When I issued the same query within the psql prompt, the query results >returned immediately. > >My application is hanging at the db.createStatement().executeQuery("SELECT >..."); statement >I turned on level 4 debugging and straced -f my postmaster and saw that it >was hanging in a recv statement in file descriptor 5 which is the >/usr/local/pgsql/data/global/pg_control file. > >Debug and strace output at the hang is: > >...... >[pid 24706] 11:40:25 brk(0x8237000) = 0x8237000 >[pid 24706] 11:40:25 write(2, "DEBUG: CommitTransactionCommand"..., >33DEBUG: CommitTransactionCommand >) = 33 >[pid 24706] 11:40:25 brk(0x8213000) = 0x8213000 >[pid 24706] 11:40:25 send(5, >"\0051\0\0\0\0051\0\0\0\0051\0\0\0\0051\0\0\0\0051\0\0\0"..., 1327, 0) = >1327 >[pid 24706] 11:40:25 recv(5, > >(It hangs here for 20 seconds) > >"Qcommit\0", 8192, 0) = 8 >[pid 24827] 11:43:09 write(2, "DEBUG: StartTransactionCommand\n", 32DEBUG: >StartTransactionCommand >) = 32 >[pid 24827] 11:43:09 write(2, "DEBUG: query: commit\n", 22DEBUG: query: >commit) = 22 >... > > >Top appears to reveal no memory problems, I have all my limits set to >unlimited. I see a very high level of CPU activity during this period. > >Does anyone have any thoughts or recommendations? > >Thanks for any help you can give me! > >Scott Jackson > >---------------------------(end of broadcast)--------------------------- >TIP 6: Have you searched our list archives? > >http://archives.postgresql.org > > >
В списке pgsql-jdbc по дате отправления: