Thread: psqlODBC 09.05.0300 (x86) - problem with SQL_QUERY_TIMEOUT
Hello, it looks like SQL_QUERY_TIMEOUT is on by default - causing dead connections with long running statements: "SQLSTATE = 00000 Native Error = 26 Could not send Query(connection dead)" I tested with 9.3.400 (without problems) as well. How can I disable it? Adding "statement_timeout = 0" to the "Connect Settings" field on Advance Options Page 2? Could you tell me, please? Thank you very much! Best regards, Peter
> From: pgsql-odbc-owner@postgresql.org > [mailto:pgsql-odbc-owner@postgresql.org] On Behalf Of Jan-Peter Seifert > it looks like SQL_QUERY_TIMEOUT is on by default - causing dead connections > with long running statements: > > "SQLSTATE = 00000 Native Error = 26 Could not send Query(connection dead)" No, SQL_QUERY_TIMEOUT is off by default. I've just confirmed that "SET statement_timeout" is not issued when I executed "SELECT pg_sleep(60);" with isql command ofthe unixODBC. In addition, SQL_QUERY_TIMEOUT just terminates the query, retaining the database connection. Your messageseems to show that the connection was terminated. Could you show the mylog output? Regards Takayuki Tsunakawa
Hello Tsunakawa-san, thank you very much for your reply! > No, SQL_QUERY_TIMEOUT is off by default. How would you set it then - via connect settings? > I've just confirmed that "SET statement_timeout" is not issued when I executed "SELECT pg_sleep(60);" with isql commandof the unixODBC. In addition, SQL_QUERY_TIMEOUT just terminates the query, retaining the database connection. Yourmessage seems to show that the connection was terminated. You are right. It seems that there was a network problem when I tried the update before. It completed using v9.5.300 withoutproblems. > Could you show the mylog output? I have the mylog output. However, there was another problem. When using 9.5.300 all the the check boxes for the logs weredisabled (greyed out) - even when creating a new ODBC DSN. When using 9.3.400 everything is okay. Best regards, Peter
> From: pgsql-odbc-owner@postgresql.org > [mailto:pgsql-odbc-owner@postgresql.org] On Behalf Of Jan-Peter Seifert > > No, SQL_QUERY_TIMEOUT is off by default. > > How would you set it then - via connect settings? psqlODBC follows the ODBC standard; i.e. the function call SQLSetStmtAttr(SQL_QUERY_TIMEOUT, timeout_seconds) does that. psqlODBC also allows setting with SQLSetConnectAttr(). ADO's CommandTimeout uses these functions behind the scenes. > > I've just confirmed that "SET statement_timeout" is not issued when I > executed "SELECT pg_sleep(60);" with isql command of the unixODBC. In > addition, SQL_QUERY_TIMEOUT just terminates the query, retaining the > database connection. Your message seems to show that the connection was > terminated. > > You are right. It seems that there was a network problem when I tried the > update before. It completed using v9.5.300 without problems. I'm relieved to hear that. > > Could you show the mylog output? > > I have the mylog output. However, there was another problem. When using > 9.5.300 all the the check boxes for the logs were disabled (greyed out) > - even when creating a new ODBC DSN. When using 9.3.400 everything is okay. I've just reproduced the grey-out phenomenon, too. It doesn't happen with at least 09.05.0100. Having a quick look at thediff between 09.05.0100 and 09.05.0300, I couldn't find any suspicious modification. I'd like to ask for help from otherdevelopers. Regards Takayuki Tsunakawa
Hi, Inoue-san, > I have the mylog output. However, there was another problem. When using > 9.5.300 all the the check boxes for the logs were disabled (greyed out) > - even when creating a new ODBC DSN. When using 9.3.400 everything is okay. The attached is a patch to fix this. Could you commit this? Regards Takayuki Tsunakawa
Attachment
Hi, On 2016/07/12 11:55, Tsunakawa, Takayuki wrote: > Hi, Inoue-san, > >> I have the mylog output. However, there was another problem. When using >> 9.5.300 all the the check boxes for the logs were disabled (greyed out) >> - even when creating a new ODBC DSN. When using 9.3.400 everything is okay. > The attached is a patch to fix this. Could you commit this? Hmm, I already committed the fix. regards, Hiroshi Inoue
> From: Inoue, Hiroshi [mailto:h-inoue@dream.email.ne.jp] > Hmm, I already committed the fix. Oh, I should have checked the git log... Thanks. Regards Takayuki Tsunakawa
Hello, > Hmm, I already committed the fix. it's good to know that this bug has been fixed. Thank you very much! Best regards, Peter