Thread: UseDeclareFetch=1, Fetch=100, UseServerSidePrepare=1 causes Windows client to intermittently hang
UseDeclareFetch=1, Fetch=100, UseServerSidePrepare=1 causes Windows client to intermittently hang
From
ljwilson
Date:
This weekend I tested at a client site the psqlODBC Windows drivers as of 2013/05/21 (what will eventually be 09.02.0100) from the github source. When printing invoices, intermittently a client (I could reproduce on all three of the test machines) would hang (freeze--must kill with taskmanager). The client might have to print just one, or 13, or as high as 61 individual jobs before the hang would occur. The fix is to set the UseServerSidePrepare back to 0--then no hangs can be reproduced. Enabling mylog makes it much harder to reproduce the hang, but I have been able to produce 4 mylog files using my own test machine that log up to the hang. Here is the end of one of the mylogs when the hang happened: [2228-302.275][SQLFreeStmt][2228-302.275]PGAPI_FreeStmt: entering...hstmt=01B3AA70, fOption=0 [2228-302.275]SC_recycle_statement: self= 01B3AA70 [2228-302.276]QResult: enter DESTRUCTOR [2228-302.276]QResult: in QR_close_result [2228-302.276]CC_cursor_count: self=00B7A6A0, num_stmts=64 [2228-302.276]CC_cursor_count: returning 13 [2228-302.276]CC_send_query: conn=00B7A6A0, query='close "SQL_CUR01B3AA70"' [2228-302.277]in QR_Constructor [2228-302.277]exit QR_Constructor [2228-302.277]SOCK_get_id has to eat 1578 bytes [2228-302.277]SC_set_Result(1c961c0, 1b68380)[2228-302.277]QResult: enter DESTRUCTOR [980-302.277]read 6, global_socket_buffersize=4096 [980-302.278]read -1, global_socket_buffersize=4096 [980-302.278]Lasterror=10035 Thanks...jack -- View this message in context: http://postgresql.1045698.n5.nabble.com/UseDeclareFetch-1-Fetch-100-UseServerSidePrepare-1-causes-Windows-client-to-intermittently-hang-tp5757049.html Sent from the PostgreSQL - odbc mailing list archive at Nabble.com.
Re: UseDeclareFetch=1, Fetch=100, UseServerSidePrepare=1 causes Windows client to intermittently hang
From
ljwilson
Date:
To bring some closure: Hiroshi Inoue-san created a patch based on the mylogs. It passed my in-house tests, and today I tested it at the same client site where I originally discovered the issue, and no problems. ...jack -- View this message in context: http://postgresql.1045698.n5.nabble.com/UseDeclareFetch-1-Fetch-100-UseServerSidePrepare-1-causes-Windows-client-to-intermittently-hang-tp5757049p5757710.html Sent from the PostgreSQL - odbc mailing list archive at Nabble.com.
Re: UseDeclareFetch=1, Fetch=100, UseServerSidePrepare=1 causes Windows client to intermittently hang
From
Joshua Berry
Date:
Hi Jack,
On Mon, May 27, 2013 at 5:38 PM, ljwilson <ljwilson@digitalav.com> wrote:
This weekend I tested at a client site the psqlODBC Windows drivers as of
2013/05/21 (what will eventually be 09.02.0100) from the github source. When
printing invoices, intermittently a client (I could reproduce on all three
of the test machines) would hang (freeze--must kill with taskmanager). The
client might have to print just one, or 13, or as high as 61 individual jobs
before the hang would occur. The fix is to set the UseServerSidePrepare back
to 0--then no hangs can be reproduced. Enabling mylog makes it much harder
to reproduce the hang, but I have been able to produce 4 mylog files using
my own test machine that log up to the hang.
When the client system would hang, would it cause the PG backend to crash and restart? Also, were other clients concurrently connected to the same backend affected as well?
I'm hoping that the issue that you're having is related to the issue I've been having such that a single patch will cure both!
Kind Regards,
-Joshua
Re: UseDeclareFetch=1, Fetch=100, UseServerSidePrepare=1 causes Windows client to intermittently hang
From
ljwilson
Date:
Joshua, > When the client system would hang, would it cause the PG backend to crash > and restart? No, the PG backend was unaffected. > Also, were other clients concurrently connected to the same > backend affected as well? No, other clients were not impacted. > I'm hoping that the issue that you're having is related to the issue I've > been having such that a single patch will cure both! > Kind Regards, > -Joshua Me too! ...jack -- View this message in context: http://postgresql.1045698.n5.nabble.com/UseDeclareFetch-1-Fetch-100-UseServerSidePrepare-1-causes-Windows-client-to-intermittently-hang-tp5757049p5757795.html Sent from the PostgreSQL - odbc mailing list archive at Nabble.com.