[Pgpool-general] Possible bug with pgpool-3.0.0 + patches and PQprepare()

[Tatsuo Ishii]

> Hi, this problem showed up in a real appliction , but I've reproduced it
> with a simple test program, attached. Upon executing PQprepare(),
> the test app hangs when using pgpool. If I point the test app at the
> database server directly, it does not hang.
> 
> In the course of investigating this, I captured packet traces from
> the pgpool and non-pgpool cases, but they didn't make much sense
> to me in ethereal (they're different, but I couldn't see an obvious smoking
> gun as to what went wrong). If it would help, I can go back and take
> similar packet captures from this test program and send them along.
> 
> This is the debug output from pgpool  when I run the command:

Thanks. It seems prepared queries are broken with master/slave
mode. Include patches shoud fix the problem. Note that the patches are
gainst 3.0. If you have already applied previous patches, and want to
apply changes incremnetaly, please use following part in the patches.

*** 1660,1665 ****
--- 1663,1675 ----
  		default:
  			if (MAJOR(backend) == PROTO_MAJOR_V3)
  			{
+ 				if (MASTER_SLAVE)
+ 				{
+ 					master_slave_was_enabled = 1;
+ 					MASTER_SLAVE = 0;
+ 					master_slave_dml = 1;
+ 				}
+ 
  				status = SimpleForwardToBackend(fkind, frontend, backend);
  				if (pool_flush(MASTER(backend)))
  					status = POOL_ERROR;
--
Tatsuo Ishii
SRA OSS, Inc. Japan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: diff
Type: application/octet-stream
Size: 1529 bytes
Desc: not available
Url : http://pgfoundry.org/pipermail/pgpool-general/attachments/20060210/84fe3f43/diff.obj