[Pgcluster-general] Connection to load balancer refused

Peter Vallero pjvallero at yahoo.com
Fri Aug 31 14:45:55 UTC 2007 

OK, I understand, and yes indeed I did have the cluster dbs on 5434 and pglb on 5433. 

 However, I'm pretty certain that pglb did not die.  I saw 3 instances of the process running, and this was the (end of the) output from the pglb.log (I was running pglb -l -v -n):

Thu Aug 30 21:17:09 2007  pool_process_query():read kind from backend pending data Z len: 5 po: 269
Thu Aug 30 21:17:09 2007  ReadyForQuery(): message length: 5
Thu Aug 30 21:17:09 2007  ReadyForQuery(): transaction state: I
Thu Aug 30 21:17:11 2007  ProcessFrontendResponse():read kind from frontend X(58)
Thu Aug 30 23:31:44 2007  PGRset_status_on_cluster_tbl():host:db01 port:5433 max:300 use:0 status1
Thu Aug 30 23:31:44 2007  PGRset_status_on_cluster_tbl():host:db02 port:5433 max:300 use:0 status1
Thu Aug 30 23:31:44 2007  init_pglb():Child_Tbl size is[464400]

and as I mentioned before, when I connect (using psql from db01) on port 5433, it allows and I see data flowing through the pglb.log.  For instance:

Thu Aug 30 20:33:40 2007  ProcessFrontendResponse():read kind from frontend Q(51)
Thu Aug 30 20:33:40 2007  Query(): select * from person
;
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend N
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend T
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 803 po: 220
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 733 po: 290
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 670 po: 353
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 601 po: 422
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 530 po: 493
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 464 po: 559
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 397 po: 626
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 332 po: 691
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 266 po: 757
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 201 po: 822
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 140 po: 883
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 71 po: 952
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 4 po: 1019
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend D
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data D len: 159 po: 67
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data C len: 93 po: 133
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data N len: 81 po: 145
Thu Aug 30 20:33:40 2007  pool_process_query():read kind from backend pending data Z len: 5 po: 221
Thu Aug 30 20:33:40 2007  ReadyForQuery(): message length: 5
Thu Aug 30 20:33:40 2007  ReadyForQuery(): transaction state: I
Thu Aug 30 20:33:45 2007  ProcessFrontendResponse():read kind from frontend X(58)


It's only when I try to connect to db01:5433 from other machines that the connections are not allowed.

Thanks again,

Peter...

----- Original Message ----
From: Tarhon-Onu Victor <mituc at activemediatech.com>
To: pgcluster-general at pgfoundry.org
Sent: Friday, August 31, 2007 2:57:38 AM
Subject: Re: [Pgcluster-general] Connection to load balancer refused


On Fri, 31 Aug 2007, Peter Vallero wrote:

> Thanks for the reply.  Are you saying that both the cluster db and the 
> pglb on db01 should listen on port 5433?  Is that allowed?

    No, it is not possible for two processes to bind on the same 
port/protocol (like, 5433/tcp).

    You said your configuration is:

    db01
        pglb (listening on port 5433)
        cluster db 1 (listening on port 5434)
        replication server 1
 
    db02
        cluster db 2 (listening on port 5434)
        replication server 2

    Your problem description:
    1. on db01, on port 5433 you have the load balancer;
    2. you can't connect to 5433 on db01.

    What can be the possible problems:
    a) the pglb is configured to bind on another port (like, 5432?);
    or:
    b) pglb isn't started at all (maybe you lauch it but it crashes 
shortly after due to some configuration problems, unavailable resources 
like ports, and so on).

-- 
Victor Tarhon-Onu
Lead Systems Administrator
ActiveMedia Technology
http://www.activemediatech.com
.............................................................................
This communication contains information that is confidential and may also be
privileged. It is for the exclusive use of the intended recipient(s). If you
are not the intended recipient(s) please note that any form of distribution,
copying or use of this communication or the information in it is strictly
prohibited and may be unlawful. If you have received this communication in
error kindly notify the sender immediately and delete it from your system.
_______________________________________________
Pgcluster-general mailing list
Pgcluster-general at pgfoundry.org
http://pgfoundry.org/mailman/listinfo/pgcluster-general

-- 
Victor Tarhon-Onu
Lead Systems Administrator
ActiveMedia Technology
http://www.activemediatech.com
.............................................................................
This communication contains information that is confidential and may also be
privileged. It is for the exclusive use of the intended recipient(s). If you
are not the intended recipient(s) please note that any form of distribution,
copying or use of this communication or the information in it is strictly
prohibited and may be unlawful. If you have received this communication in
error kindly notify the sender immediately and delete it from your system.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://pgfoundry.org/pipermail/pgcluster-general/attachments/20070831/7680061e/attachment.html

More information about the Pgcluster-general mailing list