[pgpool-general: 4672] Re: ERROR: unable to read data from frontend

[Yugo Nagata]

Wolf,

Though some "ERROR: unable to read data from frontend" problems are fixed[1][2],
pgpool still emits this error message when it receives EOF from the frontend.
For example, I can reproduce this error by `kill -9` psql which is connecting
to pgpool. This is expected behavior.

[1] http://git.postgresql.org/gitweb/?p=pgpool2.git;a=commit;h=d5cd159001d6616d8b12346f80831db026ad4b7c
[2] http://git.postgresql.org/gitweb/?p=pgpool2.git;a=commit;h=c0ed0aac08631786116169c4ffac7fe19f3667b0

What does your application when the error occurs? If it isn't the
application problem, could you tell me how to reproduce this error
in your environment?

On Mon, 9 May 2016 15:34:24 +0000
Wolf Schwurack <wolf at uen.org> wrote:

> Yugo
> 
> I tried to send the pgpool.log file but it exceeds the size limits Here is a few lines from pgpool.log
> 2016-04-21 14:33:47: pid 19648: ERROR:  unable to read data from frontend
> 2016-04-21 14:33:47: pid 19648: DETAIL:  EOF encountered with frontend
> 2016-04-21 14:33:47: pid 19764: ERROR:  unable to read data from frontend
> 2016-04-21 14:33:47: pid 19764: DETAIL:  EOF encountered with frontend
> 
> 
> We are using simple protocol and memory query cache is set to off.
> 
> Wolf
> 
> 
> -----Original Message-----
> From: Yugo Nagata [mailto:nagata at sraoss.co.jp] 
> Sent: Sunday, May 08, 2016 9:23 PM
> To: Wolf Schwurack <wolf at uen.org>
> Cc: 'pgpool-general at pgpool.net' <pgpool-general at pgpool.net>
> Subject: Re: [pgpool-general: 4662] ERROR: unable to read data from frontend
> Importance: High
> 
> On Thu, 5 May 2016 13:17:05 +0000
> Wolf Schwurack <wolf at uen.org> wrote:
> 
> > We been getting this error "ERROR: unable to read data from frontend" I did a search on the error it shows that it was bug in version 3.4.0 but that it's been fixed.
> > 
> > We are on version pgpool II version 3.4.4 Should the fix be in this version? When the error happens we are getting TIME_WAIT on the application server. We had up to 170 TIME_WAIT's in a minute's time.
> > 
> > We have two pgpool servers and two PostgreSQL servers running streaming replication
> > 
> > Servers: Ubuntu 14.04.4
> > PostgreSQL: 9.5.2
> > Pgpool: 3.4.4
> 
> Do you use extended protocol (JDBC) or simple protocol? 
> And, do you use in-memory query cache?
> 
> If possible, could you please provide pgpool log file.
> 
> > 
> > 
> > Wolfgang Schwurack
> > Database/System Administrator
> > Utah Education Network
> > 801-587-9444
> > wolf at uen.org
> > 
> > 
> > 
> 
> 
> -- 
> Yugo Nagata <nagata at sraoss.co.jp>


-- 
Yugo Nagata <nagata at sraoss.co.jp>