Обсуждение: Postgres 9.3 beta log

On Sat, Aug 10, 2013 at 06:38:57AM +0000, Arvind Singh wrote:
> as you can see it is for a single record "login88193" it places
>
> PARSER STATISTICS
> PARSE ANALYSIS STATISTICS
> REWRITER STATISTICS
>
> I know it appers because i have enabled log_parser_* options in log
>
> But what i request to know is the meaning of these statistics and how usefull
> is can be for a server load  evaluation

You should look get the getrusage manual pages for the meaning of the
values, and perhaps this for the meaning of these stages of a query:

    http://www.postgresql.org/developer/backend/

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

  + It's impossible for everything to be true. +

On 08/10/2013 04:21 PM, Bruce Momjian wrote:
> On Sat, Aug 10, 2013 at 06:38:57AM +0000, Arvind Singh wrote:
>> as you can see it is for a single record "login88193" it places
>>
>> PARSER STATISTICS
>> PARSE ANALYSIS STATISTICS
>> REWRITER STATISTICS
>>
>> I know it appers because i have enabled log_parser_* options in log
>>
>> But what i request to know is the meaning of these statistics and how usefull
>> is can be for a server load  evaluation
>
> You should look get the getrusage manual pages for the meaning of the
> values, and perhaps this for the meaning of these stages of a query:
>
>     http://www.postgresql.org/developer/backend/

true - but it seems a bit dubious that all of them show basically zero
(or a constant time)



Stefan

Hi,

I have a pair of PG servers, a master and a replica, all read-write queries are handled by the master, read-only ones
areby the replica.   

From time to time the replica itself is too busy, all read-only queries will get  inconsistent results because of
replicationlag,  sometimes it can be longer than 6 minutes.  I am thinking to add multiple replicas to off-load
read-onlyqueries, can you please suggest a way to monitor and failover the read-only query when the replication lag in
areplica is more than 5 minutes? 

regards

On Sat, Aug 10, 2013 at 09:52:48PM +0200, Stefan Kaltenbrunner wrote:
> On 08/10/2013 04:21 PM, Bruce Momjian wrote:
> > On Sat, Aug 10, 2013 at 06:38:57AM +0000, Arvind Singh wrote:
> >> as you can see it is for a single record "login88193" it places
> >>
> >> PARSER STATISTICS
> >> PARSE ANALYSIS STATISTICS
> >> REWRITER STATISTICS
> >>
> >> I know it appers because i have enabled log_parser_* options in log
> >>
> >> But what i request to know is the meaning of these statistics and how usefull
> >> is can be for a server load  evaluation
> >
> > You should look get the getrusage manual pages for the meaning of the
> > values, and perhaps this for the meaning of these stages of a query:
> >
> >     http://www.postgresql.org/developer/backend/
>
> true - but it seems a bit dubious that all of them show basically zero
> (or a constant time)

Yes, in an ideal world, the majority of time should be spent in the
executor.  Those earlier stages are designed to set thing up for the
executor.

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

  + It's impossible for everything to be true. +

Hi,

I found 'LOG:  invalid magic number 0000 in log file 169, segment 77, offset 4325376" from the replica's log:

LOG:  entering standby mode
LOG:  redo starts at A8/BE81B200
LOG:  consistent recovery state reached at A9/4CFFFFF8
LOG:  database system is ready to accept read only connections
LOG:  invalid magic number 0000 in log file 169, segment 77, offset 4325376
LOG:  streaming replication successfully connected to primary

Can you please advise what it means? can it be ignored?

regards

Hi,

I want to archive WAL log files in the replication server as well, the postgresql.conf is added with following new
lines,after restarting PG, the /var/pgsql/data/archive/ is still empty after 1 hour: 


### new lines added today
#
# to enable the replica as Hot Standby
hot_standby = on
#
# to enable WAL archive (coped from master's conf)
wal_level = hot_standby
max_wal_senders = 5
wal_keep_segments = 500
wal_sync_method = fsync
fsync = on
archive_mode = on
archive_command = 'test ! -f /var/pgsql/data/archive/%f && cp %p /var/pgsql/data/archive/%f'
###

ls -la /var/pgsql/data/archive/
total 8
drwx------ 2 postgres root 4096 Aug 11 16:12 ./
drwx------ 3 postgres root 4096 Aug 10 10:59 ../

PG: 9.2.4
OS: Ubuntu 12.04

I am new to Streaming Replication, can you please advise why there is no WAL archive files found in the archive folder?

regards

On Sun, Aug 11, 2013 at 5:51 AM, ascot.moss@gmail.com
<ascot.moss@gmail.com> wrote:
> Hi,
>
> I have a pair of PG servers, a master and a replica, all read-write queries are handled by the master, read-only ones
areby the replica. 
>
> From time to time the replica itself is too busy, all read-only queries will get  inconsistent results because of
replicationlag,  sometimes it can be longer than 6 minutes.  I am thinking to add multiple replicas to off-load
read-onlyqueries, can you please suggest a way to monitor and failover the read-only query when the replication lag in
areplica is more than 5 minutes? 
I assume that you could use pgpool for that. It has some monitoring
features for replication delay and it can do read-only load balancing
among several servers. You also shouldn't need to change your
application.
Have a look at its documentation to make an opinion:
http://pgpool.projects.pgfoundry.org/pgpool-II/doc/pgpool-en.html

Regards,
--
Michael

> On Sun, Aug 11, 2013 at 5:51 AM, ascot.moss@gmail.com
> <ascot.moss@gmail.com> wrote:
>> Hi,
>>
>> I have a pair of PG servers, a master and a replica, all read-write queries are handled by the master, read-only
onesare by the replica. 
>>
>> From time to time the replica itself is too busy, all read-only queries will get  inconsistent results because of
replicationlag,  sometimes it can be longer than 6 minutes.  I am thinking to add multiple replicas to off-load
read-onlyqueries, can you please suggest a way to monitor and failover the read-only query when the replication lag in
areplica is more than 5 minutes? 
> I assume that you could use pgpool for that. It has some monitoring
> features for replication delay and it can do read-only load balancing
> among several servers. You also shouldn't need to change your
> application.
> Have a look at its documentation to make an opinion:
> http://pgpool.projects.pgfoundry.org/pgpool-II/doc/pgpool-en.html

The url above is obsoleted. Please visit:

http://www.pgpool.net/docs/latest/pgpool-en.html
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese: http://www.sraoss.co.jp

Thanks so much.

On 11 Aug 2013, at 9:50 PM, Michael Paquier wrote:

> On Sun, Aug 11, 2013 at 5:51 AM, ascot.moss@gmail.com
> <ascot.moss@gmail.com> wrote:
>> Hi,
>>
>> I have a pair of PG servers, a master and a replica, all read-write queries are handled by the master, read-only
onesare by the replica. 
>>
>> From time to time the replica itself is too busy, all read-only queries will get  inconsistent results because of
replicationlag,  sometimes it can be longer than 6 minutes.  I am thinking to add multiple replicas to off-load
read-onlyqueries, can you please suggest a way to monitor and failover the read-only query when the replication lag in
areplica is more than 5 minutes? 
> I assume that you could use pgpool for that. It has some monitoring
> features for replication delay and it can do read-only load balancing
> among several servers. You also shouldn't need to change your
> application.
> Have a look at its documentation to make an opinion:
> http://pgpool.projects.pgfoundry.org/pgpool-II/doc/pgpool-en.html
>
> Regards,
> --
> Michael