On Aug 25, 2013, at 2:04 AM, Dzmitry wrote:
> Ok,thank you, it's a good point. Need to review & make fixes with what I
> have.
Yes, as suggested I would try just reducing first.
Where connection pooling could help is if your load across all those rails apps
is not spread out evenly, but varie
Ok,thank you, it's a good point. Need to review & make fixes with what I
have.
Thanks,
Dzmitry
On 8/24/13 6:14 PM, "Scott Ribe" wrote:
>On Aug 19, 2013, at 9:55 AM, Dzmitry wrote:
>
>> No, I am not using pgbouncer, I am using pgpool.
>>
>> Total - I have 440 connections to postgres(I h
On Aug 19, 2013, at 9:55 AM, Dzmitry wrote:
> No, I am not using pgbouncer, I am using pgpool.
>
> Total - I have 440 connections to postgres(I have rails application
> running on some servers - each application setup 60 connections to DB and
> keep if forever(until will not be killed), also I h
No, I am not using pgbouncer, I am using pgpool.
Total - I have 440 connections to postgres(I have rails application
running on some servers - each application setup 60 connections to DB and
keep if forever(until will not be killed), also I have some machines that
do background processing, that ke
On Aug 19, 2013, at 9:22 AM, Dzmitry wrote:
> I am using pgpool to balance load to slave servers.
So, to be clear, you're not using pgbouncer after all?
--
Scott Ribe
scott_r...@elevated-dev.com
http://www.elevated-dev.com/
(303) 722-0567 voice
--
Sent via pgsql-admin mailing list (pgsq
On Aug 19, 2013, at 9:21 AM, Dzmitry wrote:
> I have 2 machines, each machine run 1 process, with 20 threads in each -
> total 40 threads. Every thread connected to postgres and insert data. I am
> using technology, that keep connection opens. It means every thread open 1
> connection when it sta
Sorry for confusion.
I have 2 machines, each machine run 1 process, with 20 threads in each -
total 40 threads. Every thread connected to postgres and insert data. I am
using technology, that keep connection opens. It means every thread open 1
connection when it starts, and will closed it - only wh
On Aug 19, 2013, at 9:07 AM, Dzmitry wrote:
> Yes- I need so many threads...
This is not at all clear from your answer. At most, 1 thread per logical core
can execute at any one time. And while many can be waiting on IO instead of
executing, there is still a rather low limit to how many thread
Yes- I need so many threads, every night I need load jobs from xml to DB.
I need do it as fast as I can, currently it take 4h to load all of
them(around 100 jobs).
CPU/IO wait/percent is about 25%. Do you know how can I check other params
related to IO wait ?
Thanks,
Dzmitry
On 8/19/13
I am already using pgpool, is it bad have 550 connections, is it much ?
Thank you guys for all your help.
Thanks,
Dzmitry
On 8/19/13 4:43 PM, "Scott Ribe" wrote:
>On Aug 19, 2013, at 7:23 AM, Stéphane Schildknecht
> wrote:
>
>> As Laurenz said, you should have a look at documentation.
>>
Am 19.08.2013 10:07, schrieb Dzmitry:
> I am doing a lot of writes to DB in 40 different threads – so every
> thread check if record exists – if not => insert record, if exists =>
> update record.
> During this update, my disk IO almost always – 100% and sometimes it
> crash my DB with following m
On Aug 19, 2013, at 7:23 AM, Stéphane Schildknecht
wrote:
> As Laurenz said, you should have a look at documentation.
>
> It explains how you can lower the risk OOMKiller kills your PostgreSQL
> processes.
> 1. You can set vm.overcommit_memory to 2 in sysctl.conf
> 2. You can adjust the value
Le 19/08/2013 15:06, Dzmitry a écrit :
Thank you guys ! Found in logs(db-slave1 is my replica that use streaming
replication):
Aug 18 15:49:38 db-slave1 kernel: [25094456.525703] postgres invoked
oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0
Aug 18 15:49:38 db-slave1 kernel:
Thank you guys ! Found in logs(db-slave1 is my replica that use streaming
replication):
Aug 18 15:49:38 db-slave1 kernel: [25094456.525703] postgres invoked
oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0
Aug 18 15:49:38 db-slave1 kernel: [25094456.525708] postgres cpuset=/
mems
>> Dzmitry wrote:
>>> On 8/19/13 11:36 AM, "Stéphane Schildknecht"
>>> wrote:
Le 19/08/2013 10:07, Dzmitry a écrit :
> I have postgres server running on ubuntu 12,Intel Xeon 8 CPUs 29 GB
>RAM.
> With following settings:
> max_connections = 550
> shared_buffers = 12GB
>
Hi !
Since Maverick Ubuntu developers disabled logging to /var/log/messages
by default.
You should check /var/log/syslog instead.
--
Mael
2013/8/19 Dzmitry :
> Do you mean postgres log file(in postgres.conf)
>
> log_filename = 'postgresql-%Y-%m-%d_%H%M%S.log'
> log_min_messages = warning
>
> Or
Do you mean postgres log file(in postgres.conf)
log_filename = 'postgresql-%Y-%m-%d_%H%M%S.log'
log_min_messages = warning
Or /var/log/messages ? Because I haven't this file :(
Thanks,
Dzmitry
On 8/19/13 12:26 PM, "Albe Laurenz" wrote:
>Dzmitry wrote:
>> On 8/19/13 11:36 AM, "Stéphane
Dzmitry wrote:
> On 8/19/13 11:36 AM, "Stéphane Schildknecht"
> wrote:
>> Le 19/08/2013 10:07, Dzmitry a écrit :
>>> I have postgres server running on ubuntu 12,Intel Xeon 8 CPUs 29 GB RAM.
>>> With following settings:
>>> max_connections = 550
>>> shared_buffers = 12GB
>>> temp_buffers = 8MB
>>
I don't think it's the case. I am using newrelic for monitoring my DB
servers(I have one master and 2 slaves - all use the same configuration) -
memory is not going above 12.5GB, so I have a good reserve, also I don't
see any swapping there :(
Thanks,
Dzmitry
On 8/19/13 11:36 AM, "Stéphane
Le 19/08/2013 10:07, Dzmitry a écrit :
Hey folks,
I have postgres server running on ubuntu 12,Intel Xeon 8 CPUs 29 GB RAM.
With following settings:
max_connections = 550
shared_buffers = 12GB
temp_buffers = 8MB
max_prepared_transactions = 0
work_mem = 50MB
maintenance_work_mem = 1GB
fsync = on
w
Hey folks,
I have postgres server running on ubuntu 12, Intel Xeon 8 CPUs 29 GB RAM.
With following settings:
max_connections = 550
shared_buffers = 12GB
temp_buffers = 8MB
max_prepared_transactions = 0
work_mem = 50MB
maintenance_work_mem = 1GB
fsync = on
wal_buffers = 16MB
commit_delay = 50
c
21 matches
Mail list logo
