Postgres Pro
Downloads
Home   >   mailing lists

Thread: [GENERAL] Begginers question

[GENERAL] Begginers question

From
Alex Samad
Date:
Hi

So I have been playing with an streaming  cluster. I have the replication working I believe.

But whilst attempting to do an import of my original DB, I filled up my disk pg_xlog directory.

Strangley I have tried this before and not filled this up.

so 
1) why did it fill up this time and not previously
I add this
archive_command = '/bin/true'
wal_keep_segments = 1000 # <<< I'm guessing its this

2) how do I fix up, can I just remove the files from the pg_xlog directory 

A

Re: [GENERAL] Begginers question

From
Michael Paquier
Date:
On Wed, Aug 16, 2017 at 2:32 PM, Alex Samad <alex@samad.com.au> wrote:
> 1) why did it fill up this time and not previously
> I add this
> archive_command = '/bin/true'
> wal_keep_segments = 1000 # <<< I'm guessing its this
>
> 2) how do I fix up, can I just remove the files from the pg_xlog directory

Don't do that. those files are managed by Postgres so you may finish
with a corrupted cluster. Instead you should lower the value of
wal_keep_segments, reload the server parameters, and then enforce two
checkpoints to force WAL segments to be recycled. Note that this
depends also on the values of checkpoint_segments
(max_wal_size/min_wal_size in Postgres 9.5 and onwards).
--
Michael

Re: [GENERAL] Begginers question

From
Alex Samad
Date:


On 16 August 2017 at 16:16, Michael Paquier <michael.paquier@gmail.com> wrote:
On Wed, Aug 16, 2017 at 2:32 PM, Alex Samad <alex@samad.com.au> wrote:
> 1) why did it fill up this time and not previously
> I add this
> archive_command = '/bin/true'
> wal_keep_segments = 1000 # <<< I'm guessing its this
>
> 2) how do I fix up, can I just remove the files from the pg_xlog directory

Don't do that. those files are managed by Postgres so you may finish
with a corrupted cluster. Instead you should lower the value of

Too late, its okay its a learning experience.
 
wal_keep_segments, reload the server parameters, and then enforce two
checkpoints to force WAL segments to be recycled. Note that this

how do I force check points
 
depends also on the values of checkpoint_segments
(max_wal_size/min_wal_size in Postgres 9.5 and onwards).
--
Michael

thanks

Re: [GENERAL] Begginers question

From
Achilleas Mantzios
Date:
On 16/08/2017 13:46, Alex Samad wrote:


On 16 August 2017 at 16:16, Michael Paquier <michael.paquier@gmail.com> wrote:
On Wed, Aug 16, 2017 at 2:32 PM, Alex Samad <alex@samad.com.au> wrote:
> 1) why did it fill up this time and not previously
> I add this
> archive_command = '/bin/true'
> wal_keep_segments = 1000 # <<< I'm guessing its this
>
> 2) how do I fix up, can I just remove the files from the pg_xlog directory

Don't do that. those files are managed by Postgres so you may finish
with a corrupted cluster. Instead you should lower the value of

Too late, its okay its a learning experience.
 
wal_keep_segments, reload the server parameters, and then enforce two
checkpoints to force WAL segments to be recycled. Note that this

how do I force check points
checkpoint ;
(the ; is not meant as a smiley or whatever )
 
depends also on the values of checkpoint_segments
(max_wal_size/min_wal_size in Postgres 9.5 and onwards).
--
Michael

thanks


-- 
Achilleas Mantzios
IT DEV Lead
IT DEPT
Dynacom Tankers Mgmt

Re: [GENERAL] Begginers question

From
Alex Samad
Date:
Great I will add it to my notes.

Thanks

On 16 August 2017 at 20:55, Achilleas Mantzios <achill@matrix.gatewaynet.com> wrote:
On 16/08/2017 13:46, Alex Samad wrote:


On 16 August 2017 at 16:16, Michael Paquier <michael.paquier@gmail.com> wrote:
On Wed, Aug 16, 2017 at 2:32 PM, Alex Samad <alex@samad.com.au> wrote:
> 1) why did it fill up this time and not previously
> I add this
> archive_command = '/bin/true'
> wal_keep_segments = 1000 # <<< I'm guessing its this
>
> 2) how do I fix up, can I just remove the files from the pg_xlog directory

Don't do that. those files are managed by Postgres so you may finish
with a corrupted cluster. Instead you should lower the value of

Too late, its okay its a learning experience.
 
wal_keep_segments, reload the server parameters, and then enforce two
checkpoints to force WAL segments to be recycled. Note that this

how do I force check points
 checkpoint ;
(the ; is not meant as a smiley or whatever )
 
depends also on the values of checkpoint_segments
(max_wal_size/min_wal_size in Postgres 9.5 and onwards).
--
Michael

thanks


-- 
Achilleas Mantzios
IT DEV Lead
IT DEPT
Dynacom Tankers Mgmt

Re: [GENERAL] Begginers question

From
Alex Samad
Date:


On 16 August 2017 at 20:55, Achilleas Mantzios <achill@matrix.gatewaynet.com> wrote:
On 16/08/2017 13:46, Alex Samad wrote:


On 16 August 2017 at 16:16, Michael Paquier <michael.paquier@gmail.com> wrote:
On Wed, Aug 16, 2017 at 2:32 PM, Alex Samad <alex@samad.com.au> wrote:
> 1) why did it fill up this time and not previously
> I add this
> archive_command = '/bin/true'
> wal_keep_segments = 1000 # <<< I'm guessing its this
>
> 2) how do I fix up, can I just remove the files from the pg_xlog directory

How can I work out how much space this is going to use, 
I see in the comments a segment is 16M, so 1000 would have been 16G, So I thought I would set it at 100 so 1.6G. but it ended up using about 4.5G. So now I am confused
 

Don't do that. those files are managed by Postgres so you may finish
with a corrupted cluster. Instead you should lower the value of

Too late, its okay its a learning experience.
 
wal_keep_segments, reload the server parameters, and then enforce two
checkpoints to force WAL segments to be recycled. Note that this

how do I force check points
 checkpoint ;
(the ; is not meant as a smiley or whatever )
 
depends also on the values of checkpoint_segments
(max_wal_size/min_wal_size in Postgres 9.5 and onwards).
--
Michael

thanks


-- 
Achilleas Mantzios
IT DEV Lead
IT DEPT
Dynacom Tankers Mgmt