Thread: Moving pg_xlog

Moving pg_xlog

From
Robert Inder
Date:
I'm running  Postgres9.4 in master/hot-standby mode on a few pairs of servers.

While recovering from A Bit Of Bother last week, I came across a
posting saying that pg_xlog should be on a separate partition.

I tried to find out more about this, by consulting the PostgresQL
documentation (i.e.
https://www.postgresql.org/docs/9.4/static/index.html )
But all I could find was a mention that "It is advantageous if the log
is located on a different disk from the main database files".

The questions:
1. WHY is this good?  Is it (just) to stop pg_xlog filling the
database disk/partition?  Or are there performance implications?
SPECIFICALLY: my database is currently in "/", which is on SSD.  Is it
better to move pg_xlog to another partition on the same SSD?  Or to a
physical disk or SAN?

2. What are the implications for doing a base backup?  I believe I
read that putting pg_xlog on a different partition meant it would be
omitted from a file-system bulk copy (e.g. rsync), and this was a GOOD
thing, because the copy operation would be faster -- not copying
pg_xlog would not prevent the standby server from starting, because
the information it needed would be in the WAL files that would be
shipped separately.  Have I got that right?

Finally, the suggestion.

I'd really like to read an explicit discussion of this in the official
documentation, rather than just glean what I can from answers to
questions.
The possibility of moving pg_xlog to another disk is mentioned in the
documentation, but I almost missed it because it is in "the wrong
place".  It is in Section 29.5 -- "Reliability and the Write Ahead
Log" / "WAL Internals".  But I wasn't interested in anything INTERNAL:
I wanted to know where I should try to locate it/them.  So I'd looked
in "the obvious places" -- Section 18 (Server configuration), and in
particular 18.2 "File Locations".  Could I suggest that the motivation
for doing this, and the consequences for backups, should be discussed
in "the right place" -- in or near the section that talks about file
locations in the context of server configuration.

Robert.

--
Robert Inder,                                    0131 229 1052 / 07808 492 213
Interactive Information Ltd,   3, Lauriston Gardens, Edinburgh EH3 9HH
Registered in Scotland, Company no. SC 150689
                                           Interactions speak louder than words


Re: Moving pg_xlog

From
Achilleas Mantzios
Date:
On 01/12/2016 15:55, Robert Inder wrote:
> I'm running  Postgres9.4 in master/hot-standby mode on a few pairs of servers.
>
> While recovering from A Bit Of Bother last week, I came across a
> posting saying that pg_xlog should be on a separate partition.
>
> I tried to find out more about this, by consulting the PostgresQL
> documentation (i.e.
> https://www.postgresql.org/docs/9.4/static/index.html )
> But all I could find was a mention that "It is advantageous if the log
> is located on a different disk from the main database files".
>
> The questions:
> 1. WHY is this good?  Is it (just) to stop pg_xlog filling the
> database disk/partition?  Or are there performance implications?
> SPECIFICALLY: my database is currently in "/", which is on SSD.  Is it
> better to move pg_xlog to another partition on the same SSD?  Or to a
> physical disk or SAN?

Performance is the reason. You would benefit from moving pg_xlog to a different controller with its own write cache or
toa different SSD with a write cache which is capacitor-backed. So in  
enterprise/server-class setups the above would boost the performance. Using the same SSD with a different partition
won'tgive you much. 

>
> 2. What are the implications for doing a base backup?  I believe I
> read that putting pg_xlog on a different partition meant it would be
> omitted from a file-system bulk copy (e.g. rsync), and this was a GOOD
> thing, because the copy operation would be faster -- not copying
> pg_xlog would not prevent the standby server from starting, because
> the information it needed would be in the WAL files that would be
> shipped separately.  Have I got that right?
Rsync does cross fs boundaries unless you give it the -x option. It is true that the files in pg_xlog won't be useful
tobe taken in the backup. However the wal files to be shipped separately is not  
smth done by itself, you need to enable/implement WAL archiving. What you describe seems to be the "legacy"
old-fashionedway circa 9.0. pg_basebackup (9.1) is more convenient, can create complete  
standalone copies (without the need of any additional wals), can use wal streaming so that you don't depend on wal
archivingor wal_keep_segment, supports repl slots, can create a ready to go hot  
standby, etc.

> Finally, the suggestion.
>
> I'd really like to read an explicit discussion of this in the official
> documentation, rather than just glean what I can from answers to
> questions.
> The possibility of moving pg_xlog to another disk is mentioned in the
> documentation, but I almost missed it because it is in "the wrong
> place".  It is in Section 29.5 -- "Reliability and the Write Ahead
> Log" / "WAL Internals".  But I wasn't interested in anything INTERNAL:
> I wanted to know where I should try to locate it/them.  So I'd looked
> in "the obvious places" -- Section 18 (Server configuration), and in
> particular 18.2 "File Locations".  Could I suggest that the motivation
> for doing this, and the consequences for backups, should be discussed
> in "the right place" -- in or near the section that talks about file
> locations in the context of server configuration.

All I can tell you is I haven't found one single piece of free (or not so free) software with more complete
documentationthan pgsql. 

> Robert.
>


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



Re: Moving pg_xlog

From
Michael Paquier
Date:
On Thu, Dec 01, 2016 at 05:48:51PM +0200, Achilleas Mantzios wrote:
> On 01/12/2016 15:55, Robert Inder wrote:
> > I'm running  Postgres9.4 in master/hot-standby mode on a few pairs of servers.
> >
> > While recovering from A Bit Of Bother last week, I came across a
> > posting saying that pg_xlog should be on a separate partition.
> >
> > I tried to find out more about this, by consulting the PostgresQL
> > documentation (i.e.
> > https://www.postgresql.org/docs/9.4/static/index.html )
> > But all I could find was a mention that "It is advantageous if the log
> > is located on a different disk from the main database files".
> >
> > The questions:
> > 1. WHY is this good?  Is it (just) to stop pg_xlog filling the
> > database disk/partition?  Or are there performance implications?
> > SPECIFICALLY: my database is currently in "/", which is on SSD.  Is it
> > better to move pg_xlog to another partition on the same SSD?  Or to a
> > physical disk or SAN?
>
> Performance is the reason. You would benefit from moving pg_xlog to a
> different controller with its own write cache or to a different SSD with a
> write cache which is capacitor-backed. So in enterprise/server-class setups
> the above would boost the performance. Using the same SSD with a different
> partition won't give you much.

For performance, on-disk write pattern of data in pg_xlog is sequential
writes, while there will be likely random writes on the main data folder.
On top of that, moving them to a different partition gives more flexibility
in the way to tune checkpoint-related parameters using the partition space
as a constraint for retention policy and checkpoint timings.
--
Michael

Attachment

Re: Moving pg_xlog

From
Jeff Janes
Date:
On Thu, Dec 1, 2016 at 5:55 AM, Robert Inder <robert@interactive.co.uk> wrote:
I'm running  Postgres9.4 in master/hot-standby mode on a few pairs of servers.

While recovering from A Bit Of Bother last week, I came across a
posting saying that pg_xlog should be on a separate partition.

I tried to find out more about this, by consulting the PostgresQL
documentation (i.e.
https://www.postgresql.org/docs/9.4/static/index.html )
But all I could find was a mention that "It is advantageous if the log
is located on a different disk from the main database files".

The questions:
1. WHY is this good?  Is it (just) to stop pg_xlog filling the
database disk/partition? 

More like the reverse. Running the data partition out of space is bad. 
Running the pg_xlog partition out of space is worse. Running both partitions out of space at the same time is worse yet, which of course you will do if they are the same partition and that one partition runs out of space.

 
Or are there performance implications?
SPECIFICALLY: my database is currently in "/", which is on SSD.  Is it
better to move pg_xlog to another partition on the same SSD?  Or to a
physical disk or SAN?


If you have something with fast fsyncs (battery backed write cache, maybe SSD), but that is not big enough to hold your entire database, then you would want to put your pg_xlog on that, and the rest of the database on the rest.  (if you are doing OLTP, anyway).

On some kernels and some file systems, having a constant stream of fsyncs (from pg_xlog) interacts poorly with having ordinary non-immediately-synced writes (from the regular data files) on the same partition.
 


2. What are the implications for doing a base backup?  I believe I
read that putting pg_xlog on a different partition meant it would be
omitted from a file-system bulk copy (e.g. rsync),

rsync has lots of options to control what happens with symbolic links and mount points.  Or to exclude certain directories, symbolic links and mount points not withstanding.
 
and this was a GOOD
thing, because the copy operation would be faster -- not copying
pg_xlog would not prevent the standby server from starting, because
the information it needed would be in the WAL files that would be
shipped separately.  Have I got that right?

Finally, the suggestion.

I'd really like to read an explicit discussion of this in the official
documentation, rather than just glean what I can from answers to
questions.

The official documentation cannot have a dissertation on every combination of hardware, OS, file-system type, version of that file-system, and your usage pattern.  That is inherently the realm of the wiki or the blogs.

Cheers,

Jeff

Re: Moving pg_xlog

From
"David G. Johnston"
Date:
On Thu, Dec 1, 2016 at 7:59 PM, Jeff Janes <jeff.janes@gmail.com> wrote:
On Thu, Dec 1, 2016 at 5:55 AM, Robert Inder <robert@interactive.co.uk> wrote:

I'd really like to read an explicit discussion of this in the official
documentation, rather than just glean what I can from answers to
questions.

The official documentation cannot have a dissertation on every combination of hardware, OS, file-system type, version of that file-system, and your usage pattern.  That is inherently the realm of the wiki or the blogs.


​The documentation has enough information at this level of detail that I wouldn't object to adding commentary addressing the above should someone take the time to write it.​

Given that the location of pg_xlog is not "configurable" placing such commentary in Server Configuration would be a no-go, however.  At a quick glance a new section under "Server Setup and Operation - Creating a Database Cluster" would probably be a better home.  It already discusses Secondary File Systems and in many ways this is just an extension of that discussion.

David J.

Re: Moving pg_xlog

From
Melvin Davidson
Date:


On Thu, Dec 1, 2016 at 10:17 PM, David G. Johnston <david.g.johnston@gmail.com> wrote:
On Thu, Dec 1, 2016 at 7:59 PM, Jeff Janes <jeff.janes@gmail.com> wrote:
On Thu, Dec 1, 2016 at 5:55 AM, Robert Inder <robert@interactive.co.uk> wrote:

I'd really like to read an explicit discussion of this in the official
documentation, rather than just glean what I can from answers to
questions.

The official documentation cannot have a dissertation on every combination of hardware, OS, file-system type, version of that file-system, and your usage pattern.  That is inherently the realm of the wiki or the blogs.


​The documentation has enough information at this level of detail that I wouldn't object to adding commentary addressing the above should someone take the time to write it.​

Given that the location of pg_xlog is not "configurable" placing such commentary in Server Configuration would be a no-go, however.  At a quick glance a new section under "Server Setup and Operation - Creating a Database Cluster" would probably be a better home.  It already discusses Secondary File Systems and in many ways this is just an extension of that discussion.

David J.

 
>Given that the location of pg_xlog is not "configurable"

 Well, while the location of pg_xlog is not currently configurable, on Linux system the way to do it is  to:
 1. stop PostgreSQL
 2. move the pg_xlog directory to a separate partition
 3. create a symbolic link to point to the new partition
 4. restart PostgreSQL

--
Melvin Davidson
I reserve the right to fantasize.  Whether or not you
wish to share my fantasy is entirely up to you.

Re: Moving pg_xlog

From
Michael Paquier
Date:
On Fri, Dec 2, 2016 at 1:04 PM, Melvin Davidson <melvin6925@gmail.com> wrote:
>  Well, while the location of pg_xlog is not currently configurable, on Linux system the way to do it is  to:
>  1. stop PostgreSQL
>  2. move the pg_xlog directory to a separate partition
>  3. create a symbolic link to point to the new partition
>  4. restart PostgreSQL

Similar flow on Windows, just use a junction point for the link.
--
Michael


Re: Moving pg_xlog

From
Lucas Possamai
Date:
2016-12-02 17:10 GMT+13:00 Michael Paquier <michael.paquier@gmail.com>:
On Fri, Dec 2, 2016 at 1:04 PM, Melvin Davidson <melvin6925@gmail.com> wrote:
>  Well, while the location of pg_xlog is not currently configurable, on Linux system the way to do it is  to:
>  1. stop PostgreSQL
>  2. move the pg_xlog directory to a separate partition
>  3. create a symbolic link to point to the new partition
>  4. restart PostgreSQL

Similar flow on Windows, just use a junction point for the link.
--
Michael

I've done this on my Postgres 9.2 DB server running CentOS 6.7...

And it's pretty much what the guys told you already:


>  1. stop PostgreSQL
>  2. move the pg_xlog directory to a separate partition
>  3. create a symbolic link to point to the new partition
>  4. restart PostgreSQL

In my case, it significantly improved I/O performance.

Lucas

Re: Moving pg_xlog

From
Steven Winfield
Date:

2016-12-02 17:10 GMT+13:00 Michael Paquier <michael.paquier@gmail.com>:

On Fri, Dec 2, 2016 at 1:04 PM, Melvin Davidson <melvin6925@gmail.com> wrote:
>  Well, while the location of pg_xlog is not currently configurable, on Linux system the way to do it is  to:
>  1. stop PostgreSQL
>  2. move the pg_xlog directory to a separate partition
>  3. create a symbolic link to point to the new partition
>  4. restart PostgreSQL

It’s also worth mentioning that the xlog directory can be specified at cluster creation time using ‘initdb -X <directory>’, which (AFAICT) just creates the symlink for you.

 

Steve.

 

 

Dr. Steven Winfield

Scientist

D: +44 (0)1223 755 776

 

Cantab email sig

 

 

Attachment

Re: Moving pg_xlog

From
Jeff Janes
Date:
On Thu, Dec 1, 2016 at 6:17 PM, Michael Paquier <michael.paquier@gmail.com> wrote:
On Thu, Dec 01, 2016 at 05:48:51PM +0200, Achilleas Mantzios wrote:
>
> Performance is the reason. You would benefit from moving pg_xlog to a
> different controller with its own write cache or to a different SSD with a
> write cache which is capacitor-backed. So in enterprise/server-class setups
> the above would boost the performance. Using the same SSD with a different
> partition won't give you much.

For performance, on-disk write pattern of data in pg_xlog is sequential
writes, while there will be likely random writes on the main data folder.

This is only the case if you have a write cache, or are doing bulk loads.  With small transactions and without a write cache, the need for constant syncs totally destroys the benefits of sequential writes.

Cheers,

Jeff

Re: Moving pg_xlog

From
Robert Inder
Date:
Thanks, everyone, for your comments.

I think I've got a clearer idea of what's going on now...

Robert.


On 1 December 2016 at 13:55, Robert Inder <robert@interactive.co.uk> wrote:
> I'm running  Postgres9.4 in master/hot-standby mode on a few pairs of servers.
>
> While recovering from A Bit Of Bother last week, I came across a
> posting saying that pg_xlog should be on a separate partition.
>
> I tried to find out more about this, by consulting the PostgresQL
> documentation (i.e.
> https://www.postgresql.org/docs/9.4/static/index.html )
> But all I could find was a mention that "It is advantageous if the log
> is located on a different disk from the main database files".
>
> The questions:
> 1. WHY is this good?  Is it (just) to stop pg_xlog filling the
> database disk/partition?  Or are there performance implications?
> SPECIFICALLY: my database is currently in "/", which is on SSD.  Is it
> better to move pg_xlog to another partition on the same SSD?  Or to a
> physical disk or SAN?
>
> 2. What are the implications for doing a base backup?  I believe I
> read that putting pg_xlog on a different partition meant it would be
> omitted from a file-system bulk copy (e.g. rsync), and this was a GOOD
> thing, because the copy operation would be faster -- not copying
> pg_xlog would not prevent the standby server from starting, because
> the information it needed would be in the WAL files that would be
> shipped separately.  Have I got that right?
>
> Finally, the suggestion.
>
> I'd really like to read an explicit discussion of this in the official
> documentation, rather than just glean what I can from answers to
> questions.
> The possibility of moving pg_xlog to another disk is mentioned in the
> documentation, but I almost missed it because it is in "the wrong
> place".  It is in Section 29.5 -- "Reliability and the Write Ahead
> Log" / "WAL Internals".  But I wasn't interested in anything INTERNAL:
> I wanted to know where I should try to locate it/them.  So I'd looked
> in "the obvious places" -- Section 18 (Server configuration), and in
> particular 18.2 "File Locations".  Could I suggest that the motivation
> for doing this, and the consequences for backups, should be discussed
> in "the right place" -- in or near the section that talks about file
> locations in the context of server configuration.
>
> Robert.
>
> --
> Robert Inder,                                    0131 229 1052 / 07808 492 213
> Interactive Information Ltd,   3, Lauriston Gardens, Edinburgh EH3 9HH
> Registered in Scotland, Company no. SC 150689
>                                            Interactions speak louder than words



--
Robert Inder,                                    0131 229 1052 / 07808 492 213
Interactive Information Ltd,   3, Lauriston Gardens, Edinburgh EH3 9HH
Registered in Scotland, Company no. SC 150689
                                           Interactions speak louder than words