<listitem>

<para>

We should implement our own allocator of file space. Usually, to reduce fragmentation, fixed size block allocator is used.

it means that we allocatespace using some fixed quantum. For example if compressed page size is 932 bytes, then we will

allocate 1024block for it in the file.

It means that we allocatesplace using some fixed quantum. For example if compressed page size is 932 bytes, then we will

allocate 1024bytes for it in the file.

</para>

</listitem>

</varlistentry>

</variablelist>

<para>

As far as page location is not fixed and pagean be moved, we can notany moreaccess page directly by its address and need

to use extra level of indirection to map logical address of the page to it physical locationin the disk.

It is done usingmemory-mapped files. In most casesthis mapped will be kept in memory (size of the map is 1000 times smaller size

As far as page location is not fixed and pagecan be moved, we can notanymoreaccess the page directly by its address and need

to use extra level of indirection to map logical address of the page to it's physical locationon the disk.

It is done usingmapping files. In most casesthe mapping will be kept in memory (size of the map is 1000 times smaller size

of the file) and address translation adds almost no overhead to page access time.

But we need to maintainthis extra files: flush them during checkpoint, remove when table is dropped, include them in backup and

But we need to maintainthese extra files: flush them during checkpoint, remove when table is dropped, include them in backup and

so on...

</para>

<para>

PostgreSQL stores relation in a set of files, size of each file is not exceeding 2Gb. Separate page map is constructed for each file.

Garbage collection in CFS is done by several background workers. Number of this workers and pauses in their work can be

configured by database administrator.This workers are splitting work based on inode hash, so them do not conflict with each other.

configured by database administrator.These workers are splitting work based on inode hash, so them do not conflict with each other.

Each file is proceeded separately. The files is blocked for access at the time of garbage collection but complete relation is not

blocked. To ensure data consistency GC creates copies of original data and map files. Once them are flushed to the disk,

new version of data file is atomically renamed to original file name. And then new page map data is copied to memory-mapped file

<para>

CFS can be build with several compression libraries: PostgreSQL lz, zlib, lz4, snappy, lzfse...

But this is build time choice: it is not possible now to dynamically choose compression algorithm.

CFS storesin tablespaceinformation aboutused compression algorithm andproduce error if PostgreSQL isbuild with different

CFS stores information aboutthe compression algorithmused in a tablespaceandproduces error if PostgreSQL isbuilt with different

library.

</para>

<para>

Compression can be enabled for particular tablespaces. System relations are not compressed in any case.

It is not currently possible to alter tablespace compression option, i.e. it is not possible to compressexisted tablespace

orvisa versa - decompress compressed tablespace.

It is not currently possible to alter tablespace compression option, i.e. it is not possible to compressexisting tablespace

orvice versa - decompress compressed tablespace.

</para>

<para>

So to use compression/encryption you need to createtable space with <varname>compression=true</varname> option.

You can make thistable space default tablespace - in this case all tables will be implicitly created in this database:

So to use compression/encryption you need to createa tablespace with <varname>compression=true</varname> option.

You can make thistablespace your default tablespace - in this case all tables will be implicitly created in this database:

</para>

<programlisting>

The file is locked at the period of defragmentation, preventing any access to this part of relation.

When defragmentation is completed, garbage collection waits <varname>cfs_gc_delay</varname> milliseconds and continue directory traversal.

After the end of traversal, GC waits <varname>cfs_gc_period</varname> milliseconds and starts new GC iteration.

If there are more than one GC workers, thenthem split work based on hash of file inode.

If there are more than one GC workers, thenthey split work based on hash of file inode.

</para>

<para>