[Patch] Make block and file size for WAL and relations defined atcluster creation

Поиск
Список
Период
Сортировка
От Remi Colinet
Тема [Patch] Make block and file size for WAL and relations defined atcluster creation
Дата
Msg-id CADdR5nxSfGrBqm6tqFmzNPWiO_dTMqwM=eHrLkSyX81fHgfeWw@mail.gmail.com
обсуждение исходный текст
Ответы Re: [Patch] Make block and file size for WAL and relations defined atcluster creation  (Robert Haas <robertmhaas@gmail.com>)
Список pgsql-hackers
Hello,

So far, the relation's block and file sizes have been defined statically at server build time.
This is also the case for the WAL block size.

This means that it is not possible to share the same Postgresql binary if using different block and file sizes for the WAL and the relations, and across different server instances/databases.


Recently, the value definition of the WAL file size has been converted from server build time to cluster creation time. The current patch goes further in this direction with the relation block and file sizes and with the WAL block size. And some more could be done with LOBLKSIZE for instance.


Below patch makes block and file sizes defined at cluster creation for both the WAL and the relations. This avoids having different server builds for each possible combination of block size and file sizes.


With the patch, the values of the block and file sizes are kept in the control file (This has been the case so far) and are provided to initdb when creating the cluster. If no value is specified, the default values are used.

Values which can be defined at cluster creation time are:

- the WAL block size
- the WAL file size
- the relation block size
- the relation file size

I noticed that the names of these parameters can slightly vary throughout the source code for the name, the unit and the case used.

Such names are:

BLCKSZ: the relation block size in bytes
RELSEG_SIZE: maximum number of blocks allowed in one disk file
XLOG_BLCKSZ: the WAL block size in bytes
XLOG_SEG_SIZE: the WAL file size in bytes
blcksz (in control file): the relation block size in bytes (same as BLCKSZ)
relseg_size (in control file): the relation file size in blocks (same as RELSEG_SIZE)
xlog_blcksz (in control file): WAL block size in bytes (same as XLOG_BLCKSZ)
xlog_seg_size (in control file); the WAL file size in bytes (same as XLOG_SEG_SIZE)
WalSegSz (in pg_resetwal.c): the WAL segment size in bytes
wal_segment_size (in xlog.c): the WAL segment size in bytes
segment_size (in guc.c): the relation segment size

For the current patch, I defined common names to be used throughout in the source code, whether this in the server or in the different utilities with units in both blocks and bytes.

These are:

- wal_blck_size: which replaces XLOG_BLCKSZ
- wal_file_blck
- wal_file_size which is wal_blck_size * wal_file_blck. It replaces XLOG_SEG_SIZE and wal_segment_size

- rel_blck_size: which replaces BLCKSZ
- rel_file_blck: it replaces RELSEG_SIZE and segment_size
- rel_file_size which is rel_blck_size * rel_file_blck.

Lower case letters are used to remind that these values are not statically defined at compile time.

This is so far a POC to show that the change is not very big and worth the extra code needed. The patch is made of only small changes unless a few files which require some more work with palloc/pfree.

The patch is rather simple despite it modifies many different files.

I've tested the patch with different combination of block and file sizes for the WAL and the relations.

Feel free to comment.

Regards
Remi


Patch diffstat:

[root@rco v1]# diffstat blkfilesizes_v1.patch
 TODO                                            |   56 ++
 configure.in                                    |   94 ---
 contrib/amcheck/verify_nbtree.c                 |    4
 contrib/bloom/blinsert.c                        |   14
 contrib/bloom/bloom.h                           |   26 -
 contrib/bloom/blutils.c                         |    6
 contrib/bloom/blvacuum.c                        |    6
 contrib/file_fdw/file_fdw.c                     |    6
 contrib/pageinspect/brinfuncs.c                 |    8
 contrib/pageinspect/btreefuncs.c                |    6
 contrib/pageinspect/rawpage.c                   |   12
 contrib/pg_prewarm/pg_prewarm.c                 |    4
 contrib/pg_standby/pg_standby.c                 |    7
 contrib/pgstattuple/pgstatapprox.c              |    6
 contrib/pgstattuple/pgstatindex.c               |    4
 contrib/pgstattuple/pgstattuple.c               |   10
 contrib/postgres_fdw/deparse.c                  |    2
 contrib/postgres_fdw/postgres_fdw.c             |    2
 param.sh                                        |    1
 src/backend/access/brin/brin_pageops.c          |    4
 src/backend/access/common/bufmask.c             |    4
 src/backend/access/common/reloptions.c          |    8
 src/backend/access/gin/ginbtree.c               |   12
 src/backend/access/gin/gindatapage.c            |   18
 src/backend/access/gin/ginentrypage.c           |    2
 src/backend/access/gin/ginfast.c                |    6
 src/backend/access/gin/ginget.c                 |    6
 src/backend/access/gin/ginvacuum.c              |    2
 src/backend/access/gin/ginxlog.c                |    4
 src/backend/access/gist/gistbuild.c             |    8
 src/backend/access/gist/gistbuildbuffers.c      |   10
 src/backend/access/gist/gistscan.c              |    1
 src/backend/access/hash/hash.c                  |    7
 src/backend/access/hash/hashpage.c              |    4
 src/backend/access/heap/README.HOT              |    2
 src/backend/access/heap/heapam.c                |   17
 src/backend/access/heap/pruneheap.c             |   39 +
 src/backend/access/heap/rewriteheap.c           |    4
 src/backend/access/heap/syncscan.c              |    2
 src/backend/access/heap/visibilitymap.c         |    8
 src/backend/access/nbtree/nbtpage.c             |    2
 src/backend/access/nbtree/nbtree.c              |   18
 src/backend/access/nbtree/nbtsearch.c           |    5
 src/backend/access/nbtree/nbtsort.c             |   10
 src/backend/access/spgist/spgdoinsert.c         |    4
 src/backend/access/spgist/spginsert.c           |    2
 src/backend/access/spgist/spgscan.c             |    1
 src/backend/access/spgist/spgtextproc.c         |   10
 src/backend/access/spgist/spgutils.c            |    4
 src/backend/access/transam/README               |    2
 src/backend/access/transam/clog.c               |   10
 src/backend/access/transam/commit_ts.c          |    4
 src/backend/access/transam/generic_xlog.c       |   44 +
 src/backend/access/transam/multixact.c          |   12
 src/backend/access/transam/slru.c               |   22
 src/backend/access/transam/subtrans.c           |    5
 src/backend/access/transam/timeline.c           |    2
 src/backend/access/transam/twophase.c           |    2
 src/backend/access/transam/xlog.c               |  603 ++++++++++++++----------
 src/backend/access/transam/xlogarchive.c        |   12
 src/backend/access/transam/xlogfuncs.c          |   10
 src/backend/access/transam/xloginsert.c         |   48 +
 src/backend/access/transam/xlogreader.c         |  141 +++--
 src/backend/access/transam/xlogutils.c          |   34 -
 src/backend/bootstrap/bootstrap.c               |   33 -
 src/backend/commands/async.c                    |   15
 src/backend/commands/tablecmds.c                |    2
 src/backend/commands/vacuumlazy.c               |    4
 src/backend/executor/execGrouping.c             |    1
 src/backend/nodes/tidbitmap.c                   |  152 +++++-
 src/backend/optimizer/path/costsize.c           |   10
 src/backend/optimizer/util/plancat.c            |    2
 src/backend/postmaster/checkpointer.c           |    4
 src/backend/replication/basebackup.c            |   30 -
 src/backend/replication/logical/logical.c       |    2
 src/backend/replication/logical/reorderbuffer.c |   18
 src/backend/replication/slot.c                  |    2
 src/backend/replication/walreceiver.c           |   14
 src/backend/replication/walreceiverfuncs.c      |    4
 src/backend/replication/walsender.c             |   30 -
 src/backend/storage/buffer/buf_init.c           |    4
 src/backend/storage/buffer/bufmgr.c             |    8
 src/backend/storage/buffer/freelist.c           |    6
 src/backend/storage/buffer/localbuf.c           |    6
 src/backend/storage/file/buffile.c              |   20
 src/backend/storage/file/copydir.c              |    2
 src/backend/storage/freespace/README            |    8
 src/backend/storage/freespace/freespace.c       |   36 -
 src/backend/storage/freespace/indexfsm.c        |    7
 src/backend/storage/lmgr/predicate.c            |    2
 src/backend/storage/page/bufpage.c              |   27 -
 src/backend/storage/smgr/md.c                   |  104 ++--
 src/backend/tcop/postgres.c                     |    2
 src/backend/utils/adt/selfuncs.c                |    2
 src/backend/utils/init/globals.c                |   20
 src/backend/utils/init/miscinit.c               |    6
 src/backend/utils/init/postinit.c               |   23
 src/backend/utils/misc/guc.c                    |  175 ++++--
 src/backend/utils/misc/pg_controldata.c         |    4
 src/backend/utils/sort/logtape.c                |   49 -
 src/backend/utils/sort/tuplesort.c              |    6
 src/bin/initdb/initdb.c                         |  304 +++++++++---
 src/bin/pg_basebackup/pg_basebackup.c           |   18
 src/bin/pg_basebackup/pg_receivewal.c           |   26 -
 src/bin/pg_basebackup/pg_recvlogical.c          |   11
 src/bin/pg_basebackup/receivelog.c              |   28 -
 src/bin/pg_basebackup/streamutil.c              |   76 +--
 src/bin/pg_basebackup/streamutil.h              |    6
 src/bin/pg_basebackup/walmethods.c              |   14
 src/bin/pg_controldata/pg_controldata.c         |   16
 src/bin/pg_resetwal/pg_resetwal.c               |  125 +++-
 src/bin/pg_rewind/copy_fetch.c                  |    9
 src/bin/pg_rewind/filemap.c                     |   11
 src/bin/pg_rewind/libpq_fetch.c                 |    7
 src/bin/pg_rewind/parsexlog.c                   |   26 -
 src/bin/pg_rewind/pg_rewind.c                   |   33 -
 src/bin/pg_test_fsync/pg_test_fsync.c           |   71 +-
 src/bin/pg_upgrade/controldata.c                |    7
 src/bin/pg_upgrade/file.c                       |   15
 src/bin/pg_upgrade/pg_upgrade.c                 |    3
 src/bin/pg_waldump/pg_waldump.c                 |   69 +-
 src/common/controldata_utils.c                  |   98 +++
 src/include/access/brin_page.h                  |    2
 src/include/access/ginblock.h                   |    6
 src/include/access/gist_private.h               |   20
 src/include/access/hash.h                       |    5
 src/include/access/htup_details.h               |   11
 src/include/access/itup.h                       |    2
 src/include/access/nbtree.h                     |   10
 src/include/access/relscan.h                    |    7
 src/include/access/slru.h                       |    2
 src/include/access/spgist_private.h             |   22
 src/include/access/tuptoaster.h                 |    2
 src/include/access/xlog_internal.h              |    8
 src/include/access/xlogreader.h                 |    9
 src/include/access/xlogrecord.h                 |    6
 src/include/common/controldata_utils.h          |    4
 src/include/lib/simplehash.h                    |   12
 src/include/nodes/execnodes.h                   |    1
 src/include/nodes/nodes.h                       |    1
 src/include/pg_config.h.in                      |   31 -
 src/include/pg_config_manual.h                  |    8
 src/include/pg_control_def.h                    |   44 +
 src/include/storage/bufmgr.h                    |    4
 src/include/storage/bufpage.h                   |    5
 src/include/storage/checksum_impl.h             |    2
 src/include/storage/fsm_internals.h             |    5
 src/include/storage/large_object.h              |    4
 src/include/storage/md.h                        |   12
 src/include/storage/off.h                       |    2
 src/include/utils/rel.h                         |    4
 src/interfaces/libpq/libpq-int.h                |    5
 152 files changed, 2234 insertions(+), 1284 deletions(-)
[root@rco v1]# 



Вложения

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Tom Lane
Дата:
Сообщение: Re: [PATCH] GET DIAGNOSTICS FUNCTION_NAME
Следующее
От: Bear Giles
Дата:
Сообщение: Re: What does Time.MAX_VALUE actually represent?