postgresql/src/bin/pg_basebackup/walmethods.c

1022 lines
22 KiB
C
Raw Normal View History

/*-------------------------------------------------------------------------
*
* walmethods.c - implementations of different ways to write received wal
*
* NOTE! The caller must ensure that only one method is instantiated in
* any given program, and that it's only instantiated once!
*
* Portions Copyright (c) 1996-2020, PostgreSQL Global Development Group
*
* IDENTIFICATION
* src/bin/pg_basebackup/walmethods.c
*-------------------------------------------------------------------------
*/
#include "postgres_fe.h"
#include <sys/stat.h>
#include <time.h>
#include <unistd.h>
#ifdef HAVE_LIBZ
#include <zlib.h>
#endif
#include "common/file_perm.h"
#include "common/file_utils.h"
#include "pgtar.h"
#include "receivelog.h"
#include "streamutil.h"
/* Size of zlib buffer for .tar.gz */
#define ZLIB_OUT_SIZE 4096
/*-------------------------------------------------------------------------
* WalDirectoryMethod - write wal to a directory looking like pg_wal
*-------------------------------------------------------------------------
*/
/*
* Global static data for this method
*/
typedef struct DirectoryMethodData
{
char *basedir;
int compression;
bool sync;
} DirectoryMethodData;
static DirectoryMethodData *dir_data = NULL;
/*
* Local file handle
*/
typedef struct DirectoryMethodFile
{
int fd;
off_t currpos;
char *pathname;
char *fullpath;
char *temp_suffix;
#ifdef HAVE_LIBZ
gzFile gzfp;
#endif
} DirectoryMethodFile;
static const char *
dir_getlasterror(void)
{
/* Directory method always sets errno, so just use strerror */
return strerror(errno);
}
static Walfile
dir_open_for_write(const char *pathname, const char *temp_suffix, size_t pad_to_size)
{
static char tmppath[MAXPGPATH];
int fd;
DirectoryMethodFile *f;
#ifdef HAVE_LIBZ
gzFile gzfp = NULL;
#endif
snprintf(tmppath, sizeof(tmppath), "%s/%s%s%s",
dir_data->basedir, pathname,
dir_data->compression > 0 ? ".gz" : "",
temp_suffix ? temp_suffix : "");
/*
* Open a file for non-compressed as well as compressed files. Tracking
* the file descriptor is important for dir_sync() method as gzflush()
* does not do any system calls to fsync() to make changes permanent on
* disk.
*/
fd = open(tmppath, O_WRONLY | O_CREAT | PG_BINARY, pg_file_create_mode);
if (fd < 0)
return NULL;
#ifdef HAVE_LIBZ
if (dir_data->compression > 0)
{
gzfp = gzdopen(fd, "wb");
if (gzfp == NULL)
{
close(fd);
return NULL;
}
if (gzsetparams(gzfp, dir_data->compression,
Z_DEFAULT_STRATEGY) != Z_OK)
{
gzclose(gzfp);
return NULL;
}
}
#endif
/* Do pre-padding on non-compressed files */
if (pad_to_size && dir_data->compression == 0)
{
PGAlignedXLogBlock zerobuf;
int bytes;
memset(zerobuf.data, 0, XLOG_BLCKSZ);
for (bytes = 0; bytes < pad_to_size; bytes += XLOG_BLCKSZ)
{
errno = 0;
if (write(fd, zerobuf.data, XLOG_BLCKSZ) != XLOG_BLCKSZ)
{
int save_errno = errno;
close(fd);
/*
* If write didn't set errno, assume problem is no disk space.
*/
errno = save_errno ? save_errno : ENOSPC;
return NULL;
}
}
if (lseek(fd, 0, SEEK_SET) != 0)
{
int save_errno = errno;
close(fd);
errno = save_errno;
return NULL;
}
}
/*
* fsync WAL file and containing directory, to ensure the file is
* persistently created and zeroed (if padded). That's particularly
* important when using synchronous mode, where the file is modified and
* fsynced in-place, without a directory fsync.
*/
if (dir_data->sync)
{
Unified logging system for command-line programs This unifies the various ad hoc logging (message printing, error printing) systems used throughout the command-line programs. Features: - Program name is automatically prefixed. - Message string does not end with newline. This removes a common source of inconsistencies and omissions. - Additionally, a final newline is automatically stripped, simplifying use of PQerrorMessage() etc., another common source of mistakes. - I converted error message strings to use %m where possible. - As a result of the above several points, more translatable message strings can be shared between different components and between frontends and backend, without gratuitous punctuation or whitespace differences. - There is support for setting a "log level". This is not meant to be user-facing, but can be used internally to implement debug or verbose modes. - Lazy argument evaluation, so no significant overhead if logging at some level is disabled. - Some color in the messages, similar to gcc and clang. Set PG_COLOR=auto to try it out. Some colors are predefined, but can be customized by setting PG_COLORS. - Common files (common/, fe_utils/, etc.) can handle logging much more simply by just using one API without worrying too much about the context of the calling program, requiring callbacks, or having to pass "progname" around everywhere. - Some programs called setvbuf() to make sure that stderr is unbuffered, even on Windows. But not all programs did that. This is now done centrally. Soft goals: - Reduces vertical space use and visual complexity of error reporting in the source code. - Encourages more deliberate classification of messages. For example, in some cases it wasn't clear without analyzing the surrounding code whether a message was meant as an error or just an info. - Concepts and terms are vaguely aligned with popular logging frameworks such as log4j and Python logging. This is all just about printing stuff out. Nothing affects program flow (e.g., fatal exits). The uses are just too varied to do that. Some existing code had wrappers that do some kind of print-and-exit, and I adapted those. I tried to keep the output mostly the same, but there is a lot of historical baggage to unwind and special cases to consider, and I might not always have succeeded. One significant change is that pg_rewind used to write all error messages to stdout. That is now changed to stderr. Reviewed-by: Donald Dong <xdong@csumb.edu> Reviewed-by: Arthur Zakirov <a.zakirov@postgrespro.ru> Discussion: https://www.postgresql.org/message-id/flat/6a609b43-4f57-7348-6480-bd022f924310@2ndquadrant.com
2019-04-01 14:24:37 +02:00
if (fsync_fname(tmppath, false) != 0 ||
fsync_parent_path(tmppath) != 0)
{
#ifdef HAVE_LIBZ
if (dir_data->compression > 0)
gzclose(gzfp);
else
#endif
close(fd);
return NULL;
}
}
f = pg_malloc0(sizeof(DirectoryMethodFile));
#ifdef HAVE_LIBZ
if (dir_data->compression > 0)
f->gzfp = gzfp;
#endif
f->fd = fd;
f->currpos = 0;
f->pathname = pg_strdup(pathname);
f->fullpath = pg_strdup(tmppath);
if (temp_suffix)
f->temp_suffix = pg_strdup(temp_suffix);
return f;
}
static ssize_t
dir_write(Walfile f, const void *buf, size_t count)
{
ssize_t r;
DirectoryMethodFile *df = (DirectoryMethodFile *) f;
Assert(f != NULL);
#ifdef HAVE_LIBZ
if (dir_data->compression > 0)
r = (ssize_t) gzwrite(df->gzfp, buf, count);
else
#endif
r = write(df->fd, buf, count);
if (r > 0)
df->currpos += r;
return r;
}
static off_t
dir_get_current_pos(Walfile f)
{
Assert(f != NULL);
/* Use a cached value to prevent lots of reseeks */
return ((DirectoryMethodFile *) f)->currpos;
}
static int
dir_close(Walfile f, WalCloseMethod method)
{
int r;
DirectoryMethodFile *df = (DirectoryMethodFile *) f;
static char tmppath[MAXPGPATH];
static char tmppath2[MAXPGPATH];
Assert(f != NULL);
#ifdef HAVE_LIBZ
if (dir_data->compression > 0)
r = gzclose(df->gzfp);
else
#endif
r = close(df->fd);
if (r == 0)
{
/* Build path to the current version of the file */
if (method == CLOSE_NORMAL && df->temp_suffix)
{
/*
* If we have a temp prefix, normal operation is to rename the
* file.
*/
snprintf(tmppath, sizeof(tmppath), "%s/%s%s%s",
dir_data->basedir, df->pathname,
dir_data->compression > 0 ? ".gz" : "",
df->temp_suffix);
snprintf(tmppath2, sizeof(tmppath2), "%s/%s%s",
dir_data->basedir, df->pathname,
dir_data->compression > 0 ? ".gz" : "");
Unified logging system for command-line programs This unifies the various ad hoc logging (message printing, error printing) systems used throughout the command-line programs. Features: - Program name is automatically prefixed. - Message string does not end with newline. This removes a common source of inconsistencies and omissions. - Additionally, a final newline is automatically stripped, simplifying use of PQerrorMessage() etc., another common source of mistakes. - I converted error message strings to use %m where possible. - As a result of the above several points, more translatable message strings can be shared between different components and between frontends and backend, without gratuitous punctuation or whitespace differences. - There is support for setting a "log level". This is not meant to be user-facing, but can be used internally to implement debug or verbose modes. - Lazy argument evaluation, so no significant overhead if logging at some level is disabled. - Some color in the messages, similar to gcc and clang. Set PG_COLOR=auto to try it out. Some colors are predefined, but can be customized by setting PG_COLORS. - Common files (common/, fe_utils/, etc.) can handle logging much more simply by just using one API without worrying too much about the context of the calling program, requiring callbacks, or having to pass "progname" around everywhere. - Some programs called setvbuf() to make sure that stderr is unbuffered, even on Windows. But not all programs did that. This is now done centrally. Soft goals: - Reduces vertical space use and visual complexity of error reporting in the source code. - Encourages more deliberate classification of messages. For example, in some cases it wasn't clear without analyzing the surrounding code whether a message was meant as an error or just an info. - Concepts and terms are vaguely aligned with popular logging frameworks such as log4j and Python logging. This is all just about printing stuff out. Nothing affects program flow (e.g., fatal exits). The uses are just too varied to do that. Some existing code had wrappers that do some kind of print-and-exit, and I adapted those. I tried to keep the output mostly the same, but there is a lot of historical baggage to unwind and special cases to consider, and I might not always have succeeded. One significant change is that pg_rewind used to write all error messages to stdout. That is now changed to stderr. Reviewed-by: Donald Dong <xdong@csumb.edu> Reviewed-by: Arthur Zakirov <a.zakirov@postgrespro.ru> Discussion: https://www.postgresql.org/message-id/flat/6a609b43-4f57-7348-6480-bd022f924310@2ndquadrant.com
2019-04-01 14:24:37 +02:00
r = durable_rename(tmppath, tmppath2);
}
else if (method == CLOSE_UNLINK)
{
/* Unlink the file once it's closed */
snprintf(tmppath, sizeof(tmppath), "%s/%s%s%s",
dir_data->basedir, df->pathname,
dir_data->compression > 0 ? ".gz" : "",
df->temp_suffix ? df->temp_suffix : "");
r = unlink(tmppath);
}
else
{
/*
* Else either CLOSE_NORMAL and no temp suffix, or
* CLOSE_NO_RENAME. In this case, fsync the file and containing
* directory if sync mode is requested.
*/
if (dir_data->sync)
{
Unified logging system for command-line programs This unifies the various ad hoc logging (message printing, error printing) systems used throughout the command-line programs. Features: - Program name is automatically prefixed. - Message string does not end with newline. This removes a common source of inconsistencies and omissions. - Additionally, a final newline is automatically stripped, simplifying use of PQerrorMessage() etc., another common source of mistakes. - I converted error message strings to use %m where possible. - As a result of the above several points, more translatable message strings can be shared between different components and between frontends and backend, without gratuitous punctuation or whitespace differences. - There is support for setting a "log level". This is not meant to be user-facing, but can be used internally to implement debug or verbose modes. - Lazy argument evaluation, so no significant overhead if logging at some level is disabled. - Some color in the messages, similar to gcc and clang. Set PG_COLOR=auto to try it out. Some colors are predefined, but can be customized by setting PG_COLORS. - Common files (common/, fe_utils/, etc.) can handle logging much more simply by just using one API without worrying too much about the context of the calling program, requiring callbacks, or having to pass "progname" around everywhere. - Some programs called setvbuf() to make sure that stderr is unbuffered, even on Windows. But not all programs did that. This is now done centrally. Soft goals: - Reduces vertical space use and visual complexity of error reporting in the source code. - Encourages more deliberate classification of messages. For example, in some cases it wasn't clear without analyzing the surrounding code whether a message was meant as an error or just an info. - Concepts and terms are vaguely aligned with popular logging frameworks such as log4j and Python logging. This is all just about printing stuff out. Nothing affects program flow (e.g., fatal exits). The uses are just too varied to do that. Some existing code had wrappers that do some kind of print-and-exit, and I adapted those. I tried to keep the output mostly the same, but there is a lot of historical baggage to unwind and special cases to consider, and I might not always have succeeded. One significant change is that pg_rewind used to write all error messages to stdout. That is now changed to stderr. Reviewed-by: Donald Dong <xdong@csumb.edu> Reviewed-by: Arthur Zakirov <a.zakirov@postgrespro.ru> Discussion: https://www.postgresql.org/message-id/flat/6a609b43-4f57-7348-6480-bd022f924310@2ndquadrant.com
2019-04-01 14:24:37 +02:00
r = fsync_fname(df->fullpath, false);
if (r == 0)
Unified logging system for command-line programs This unifies the various ad hoc logging (message printing, error printing) systems used throughout the command-line programs. Features: - Program name is automatically prefixed. - Message string does not end with newline. This removes a common source of inconsistencies and omissions. - Additionally, a final newline is automatically stripped, simplifying use of PQerrorMessage() etc., another common source of mistakes. - I converted error message strings to use %m where possible. - As a result of the above several points, more translatable message strings can be shared between different components and between frontends and backend, without gratuitous punctuation or whitespace differences. - There is support for setting a "log level". This is not meant to be user-facing, but can be used internally to implement debug or verbose modes. - Lazy argument evaluation, so no significant overhead if logging at some level is disabled. - Some color in the messages, similar to gcc and clang. Set PG_COLOR=auto to try it out. Some colors are predefined, but can be customized by setting PG_COLORS. - Common files (common/, fe_utils/, etc.) can handle logging much more simply by just using one API without worrying too much about the context of the calling program, requiring callbacks, or having to pass "progname" around everywhere. - Some programs called setvbuf() to make sure that stderr is unbuffered, even on Windows. But not all programs did that. This is now done centrally. Soft goals: - Reduces vertical space use and visual complexity of error reporting in the source code. - Encourages more deliberate classification of messages. For example, in some cases it wasn't clear without analyzing the surrounding code whether a message was meant as an error or just an info. - Concepts and terms are vaguely aligned with popular logging frameworks such as log4j and Python logging. This is all just about printing stuff out. Nothing affects program flow (e.g., fatal exits). The uses are just too varied to do that. Some existing code had wrappers that do some kind of print-and-exit, and I adapted those. I tried to keep the output mostly the same, but there is a lot of historical baggage to unwind and special cases to consider, and I might not always have succeeded. One significant change is that pg_rewind used to write all error messages to stdout. That is now changed to stderr. Reviewed-by: Donald Dong <xdong@csumb.edu> Reviewed-by: Arthur Zakirov <a.zakirov@postgrespro.ru> Discussion: https://www.postgresql.org/message-id/flat/6a609b43-4f57-7348-6480-bd022f924310@2ndquadrant.com
2019-04-01 14:24:37 +02:00
r = fsync_parent_path(df->fullpath);
}
}
}
pg_free(df->pathname);
pg_free(df->fullpath);
if (df->temp_suffix)
pg_free(df->temp_suffix);
pg_free(df);
return r;
}
static int
dir_sync(Walfile f)
{
Assert(f != NULL);
if (!dir_data->sync)
return 0;
#ifdef HAVE_LIBZ
if (dir_data->compression > 0)
{
if (gzflush(((DirectoryMethodFile *) f)->gzfp, Z_SYNC_FLUSH) != Z_OK)
return -1;
}
#endif
return fsync(((DirectoryMethodFile *) f)->fd);
}
static ssize_t
dir_get_file_size(const char *pathname)
{
struct stat statbuf;
static char tmppath[MAXPGPATH];
snprintf(tmppath, sizeof(tmppath), "%s/%s",
dir_data->basedir, pathname);
if (stat(tmppath, &statbuf) != 0)
return -1;
return statbuf.st_size;
}
static bool
dir_existsfile(const char *pathname)
{
static char tmppath[MAXPGPATH];
int fd;
snprintf(tmppath, sizeof(tmppath), "%s/%s",
dir_data->basedir, pathname);
fd = open(tmppath, O_RDONLY | PG_BINARY, 0);
if (fd < 0)
return false;
close(fd);
return true;
}
static bool
dir_finish(void)
{
if (dir_data->sync)
{
/*
* Files are fsynced when they are closed, but we need to fsync the
* directory entry here as well.
*/
Unified logging system for command-line programs This unifies the various ad hoc logging (message printing, error printing) systems used throughout the command-line programs. Features: - Program name is automatically prefixed. - Message string does not end with newline. This removes a common source of inconsistencies and omissions. - Additionally, a final newline is automatically stripped, simplifying use of PQerrorMessage() etc., another common source of mistakes. - I converted error message strings to use %m where possible. - As a result of the above several points, more translatable message strings can be shared between different components and between frontends and backend, without gratuitous punctuation or whitespace differences. - There is support for setting a "log level". This is not meant to be user-facing, but can be used internally to implement debug or verbose modes. - Lazy argument evaluation, so no significant overhead if logging at some level is disabled. - Some color in the messages, similar to gcc and clang. Set PG_COLOR=auto to try it out. Some colors are predefined, but can be customized by setting PG_COLORS. - Common files (common/, fe_utils/, etc.) can handle logging much more simply by just using one API without worrying too much about the context of the calling program, requiring callbacks, or having to pass "progname" around everywhere. - Some programs called setvbuf() to make sure that stderr is unbuffered, even on Windows. But not all programs did that. This is now done centrally. Soft goals: - Reduces vertical space use and visual complexity of error reporting in the source code. - Encourages more deliberate classification of messages. For example, in some cases it wasn't clear without analyzing the surrounding code whether a message was meant as an error or just an info. - Concepts and terms are vaguely aligned with popular logging frameworks such as log4j and Python logging. This is all just about printing stuff out. Nothing affects program flow (e.g., fatal exits). The uses are just too varied to do that. Some existing code had wrappers that do some kind of print-and-exit, and I adapted those. I tried to keep the output mostly the same, but there is a lot of historical baggage to unwind and special cases to consider, and I might not always have succeeded. One significant change is that pg_rewind used to write all error messages to stdout. That is now changed to stderr. Reviewed-by: Donald Dong <xdong@csumb.edu> Reviewed-by: Arthur Zakirov <a.zakirov@postgrespro.ru> Discussion: https://www.postgresql.org/message-id/flat/6a609b43-4f57-7348-6480-bd022f924310@2ndquadrant.com
2019-04-01 14:24:37 +02:00
if (fsync_fname(dir_data->basedir, true) != 0)
return false;
}
return true;
}
WalWriteMethod *
CreateWalDirectoryMethod(const char *basedir, int compression, bool sync)
{
WalWriteMethod *method;
method = pg_malloc0(sizeof(WalWriteMethod));
method->open_for_write = dir_open_for_write;
method->write = dir_write;
method->get_current_pos = dir_get_current_pos;
method->get_file_size = dir_get_file_size;
method->close = dir_close;
method->sync = dir_sync;
method->existsfile = dir_existsfile;
method->finish = dir_finish;
method->getlasterror = dir_getlasterror;
dir_data = pg_malloc0(sizeof(DirectoryMethodData));
dir_data->compression = compression;
dir_data->basedir = pg_strdup(basedir);
dir_data->sync = sync;
return method;
}
void
FreeWalDirectoryMethod(void)
{
pg_free(dir_data->basedir);
pg_free(dir_data);
}
/*-------------------------------------------------------------------------
* WalTarMethod - write wal to a tar file containing pg_wal contents
*-------------------------------------------------------------------------
*/
typedef struct TarMethodFile
{
off_t ofs_start; /* Where does the *header* for this file start */
off_t currpos;
char header[512];
char *pathname;
size_t pad_to_size;
} TarMethodFile;
typedef struct TarMethodData
{
char *tarfilename;
int fd;
int compression;
bool sync;
TarMethodFile *currentfile;
char lasterror[1024];
#ifdef HAVE_LIBZ
z_streamp zp;
void *zlibOut;
#endif
} TarMethodData;
static TarMethodData *tar_data = NULL;
#define tar_clear_error() tar_data->lasterror[0] = '\0'
#define tar_set_error(msg) strlcpy(tar_data->lasterror, _(msg), sizeof(tar_data->lasterror))
static const char *
tar_getlasterror(void)
{
/*
* If a custom error is set, return that one. Otherwise, assume errno is
* set and return that one.
*/
if (tar_data->lasterror[0])
return tar_data->lasterror;
return strerror(errno);
}
#ifdef HAVE_LIBZ
static bool
tar_write_compressed_data(void *buf, size_t count, bool flush)
{
tar_data->zp->next_in = buf;
tar_data->zp->avail_in = count;
while (tar_data->zp->avail_in || flush)
{
int r;
r = deflate(tar_data->zp, flush ? Z_FINISH : Z_NO_FLUSH);
if (r == Z_STREAM_ERROR)
{
2017-08-05 00:31:01 +02:00
tar_set_error("could not compress data");
return false;
}
if (tar_data->zp->avail_out < ZLIB_OUT_SIZE)
{
size_t len = ZLIB_OUT_SIZE - tar_data->zp->avail_out;
errno = 0;
if (write(tar_data->fd, tar_data->zlibOut, len) != len)
{
/*
* If write didn't set errno, assume problem is no disk space.
*/
if (errno == 0)
errno = ENOSPC;
return false;
}
tar_data->zp->next_out = tar_data->zlibOut;
tar_data->zp->avail_out = ZLIB_OUT_SIZE;
}
if (r == Z_STREAM_END)
break;
}
if (flush)
{
/* Reset the stream for writing */
if (deflateReset(tar_data->zp) != Z_OK)
{
2017-08-05 00:31:01 +02:00
tar_set_error("could not reset compression stream");
return false;
}
}
return true;
}
#endif
static ssize_t
tar_write(Walfile f, const void *buf, size_t count)
{
ssize_t r;
Assert(f != NULL);
tar_clear_error();
/* Tarfile will always be positioned at the end */
if (!tar_data->compression)
{
r = write(tar_data->fd, buf, count);
if (r > 0)
((TarMethodFile *) f)->currpos += r;
return r;
}
#ifdef HAVE_LIBZ
else
{
if (!tar_write_compressed_data(unconstify(void *, buf), count, false))
return -1;
((TarMethodFile *) f)->currpos += count;
return count;
}
#else
else
/* Can't happen - compression enabled with no libz */
return -1;
#endif
}
static bool
tar_write_padding_data(TarMethodFile *f, size_t bytes)
{
PGAlignedXLogBlock zerobuf;
size_t bytesleft = bytes;
memset(zerobuf.data, 0, XLOG_BLCKSZ);
while (bytesleft)
{
size_t bytestowrite = Min(bytesleft, XLOG_BLCKSZ);
ssize_t r = tar_write(f, zerobuf.data, bytestowrite);
if (r < 0)
return false;
bytesleft -= r;
}
return true;
}
static Walfile
tar_open_for_write(const char *pathname, const char *temp_suffix, size_t pad_to_size)
{
int save_errno;
static char tmppath[MAXPGPATH];
tar_clear_error();
if (tar_data->fd < 0)
{
/*
* We open the tar file only when we first try to write to it.
*/
tar_data->fd = open(tar_data->tarfilename,
O_WRONLY | O_CREAT | PG_BINARY,
pg_file_create_mode);
if (tar_data->fd < 0)
return NULL;
#ifdef HAVE_LIBZ
if (tar_data->compression)
{
tar_data->zp = (z_streamp) pg_malloc(sizeof(z_stream));
tar_data->zp->zalloc = Z_NULL;
tar_data->zp->zfree = Z_NULL;
tar_data->zp->opaque = Z_NULL;
tar_data->zp->next_out = tar_data->zlibOut;
tar_data->zp->avail_out = ZLIB_OUT_SIZE;
/*
* Initialize deflation library. Adding the magic value 16 to the
* default 15 for the windowBits parameter makes the output be
* gzip instead of zlib.
*/
if (deflateInit2(tar_data->zp, tar_data->compression, Z_DEFLATED, 15 + 16, 8, Z_DEFAULT_STRATEGY) != Z_OK)
{
pg_free(tar_data->zp);
tar_data->zp = NULL;
2017-08-05 00:31:01 +02:00
tar_set_error("could not initialize compression library");
return NULL;
}
}
#endif
/* There's no tar header itself, the file starts with regular files */
}
Assert(tar_data->currentfile == NULL);
if (tar_data->currentfile != NULL)
{
2017-08-05 00:31:01 +02:00
tar_set_error("implementation error: tar files can't have more than one open file");
return NULL;
}
tar_data->currentfile = pg_malloc0(sizeof(TarMethodFile));
snprintf(tmppath, sizeof(tmppath), "%s%s",
pathname, temp_suffix ? temp_suffix : "");
/* Create a header with size set to 0 - we will fill out the size on close */
if (tarCreateHeader(tar_data->currentfile->header, tmppath, NULL, 0, S_IRUSR | S_IWUSR, 0, 0, time(NULL)) != TAR_OK)
{
pg_free(tar_data->currentfile);
tar_data->currentfile = NULL;
tar_set_error("could not create tar header");
return NULL;
}
#ifdef HAVE_LIBZ
if (tar_data->compression)
{
/* Flush existing data */
if (!tar_write_compressed_data(NULL, 0, true))
return NULL;
/* Turn off compression for header */
if (deflateParams(tar_data->zp, 0, 0) != Z_OK)
{
2017-08-05 00:31:01 +02:00
tar_set_error("could not change compression parameters");
return NULL;
}
}
#endif
tar_data->currentfile->ofs_start = lseek(tar_data->fd, 0, SEEK_CUR);
if (tar_data->currentfile->ofs_start == -1)
{
save_errno = errno;
pg_free(tar_data->currentfile);
tar_data->currentfile = NULL;
errno = save_errno;
return NULL;
}
tar_data->currentfile->currpos = 0;
if (!tar_data->compression)
{
errno = 0;
if (write(tar_data->fd, tar_data->currentfile->header, 512) != 512)
{
save_errno = errno;
pg_free(tar_data->currentfile);
tar_data->currentfile = NULL;
/* if write didn't set errno, assume problem is no disk space */
errno = save_errno ? save_errno : ENOSPC;
return NULL;
}
}
#ifdef HAVE_LIBZ
else
{
/* Write header through the zlib APIs but with no compression */
if (!tar_write_compressed_data(tar_data->currentfile->header, 512, true))
return NULL;
/* Re-enable compression for the rest of the file */
if (deflateParams(tar_data->zp, tar_data->compression, 0) != Z_OK)
{
2017-08-05 00:31:01 +02:00
tar_set_error("could not change compression parameters");
return NULL;
}
}
#endif
tar_data->currentfile->pathname = pg_strdup(pathname);
/*
* Uncompressed files are padded on creation, but for compression we can't
* do that
*/
if (pad_to_size)
{
tar_data->currentfile->pad_to_size = pad_to_size;
if (!tar_data->compression)
{
/* Uncompressed, so pad now */
tar_write_padding_data(tar_data->currentfile, pad_to_size);
/* Seek back to start */
if (lseek(tar_data->fd, tar_data->currentfile->ofs_start + 512, SEEK_SET) != tar_data->currentfile->ofs_start + 512)
return NULL;
tar_data->currentfile->currpos = 0;
}
}
return tar_data->currentfile;
}
static ssize_t
tar_get_file_size(const char *pathname)
{
tar_clear_error();
/* Currently not used, so not supported */
errno = ENOSYS;
return -1;
}
static off_t
tar_get_current_pos(Walfile f)
{
Assert(f != NULL);
tar_clear_error();
return ((TarMethodFile *) f)->currpos;
}
static int
tar_sync(Walfile f)
{
Assert(f != NULL);
tar_clear_error();
if (!tar_data->sync)
return 0;
/*
* Always sync the whole tarfile, because that's all we can do. This makes
* no sense on compressed files, so just ignore those.
*/
if (tar_data->compression)
return 0;
return fsync(tar_data->fd);
}
static int
tar_close(Walfile f, WalCloseMethod method)
{
ssize_t filesize;
int padding;
TarMethodFile *tf = (TarMethodFile *) f;
Assert(f != NULL);
tar_clear_error();
if (method == CLOSE_UNLINK)
{
if (tar_data->compression)
{
tar_set_error("unlink not supported with compression");
return -1;
}
/*
* Unlink the file that we just wrote to the tar. We do this by
* truncating it to the start of the header. This is safe as we only
* allow writing of the very last file.
*/
if (ftruncate(tar_data->fd, tf->ofs_start) != 0)
return -1;
pg_free(tf->pathname);
pg_free(tf);
tar_data->currentfile = NULL;
return 0;
}
/*
* Pad the file itself with zeroes if necessary. Note that this is
* different from the tar format padding -- this is the padding we asked
* for when the file was opened.
*/
if (tf->pad_to_size)
{
if (tar_data->compression)
{
/*
* A compressed tarfile is padded on close since we cannot know
* the size of the compressed output until the end.
*/
size_t sizeleft = tf->pad_to_size - tf->currpos;
if (sizeleft)
{
if (!tar_write_padding_data(tf, sizeleft))
return -1;
}
}
else
{
/*
* An uncompressed tarfile was padded on creation, so just adjust
* the current position as if we seeked to the end.
*/
tf->currpos = tf->pad_to_size;
}
}
/*
* Get the size of the file, and pad the current data up to the nearest
* 512 byte boundary.
*/
filesize = tar_get_current_pos(f);
padding = ((filesize + 511) & ~511) - filesize;
if (padding)
{
char zerobuf[512];
MemSet(zerobuf, 0, padding);
if (tar_write(f, zerobuf, padding) != padding)
return -1;
}
#ifdef HAVE_LIBZ
if (tar_data->compression)
{
/* Flush the current buffer */
if (!tar_write_compressed_data(NULL, 0, true))
{
errno = EINVAL;
return -1;
}
}
#endif
/*
* Now go back and update the header with the correct filesize and
* possibly also renaming the file. We overwrite the entire current header
* when done, including the checksum.
*/
print_tar_number(&(tf->header[124]), 12, filesize);
if (method == CLOSE_NORMAL)
/*
* We overwrite it with what it was before if we have no tempname,
* since we're going to write the buffer anyway.
*/
strlcpy(&(tf->header[0]), tf->pathname, 100);
print_tar_number(&(tf->header[148]), 8, tarChecksum(((TarMethodFile *) f)->header));
if (lseek(tar_data->fd, tf->ofs_start, SEEK_SET) != ((TarMethodFile *) f)->ofs_start)
return -1;
if (!tar_data->compression)
{
errno = 0;
if (write(tar_data->fd, tf->header, 512) != 512)
{
/* if write didn't set errno, assume problem is no disk space */
if (errno == 0)
errno = ENOSPC;
return -1;
}
}
#ifdef HAVE_LIBZ
else
{
/* Turn off compression */
if (deflateParams(tar_data->zp, 0, 0) != Z_OK)
{
2017-08-05 00:31:01 +02:00
tar_set_error("could not change compression parameters");
return -1;
}
/* Overwrite the header, assuming the size will be the same */
if (!tar_write_compressed_data(tar_data->currentfile->header, 512, true))
return -1;
/* Turn compression back on */
if (deflateParams(tar_data->zp, tar_data->compression, 0) != Z_OK)
{
2017-08-05 00:31:01 +02:00
tar_set_error("could not change compression parameters");
return -1;
}
}
#endif
/* Move file pointer back down to end, so we can write the next file */
if (lseek(tar_data->fd, 0, SEEK_END) < 0)
return -1;
/* Always fsync on close, so the padding gets fsynced */
if (tar_sync(f) < 0)
exit(1);
/* Clean up and done */
pg_free(tf->pathname);
pg_free(tf);
tar_data->currentfile = NULL;
return 0;
}
static bool
tar_existsfile(const char *pathname)
{
tar_clear_error();
/* We only deal with new tarfiles, so nothing externally created exists */
return false;
}
static bool
tar_finish(void)
{
char zerobuf[1024];
tar_clear_error();
if (tar_data->currentfile)
{
if (tar_close(tar_data->currentfile, CLOSE_NORMAL) != 0)
return false;
}
/* A tarfile always ends with two empty blocks */
MemSet(zerobuf, 0, sizeof(zerobuf));
if (!tar_data->compression)
{
errno = 0;
if (write(tar_data->fd, zerobuf, sizeof(zerobuf)) != sizeof(zerobuf))
{
/* if write didn't set errno, assume problem is no disk space */
if (errno == 0)
errno = ENOSPC;
return false;
}
}
#ifdef HAVE_LIBZ
else
{
if (!tar_write_compressed_data(zerobuf, sizeof(zerobuf), false))
return false;
/* Also flush all data to make sure the gzip stream is finished */
tar_data->zp->next_in = NULL;
tar_data->zp->avail_in = 0;
while (true)
{
int r;
r = deflate(tar_data->zp, Z_FINISH);
if (r == Z_STREAM_ERROR)
{
2017-08-05 00:31:01 +02:00
tar_set_error("could not compress data");
return false;
}
if (tar_data->zp->avail_out < ZLIB_OUT_SIZE)
{
size_t len = ZLIB_OUT_SIZE - tar_data->zp->avail_out;
errno = 0;
if (write(tar_data->fd, tar_data->zlibOut, len) != len)
{
/*
* If write didn't set errno, assume problem is no disk
* space.
*/
if (errno == 0)
errno = ENOSPC;
return false;
}
}
if (r == Z_STREAM_END)
break;
}
if (deflateEnd(tar_data->zp) != Z_OK)
{
2017-08-05 00:31:01 +02:00
tar_set_error("could not close compression stream");
return false;
}
}
#endif
/* sync the empty blocks as well, since they're after the last file */
if (tar_data->sync)
{
if (fsync(tar_data->fd) != 0)
return false;
}
if (close(tar_data->fd) != 0)
return false;
tar_data->fd = -1;
if (tar_data->sync)
{
Unified logging system for command-line programs This unifies the various ad hoc logging (message printing, error printing) systems used throughout the command-line programs. Features: - Program name is automatically prefixed. - Message string does not end with newline. This removes a common source of inconsistencies and omissions. - Additionally, a final newline is automatically stripped, simplifying use of PQerrorMessage() etc., another common source of mistakes. - I converted error message strings to use %m where possible. - As a result of the above several points, more translatable message strings can be shared between different components and between frontends and backend, without gratuitous punctuation or whitespace differences. - There is support for setting a "log level". This is not meant to be user-facing, but can be used internally to implement debug or verbose modes. - Lazy argument evaluation, so no significant overhead if logging at some level is disabled. - Some color in the messages, similar to gcc and clang. Set PG_COLOR=auto to try it out. Some colors are predefined, but can be customized by setting PG_COLORS. - Common files (common/, fe_utils/, etc.) can handle logging much more simply by just using one API without worrying too much about the context of the calling program, requiring callbacks, or having to pass "progname" around everywhere. - Some programs called setvbuf() to make sure that stderr is unbuffered, even on Windows. But not all programs did that. This is now done centrally. Soft goals: - Reduces vertical space use and visual complexity of error reporting in the source code. - Encourages more deliberate classification of messages. For example, in some cases it wasn't clear without analyzing the surrounding code whether a message was meant as an error or just an info. - Concepts and terms are vaguely aligned with popular logging frameworks such as log4j and Python logging. This is all just about printing stuff out. Nothing affects program flow (e.g., fatal exits). The uses are just too varied to do that. Some existing code had wrappers that do some kind of print-and-exit, and I adapted those. I tried to keep the output mostly the same, but there is a lot of historical baggage to unwind and special cases to consider, and I might not always have succeeded. One significant change is that pg_rewind used to write all error messages to stdout. That is now changed to stderr. Reviewed-by: Donald Dong <xdong@csumb.edu> Reviewed-by: Arthur Zakirov <a.zakirov@postgrespro.ru> Discussion: https://www.postgresql.org/message-id/flat/6a609b43-4f57-7348-6480-bd022f924310@2ndquadrant.com
2019-04-01 14:24:37 +02:00
if (fsync_fname(tar_data->tarfilename, false) != 0)
return false;
Unified logging system for command-line programs This unifies the various ad hoc logging (message printing, error printing) systems used throughout the command-line programs. Features: - Program name is automatically prefixed. - Message string does not end with newline. This removes a common source of inconsistencies and omissions. - Additionally, a final newline is automatically stripped, simplifying use of PQerrorMessage() etc., another common source of mistakes. - I converted error message strings to use %m where possible. - As a result of the above several points, more translatable message strings can be shared between different components and between frontends and backend, without gratuitous punctuation or whitespace differences. - There is support for setting a "log level". This is not meant to be user-facing, but can be used internally to implement debug or verbose modes. - Lazy argument evaluation, so no significant overhead if logging at some level is disabled. - Some color in the messages, similar to gcc and clang. Set PG_COLOR=auto to try it out. Some colors are predefined, but can be customized by setting PG_COLORS. - Common files (common/, fe_utils/, etc.) can handle logging much more simply by just using one API without worrying too much about the context of the calling program, requiring callbacks, or having to pass "progname" around everywhere. - Some programs called setvbuf() to make sure that stderr is unbuffered, even on Windows. But not all programs did that. This is now done centrally. Soft goals: - Reduces vertical space use and visual complexity of error reporting in the source code. - Encourages more deliberate classification of messages. For example, in some cases it wasn't clear without analyzing the surrounding code whether a message was meant as an error or just an info. - Concepts and terms are vaguely aligned with popular logging frameworks such as log4j and Python logging. This is all just about printing stuff out. Nothing affects program flow (e.g., fatal exits). The uses are just too varied to do that. Some existing code had wrappers that do some kind of print-and-exit, and I adapted those. I tried to keep the output mostly the same, but there is a lot of historical baggage to unwind and special cases to consider, and I might not always have succeeded. One significant change is that pg_rewind used to write all error messages to stdout. That is now changed to stderr. Reviewed-by: Donald Dong <xdong@csumb.edu> Reviewed-by: Arthur Zakirov <a.zakirov@postgrespro.ru> Discussion: https://www.postgresql.org/message-id/flat/6a609b43-4f57-7348-6480-bd022f924310@2ndquadrant.com
2019-04-01 14:24:37 +02:00
if (fsync_parent_path(tar_data->tarfilename) != 0)
return false;
}
return true;
}
WalWriteMethod *
CreateWalTarMethod(const char *tarbase, int compression, bool sync)
{
WalWriteMethod *method;
const char *suffix = (compression != 0) ? ".tar.gz" : ".tar";
method = pg_malloc0(sizeof(WalWriteMethod));
method->open_for_write = tar_open_for_write;
method->write = tar_write;
method->get_current_pos = tar_get_current_pos;
method->get_file_size = tar_get_file_size;
method->close = tar_close;
method->sync = tar_sync;
method->existsfile = tar_existsfile;
method->finish = tar_finish;
method->getlasterror = tar_getlasterror;
tar_data = pg_malloc0(sizeof(TarMethodData));
tar_data->tarfilename = pg_malloc0(strlen(tarbase) + strlen(suffix) + 1);
sprintf(tar_data->tarfilename, "%s%s", tarbase, suffix);
tar_data->fd = -1;
tar_data->compression = compression;
tar_data->sync = sync;
#ifdef HAVE_LIBZ
if (compression)
tar_data->zlibOut = (char *) pg_malloc(ZLIB_OUT_SIZE + 1);
#endif
return method;
}
void
FreeWalTarMethod(void)
{
pg_free(tar_data->tarfilename);
#ifdef HAVE_LIBZ
if (tar_data->compression)
pg_free(tar_data->zlibOut);
#endif
pg_free(tar_data);
}