Hi,
thanks for the additional review!
Am Donnerstag, den 14.03.2019, 11:54 +0900 schrieb Kyotaro HORIGUCHI:
> At Wed, 13 Mar 2019 16:25:15 +0900, Michael Paquier <mich...@paquier.xyz>
> wrote in <20190313072515.gb2...@paquier.xyz>
> > On Wed, Mar 13, 2019 at 07:22:28AM +0100, Fabien COELHO wrote:
> > > Does not apply because of the renaming committed by Michaël.
> > >
> > > Could you rebase?
> >
> > This stuff touches pg_checksums.c, so you may want to wait one day or
> > two to avoid extra work... I think that I'll be able to finish the
> > addition of the --enable and --disable switches by then.
I have rebased it now.
> > + /* Make sure we report at most once every tenth of a second */
> > + if ((INSTR_TIME_GET_MILLISEC(now) -
> > + INSTR_TIME_GET_MILLISEC(last_progress_update) < 100) &&
> > !force)
>
> I'm not a skilled gamer and 10Hz update is a bit hard for my
> eyes:p The second hand of old clocks ticks at 4Hz. I think it is
> enough frequently.
Ok.
> > 841/1516 MB (55%, 700 MB/s)
>
> Differently from other prgress reporting project, estimating ETC
> (estimated time to completion), which is in the most important,
> is quite easy. (pgbench does that.) And I'd like to see a
> progress bar there but it might be overdoing. I'm not sure let
> the current size occupy a part of screen width is needed. I
> don't think total size is needed to be fixed in MB and I think at
> most four or five digits are enough. (That is the same for
> speed.)
>
> If the all of aboves are involved, the line would look as the
> follows.
>
> [======================= ] ( 63% of 12.53 GB, 179 MB/s, ETC 26s)
>
> # Note that this is just an opinion.
>
> (pg_checksum runs fast at the beginning so ETC behaves somewhat
> strange in the meanwhile.)
I haven't changed that for now as it seems to be a bit more involved.
I'd like to hear other opinions on whether that is worthwhile?
> > +#define MEGABYTES 1048576
>
> 1024 * 1024 is preferable than that many digits.
Good point, changed that way.
> > + /* we handle SIGUSR1 only, and toggle the value of show_progress */
> > + if (signum == SIGUSR1)
> > + show_progress = !show_progress;
>
> SIGUSR1 *toggles* progress.
Not sure what you mean here, are you saying it should be called
toggle_progress and not show_progress? I think it was like that
originally but got changed due to review comments.
> A garbage line is left alone after turning it off. It would be
> erasable. I'm not sure which is better, though.
How about just printing a "\n" in the signal handler if we are in a
terminal? I think erasing the line might get too clever and will be a
portability hazard? I have done that now in the attached patch, let me
know what you think.
Michael
--
Michael Banck
Projektleiter / Senior Berater
Tel.: +49 2166 9901-171
Fax: +49 2166 9901-100
Email: michael.ba...@credativ.de
credativ GmbH, HRB Mönchengladbach 12080
USt-ID-Nummer: DE204566209
Trompeterallee 108, 41189 Mönchengladbach
Geschäftsführung: Dr. Michael Meskes, Jörg Folz, Sascha Heuer
Unser Umgang mit personenbezogenen Daten unterliegt
folgenden Bestimmungen: https://www.credativ.de/datenschutz
diff --git a/doc/src/sgml/ref/pg_checksums.sgml b/doc/src/sgml/ref/pg_checksums.sgml
index 6a47dda683..c790170953 100644
--- a/doc/src/sgml/ref/pg_checksums.sgml
+++ b/doc/src/sgml/ref/pg_checksums.sgml
@@ -80,6 +80,19 @@ PostgreSQL documentation
</varlistentry>
<varlistentry>
+ <term><option>-P</option></term>
+ <term><option>--progress</option></term>
+ <listitem>
+ <para>
+ Enable progress reporting. Turning this on will deliver an approximate
+ progress report during the checksum verification. Sending the
+ <systemitem>SIGUSR1</systemitem> signal will toggle progress reporting
+ on or off during the verification run (not available on Windows).
+ </para>
+ </listitem>
+ </varlistentry>
+
+ <varlistentry>
<term><option>-V</option></term>
<term><option>--version</option></term>
<listitem>
diff --git a/src/bin/pg_checksums/pg_checksums.c b/src/bin/pg_checksums/pg_checksums.c
index b7ebc11017..a771f67c18 100644
--- a/src/bin/pg_checksums/pg_checksums.c
+++ b/src/bin/pg_checksums/pg_checksums.c
@@ -14,6 +14,8 @@
#include "postgres_fe.h"
#include <dirent.h>
+#include <signal.h>
+#include <time.h>
#include <sys/stat.h>
#include <unistd.h>
@@ -21,6 +23,7 @@
#include "common/controldata_utils.h"
#include "getopt_long.h"
#include "pg_getopt.h"
+#include "portability/instr_time.h"
#include "storage/bufpage.h"
#include "storage/checksum.h"
#include "storage/checksum_impl.h"
@@ -34,9 +37,18 @@ static ControlFileData *ControlFile;
static char *only_relfilenode = NULL;
static bool verbose = false;
+static bool show_progress = false;
static const char *progname;
+/*
+ * Progress status information.
+ */
+int64 total_size = 0;
+int64 current_size = 0;
+instr_time last_progress_update;
+instr_time scan_started;
+
static void
usage(void)
{
@@ -47,6 +59,7 @@ usage(void)
printf(_(" [-D, --pgdata=]DATADIR data directory\n"));
printf(_(" -v, --verbose output verbose messages\n"));
printf(_(" -r RELFILENODE check only relation with specified relfilenode\n"));
+ printf(_(" -P, --progress show progress information\n"));
printf(_(" -V, --version output version information, then exit\n"));
printf(_(" -?, --help show this help, then exit\n"));
printf(_("\nIf no data directory (DATADIR) is specified, "
@@ -71,6 +84,83 @@ static const char *const skip[] = {
NULL,
};
+static void
+toggle_progress_report(int signum)
+{
+
+ /* we handle SIGUSR1 only, and toggle the value of show_progress */
+ if (signum == SIGUSR1)
+ {
+ /*
+ * Skip to the next line in order to avoid overwriting half of
+ * the progress report line with the final output.
+ */
+ if (show_progress && isatty(fileno(stderr)))
+ fprintf(stderr, "\n");
+ show_progress = !show_progress;
+ }
+}
+
+/*
+ * Report current progress status. Parts borrowed from
+ * PostgreSQLs' src/bin/pg_basebackup.c
+ */
+static void
+report_progress(bool force)
+{
+ instr_time now;
+ double elapsed;
+ double total_percent;
+ int64 current_speed;
+
+ char totalstr[32];
+ char currentstr[32];
+ char currspeedstr[32];
+
+ INSTR_TIME_SET_CURRENT(now);
+
+ /* Make sure we report at most once every 400 milliseconds */
+ if ((INSTR_TIME_GET_MILLISEC(now) -
+ INSTR_TIME_GET_MILLISEC(last_progress_update) < 400) && !force)
+ return;
+
+ /* Save current time */
+ last_progress_update = now;
+
+ /* Elapsed time in milliseconds since start of scan */
+ elapsed = (INSTR_TIME_GET_MILLISEC(now) -
+ INSTR_TIME_GET_MILLISEC(scan_started));
+
+ /* Calculate current percent done */
+ total_percent = total_size ? 100.0 * current_size / total_size : 0.0;
+
+ /* Don't display larger than 100% */
+ if (total_percent > 100)
+ total_percent = 100;
+
+#define MEGABYTES (1024 * 1024)
+
+ /* The same for total size */
+ if (current_size > total_size)
+ total_size = current_size / MEGABYTES;
+
+ /* Calculate current speed */
+ current_speed = (int64)(current_size / MEGABYTES) / (elapsed / 1000);
+
+ snprintf(totalstr, sizeof(totalstr), INT64_FORMAT,
+ total_size / MEGABYTES);
+ snprintf(currentstr, sizeof(currentstr), INT64_FORMAT,
+ current_size / MEGABYTES);
+ snprintf(currspeedstr, sizeof(currspeedstr), INT64_FORMAT,
+ current_speed);
+
+ fprintf(stderr, "%s/%s MB (%d%%, %s MB/s)",
+ currentstr, totalstr, (int)total_percent, currspeedstr);
+
+ /* Stay on the same line if reporting to a terminal */
+ fprintf(stderr, isatty(fileno(stderr)) ? "\r" : "\n");
+}
+
static bool
skipfile(const char *fn)
{
@@ -121,6 +211,7 @@ scan_file(const char *fn, BlockNumber segmentno)
continue;
csum = pg_checksum_page(buf.data, blockno + segmentno * RELSEG_SIZE);
+ current_size += r;
if (csum != header->pd_checksum)
{
if (ControlFile->data_checksum_version == PG_DATA_CHECKSUM_VERSION)
@@ -128,6 +219,9 @@ scan_file(const char *fn, BlockNumber segmentno)
progname, fn, blockno, csum, header->pd_checksum);
badblocks++;
}
+
+ if (show_progress)
+ report_progress(false);
}
if (verbose)
@@ -137,9 +231,10 @@ scan_file(const char *fn, BlockNumber segmentno)
close(f);
}
-static void
-scan_directory(const char *basedir, const char *subdir)
+static int64
+scan_directory(const char *basedir, const char *subdir, bool sizeonly)
{
+ int64 dirsize = 0;
char path[MAXPGPATH];
DIR *dir;
struct dirent *de;
@@ -218,16 +313,20 @@ scan_directory(const char *basedir, const char *subdir)
/* Relfilenode not to be included */
continue;
- scan_file(fn, segmentno);
+ dirsize += st.st_size;
+
+ if (!sizeonly)
+ scan_file(fn, segmentno);
}
#ifndef WIN32
else if (S_ISDIR(st.st_mode) || S_ISLNK(st.st_mode))
#else
else if (S_ISDIR(st.st_mode) || pgwin32_is_junction(fn))
#endif
- scan_directory(path, de->d_name);
+ dirsize += scan_directory(path, de->d_name, sizeonly);
}
closedir(dir);
+ return dirsize;
}
int
@@ -235,6 +334,7 @@ main(int argc, char *argv[])
{
static struct option long_options[] = {
{"pgdata", required_argument, NULL, 'D'},
+ {"progress", no_argument, NULL, 'P'},
{"verbose", no_argument, NULL, 'v'},
{NULL, 0, NULL, 0}
};
@@ -262,7 +362,7 @@ main(int argc, char *argv[])
}
}
- while ((c = getopt_long(argc, argv, "D:r:v", long_options, &option_index)) != -1)
+ while ((c = getopt_long(argc, argv, "D:r:vP", long_options, &option_index)) != -1)
{
switch (c)
{
@@ -280,6 +380,9 @@ main(int argc, char *argv[])
}
only_relfilenode = pstrdup(optarg);
break;
+ case 'P':
+ show_progress = true;
+ break;
default:
fprintf(stderr, _("Try \"%s --help\" for more information.\n"), progname);
exit(1);
@@ -349,10 +452,45 @@ main(int argc, char *argv[])
exit(1);
}
+
+#ifndef WIN32
+ /*
+ * Assign SIGUSR1 signal handler to toggle progress status information
+ */
+ pqsignal(SIGUSR1, toggle_progress_report);
+#endif
+
+ /*
+ * As progress status information may be requested even after start of
+ * operation, we need to scan the directory tree(s) twice, once to get
+ * the idea how much data we need to scan and finally to do the real
+ * legwork.
+ */
+ total_size = scan_directory(DataDir, "global", true);
+ total_size += scan_directory(DataDir, "base", true);
+ total_size += scan_directory(DataDir, "pg_tblspc", true);
+
+ /*
+ * Remember start time. Required to calculate the current speed in
+ * report_progress()
+ */
+ INSTR_TIME_SET_CURRENT(scan_started);
+
/* Scan all files */
- scan_directory(DataDir, "global");
- scan_directory(DataDir, "base");
- scan_directory(DataDir, "pg_tblspc");
+ scan_directory(DataDir, "global", false);
+ scan_directory(DataDir, "base", false);
+ scan_directory(DataDir, "pg_tblspc", false);
+
+ /*
+ * Done. Move to next line in case progress information was shown.
+ * Otherwise we clutter the summary output.
+ */
+ if (show_progress)
+ {
+ report_progress(true);
+ if (isatty(fileno(stderr)))
+ fprintf(stderr, "\n");
+ }
printf(_("Checksum scan completed\n"));
printf(_("Data checksum version: %d\n"), ControlFile->data_checksum_version);