Commit Graph

32 Commits

Author SHA1 Message Date
Bruce Momjian f9bb944cd8 pg_test_fsync: add C comment about direct I/O and write size failure
Report from Marti Raudsepp
2014-02-12 15:38:29 -05:00
Bruce Momjian 43f6388931 pg_test_fsync: expand ops/sec display
Expand ops/sec by two digits to maintain alignment on servers with fast
I/O subsystems, e.g. can now display < 10M ops/sec with consistent
alignment.
2013-08-02 00:45:19 -04:00
Heikki Linnakangas b5ed21998c Fix pg_test_fsync, broken by xloginsert scaling patch.
I didn't realize that ALIGNOF_XLOG_BUFFER was used in pg_test_fsync.
2013-07-08 13:04:07 +03:00
Bruce Momjian 06b804377c Remove undocumented -h (help) option
The -h option was not supported by many tools, and not documented, so
remove them for consistency from pg_upgrade, pg_test_fsync, and
pg_test_timing.
2013-07-01 12:40:33 -04:00
Bruce Momjian 095018bc32 pg_test_fsync: update output to show usecs/op clearer 2013-05-02 10:27:12 -04:00
Tom Lane 3c07fbf40b Use pqsignal() in contrib programs rather than calling signal(2) directly.
The semantics of signal(2) are more variable than one could wish; in
particular, on strict-POSIX platforms the signal handler will be reset
to SIG_DFL when the signal is delivered.  This demonstrably breaks
pg_test_fsync's use of SIGALRM.  The other changes I made are not
absolutely necessary today, because the called handlers all exit the
program anyway.  But it seems like a good general practice to use
pqsignal() exclusively in Postgres code, now that we have it available
everywhere.
2013-03-17 16:09:47 -04:00
Tom Lane 70ec2f8f43 Improve the documentation about commit_delay.
Clarify the docs explaining what commit_delay does, and add a
recommendation about a useful value for it, namely half of the single-page
fsync time reported by pg_test_fsync.  This is informed by testing of
the new-in-9.3 implementation of commit_delay; in prior versions it
was far harder to arrive at a useful setting.

In passing, do some wordsmithing and markup-fixing in the same general
area.

Also, change pg_test_fsync's default time-per-test from 2 seconds to 5.
The old value was about the minimum at which the results could be taken
seriously at all, and so seems a tad optimistic as a default.

Peter Geoghegan, reviewed by Noah Misch; some additional editing by me
2013-03-15 17:41:47 -04:00
Bruce Momjian bd9c8e741b Move long_options structures to the top of main() functions, for
consistency.

Per suggestion from Tom.
2012-11-30 14:49:55 -05:00
Tom Lane 427fd88552 Fix unportable format string.
Per compiler warning.
2012-10-12 13:22:11 -04:00
Simon Riggs 82e429794b Add microsecs/op display to pg_test_fsync utility
e.g. fsync  2103.613 ops/sec (   475 microsecs/op)

Peter Geoghegan
2012-10-09 08:15:23 +01:00
Bruce Momjian 927d61eeff Run pgindent on 9.2 source tree in preparation for first 9.3
commit-fest.
2012-06-10 15:20:04 -04:00
Bruce Momjian 388c2f9325 Fix pg_test_fsync Win32 problems reported by the build farm; add
comments about the alarm method used on Win32.
2012-02-15 07:10:48 -05:00
Bruce Momjian 7ae2db1d1e Try to get pg_test_thread to compile on Windows by using a sleeper
thread.
2012-02-14 21:53:17 -05:00
Bruce Momjian 0a8396e35e Change contrib/pg_test_fsync to control tests in terms of seconds per
test, rather than a number of test cycles.  Changes -o/cycles option to
-s/seconds.
2012-02-14 11:10:03 -05:00
Bruce Momjian 2bbd88f8f8 Improve labeling of pg_test_fsync open_sync test output. 2012-02-13 21:48:51 -05:00
Robert Haas ab7d4213f9 If pg_test_fsync is interrupted, clean up the temp file.
Marti Raudsepp, with additional paranoia by me.
2011-12-09 15:06:46 -05:00
Peter Eisentraut 3a8dcd9a1f Put pg_test_fsync --help and --version output in line with conventions
Foremost, it should go to stdout.
2011-05-10 21:34:26 +03:00
Peter Eisentraut 001cbb145f Avoid unused variable warnings for certain configurations 2011-04-19 20:01:51 +03:00
Bruce Momjian bf50caf105 pgindent run before PG 9.1 beta 1. 2011-04-10 11:42:00 -04:00
Bruce Momjian 303b7fcd10 Modify pg_test_fsync to match the behavior of git head in regards to
O_DIRECT behavior.
2011-03-10 20:25:41 -05:00
Tom Lane 629b3af27d Convert contrib modules to use the extension facility.
This isn't fully tested as yet, in particular I'm not sure that the
"foo--unpackaged--1.0.sql" scripts are OK.  But it's time to get some
buildfarm cycles on it.

sepgsql is not converted to an extension, mainly because it seems to
require a very nonstandard installation process.

Dimitri Fontaine and Tom Lane
2011-02-13 22:54:49 -05:00
Bruce Momjian 4fec63f94e Per Peter E, use 'kB' for kilobyte, not 'K'. 2011-01-26 16:21:19 -05:00
Bruce Momjian 00869024cb In pg_test_fsync, use K(1024) rather than k(1000) for write size units. 2011-01-26 11:02:52 -05:00
Bruce Momjian 64bc872761 Adjust pg_test_fsync to always do XLOG_BLCKSZ-sized writes, rather than
always 8k writes, per suggestion from Tom.  Also adjust open_sync output
layout.
2011-01-25 22:18:24 -05:00
Bruce Momjian 9dd7933937 Use XLOG_BLCKSZ in pg_test_fsync, rather than our own define, but verify
it is 8k as expected.
2011-01-24 20:07:05 -05:00
Bruce Momjian 9fc0192283 In test_open_sync(), writes kilobytes as intended, not by bytes. 2011-01-24 19:42:32 -05:00
Tom Lane 37eb2cd4ad More pg_test_fsync fixups.
Reduce #includes to minimum actually needed; in particular include
postgres_fe.h not postgres.h, so as to stop build failures on some
platforms.

Use get_progname() instead of hardwired program name; improve error
checking for command line syntax; bring error messages into line with
style guidelines; include strerror result in die() cases.
2011-01-22 15:01:26 -05:00
Tom Lane 3ae28ce8c4 Suppress unused-variables warning when OPEN_SYNC_FLAG isn't defined.
Per buildfarm.
2011-01-22 12:26:44 -05:00
Tom Lane cb38ab6d3b More pg_test_fsync cleanup.
Un-break Windows build (I hope) by making the HAVE_FSYNC_WRITETHROUGH
code match the backend.  Fix incorrect program help message.  static-ize
all functions.
2011-01-21 19:44:53 -05:00
Tom Lane bc616703e8 Clean up pg_test_fsync commit.
Actually rename the program, rather than just claiming we did.  Hook it
into the build system.  Get rid of useless dependency on libpq.  Clean up
#include list and messy whitespace.
2011-01-21 19:27:25 -05:00
Bruce Momjian b35bfcae93 Update C banner on new pg_test_fsync file. 2011-01-21 12:49:04 -05:00
Bruce Momjian 606a3d54fc Move test_fsync to /contrib. 2011-01-21 12:47:54 -05:00