c3b0aa04e2
Addendum to the fix for #10 : also ignore duplicate key errors, as they
...
are legal for PostgreSQL in the OLTP workload due to its transaction
isolation implementation specifics.
2015-10-02 10:29:37 +03:00
ccc252f43c
Fixed #3 in the 0.4 branch.
2015-10-02 10:26:45 +03:00
3d63e69a69
More descriptive configure message when MySQL client library is not found.
2015-09-30 19:22:05 +03:00
76fc3ef79e
Fixes #10 : oltp: pgsql query execution failed
...
Handle deadlocks correctly in the PostgreSQL driver, i.e. return
SB_DB_ERROR_DEADLOCK to the upper layer to restart the transaction and
increment the deadlocks counter.
2015-09-23 16:59:04 +03:00
26abf0c893
Fix wrong format string in error reporting.
2015-09-23 15:51:48 +03:00
b0ce007822
Added the --mysql-dry-run option.
2015-04-17 17:00:59 -07:00
07ba467bfc
Created .gitignore.
2015-03-09 21:56:59 +03:00
ad9c40f592
Fixed LP #1417545 : Minor fixes for compilation informationals/complaints on VMS
...
Fixed unsigned int comparisons with 0 in a couple of places.
2015-03-09 21:53:40 +03:00
a24fa26f0f
Bug #1415750 : --with-pgsql-libs is ignored
...
Merged a contributed patch.
2015-03-01 21:18:51 +03:00
7de7f37551
Bug #1415746 : mysql include path was set in PGSQL_CFLAGS
...
Merged a contributed patch.
2015-03-01 21:14:41 +03:00
80bc21830a
Bug #1197127 : fastmmap and slowmmap arg should be mmap
...
Replaced references to "fastmmap" and "slowwmap" with "mmap" in both
documentation and the built-in help.
2013-07-04 10:27:59 +04:00
9f15f73044
Linked the Serbo-Croatian translation of the manual contributed by Vera Djuraskovic <veradjuraskovic@webhostinggeeks.com>.
...
Also fixed bug #1071737 "Docs say --oltp-read-only should be use,
though only --oltp-read-only=on works".
2013-03-10 16:34:17 +04:00
0e7116ef5a
Bug #1073672 : --forced-shutdown does not reflect stuck transactions in
...
stats
When --forced-shutdown is used, it prints the currently collected
statistics and terminates without waiting for worker threads to
complete. This was implemented to be able to complete benchmark even
if we have some long running or stuck transactions.
The problem is that unfinished transactions are not reflected in
statistics. So, for example, if we do have stuck transactions that
have been stuck for the entire test duration, their timers will still
be running on forced shutdown and thus, statistics will only show
transactions performed by other threads, if any.
The fix is stop all timers on forced shutdown, so even stuck
transactions are accounted with (forced_shutdown_time -
trx_start_time) duration. Also printing the number of currently
running transactions at the time of forced shutdown would also make
sense.
2012-10-31 22:30:24 +04:00
cbed8d8942
Bug #1042442 : Automake fails
...
Merged configure fix for newer Automake versions contributed by
Raghavendra D Prabhu.
2012-08-30 12:46:49 +04:00
516ae92c46
Update fileio stats in async mode when completing requests rather than
...
submitting them.
2012-03-10 23:47:56 -08:00
82c96ff62d
Bug #863564 : Make fileio prepare reuse existing files
...
Changed the fileio prepare command so that any existing files left from
the previous runs are now reused rather than recreated from
scratch. This is is mostly useful for SSD drives, but may speed up the
prepare stage for regular drives too.
2011-09-30 22:40:56 +04:00
93c2cd6113
Bug #863159 : Provide percentile response time stats for intermediate
...
report in fileio test
Added response time percentile to intermediate reports in fileio.
2011-09-30 16:17:42 +04:00
6f3592ec28
Fixed compilation on Windows (patch contributed by Vladislav Vaintroub).
2011-09-26 17:48:33 +04:00
5020eea644
Make --report-interval stats use the --percentile value rather than
...
hardcoded ones.
2011-07-21 20:59:45 +04:00
80df3b3500
Bug #811105 : Add an option to dump full stats after specified time
...
intervals.
Added a new --report-checkpoints option that allows to dump full
statistics and reset all counters at specified points in time. The
argument is a list of comma-separated values representing the amount of
time in seconds elapsed from start of test when report checkpoint(s)
must be performed. Report checkpoints are off by Report checkpoints
are off by default.
2011-07-21 17:29:08 +04:00
3d52f3e5b5
Bug #797269 : Print percentile statistics in --report-interval stats
...
--report-interval now also prints 95th and 99th percentile of
transactions response times.
2011-07-19 17:53:08 +04:00
3943f66125
Make the prepare stage in fileio honour --file-extra-flags.
2011-05-18 14:24:16 +04:00
ae40add564
Delay killing the --report-interval thread until the very end of
...
run_test() to avoid mutex lock leaks.
2011-04-19 13:29:19 +04:00
c4bb8a246c
In --report-interval mode, reads/second and writes/seconds are now displayed
...
in addition to tps.
2011-01-12 13:39:47 +03:00
2052190874
Automerge.
2010-11-09 20:00:02 +03:00
bf3e2c8c0b
Fixed an example using an obsolete option in the manual
2010-11-09 18:02:10 +03:00
63f538eb35
Fix race condition. Global variables were begin accessed without a mutex locked. This caused
...
some fsync calls to use a bogus file descriptor, FATAL errors and early test failure.
2010-11-07 11:30:49 -08:00
55389ac78f
Fixed compiler warnings.
2010-11-06 17:10:38 +03:00
ff225a7490
Fixed LP bug #671684 : build broken in lp:sysbench/0.4
...
Backported autotools-related fixes from trunk.
2010-11-06 15:37:17 +03:00
de21c2a060
Fixed docs for --report-interval to mention that statistics is
...
per-interval rather than cumulative.
2010-11-06 11:03:26 +03:00
c055685f24
Fix race condition. Global variables were begin accessed without a mutex locked. This caused
...
some fsync calls to use a bogus file descriptor, FATAL errors and early test failure.
2010-11-05 18:03:23 -07:00
c33ec00456
Rewrote intermediate statistics feature contributed by Mikael Ronstrom.
...
Removed "batch mode", --report-interval should now be used instead.
Fixed compile warnings produced by gcc with -Wextra.
2010-11-01 11:35:41 +03:00
d34258cf2d
Fixed a typo in the manual.
2010-10-14 12:25:00 +04:00
51f5dd4f5f
Fixed Launchpad bug #654832 .
...
Fixed a couple of build warnings in drv_pgsql.c.
2010-10-07 01:02:41 +04:00
1425aa3c95
Merged Tim Cook's patch adding transaction rate control.
2010-09-24 02:49:21 +04:00
f81f75d427
Removed auto-generated files from version control.
...
Created .bzrignore.
2010-09-23 16:37:15 +04:00
46a950f9f9
Added intermediate-result-timer, also fixed more proper timing to avoid thread variations to impact result too much
2010-09-22 11:14:13 +02:00
ce91c1adac
Added transaction rate control option oltp test
2010-09-22 10:38:55 +02:00
f8280bd124
Increased minimum thread stack size on Windows to 128K to avoid "Can't create TCP/IP socket (10106)" error on Windows 7.
2010-04-26 11:25:11 +04:00
46b5eceb4d
Merge r67 from lp:~mdcallag/sysbench/0.4 to get a bug fix
...
Use sb_get_value_int instead of sb_get_value_flag for --seed-rng
2009-10-15 13:00:31 -07:00
3a24ba8042
Use sb_get_value_int instead of sb_get_value_flag for --send-rng
2009-10-02 16:55:04 -07:00
3c714e34e0
merge lp:~mdcallag/sysbench/0.4 to get revisions:
...
66 : add --seed-rng
67 : add --oltp-point-select-mysql-handler
68 : add --oltp-secondary and --oltp-point-select-all-cols
2009-10-01 16:12:07 -07:00
d9b13b92b2
Add --oltp-secondary option to use a secondary index on _id_ rather than a primary key.
...
Add --oltp-point-select-all-cols to make the point-select query (fetch 1 row by primary key)
fetch all columns from the row rather than only _c_. This makes it comparable to the work
done when --oltp-point-select-mysql-handler is used because HANDLER statements fetch all
columns.
2009-09-30 15:44:21 -07:00
9794383fca
Add option --oltp-point-select-mysql-handler to use HANDLER for primary key lookups
...
instead of a SELECT statement.
2009-09-30 12:37:03 -07:00
7dce4a7e54
add --send-rng option
2009-09-30 11:10:46 -07:00
df31f6767f
* Removed autogenerated ltmain.sh from the repository.
...
* Upgraded autogen.sh to invoke libtoolize to simplify
building from the Bazaar repository.
* Updated docs.
2009-07-28 15:41:45 +04:00
db2d315f74
Added the old OLTP test back
2009-05-28 20:12:23 +04:00
b773e3146b
Since Bazaar does not allow merging of two unrelated branches,
...
created the branch for the old 0.4 code base from trunk, then
reconciled with 0.4 manually (the patchwill be null-merged into
trunk).
2009-05-28 19:25:12 +04:00
c123d1deff
test commit
2009-05-21 05:52:09 -07:00
2fcffcf014
Look for DocBook XML catalog in multiple locations instead of using a hardcoded path
2009-05-19 18:24:39 +00:00
