bonnie++ (1.90d) unstable; urgency=low * Updated standards version to 3.2.1. * Done some serious work on porting to OS/2 and NT. This may break compile on some versions of UNIX. If so I'll fix it as soon as it gets reported. -- Russell Coker Mon, 11 Dec 2000 22:31:20 +0100 bonnie++ (1.90c) unstable; urgency=low * Rewrote bon_csv2html in C++ and made it assign colors to the fields to show how fast or slow the values are (red for slow, green for fast). -- Russell Coker Mon, 11 Dec 2000 22:31:20 +0100 bonnie++ (1.90b) unstable; urgency=low * Added first stage of support for synchronising bonnie++ instances over a network. This required changing the way the -y option operates. Also created a new Sync class to do this generically. * Added code to check the latency of operations in micro-seconds (us). -- Russell Coker Tue, 5 Dec 2000 17:04:27 +0100 bonnie++ (1.90a) unstable; urgency=low * This is the start of the new 2.00 series of Bonnie++. The aim is to do serious performance testing of RAID arrays. All programs will be multi- threaded. * Added the multi-threaded zcav that I had removed from the 1.00 series and make Bonnie++ use threads instead of fork(). Next step is to add extra concurency through threads. * Changed the format of the csv files. Added a format version number and a place for the version of Bonnie++ that created the data. Made the bon_csv2txt and bon_csv2html programs understand the new format. -- Russell Coker Wed, 29 Nov 2000 23:53:13 +0100 bonnie++ (1.00f) unstable; urgency=low * Fixed the bugs in timing of seeks. * Changed the number of seeks from 8000 to 8192. * Now the minimum time for a test that will be considered valid is 500ms (for bonnie++ and zcav), it was previously 1000ms but I have been convinced that system clocks are accurate enough for this. * Changed the default number of files created for file creation tests from 30K to 16K, this change makes the test time bearable on obsolete file systems like UFS but relies on the above change to work on ReiserFS on Pentium4 class CPUs. * Changed the default file size to 300M for IO tests, this reflects the increase in RAM sizes over the last year. * Added some more compile warnings and fixed some more trivial bugs. * Made the loops: line in zcav output always be prefixed by a '#' for a gnuplot comment. * Made zcav correctly display the position in megabytes instead of block-size units. -- Russell Coker Mon, 27 Nov 2000 09:45:30 +0100 bonnie++ (1.00e) unstable; urgency=low * Now exit on sync errors. * When directory can't be synced it will display one warning and not try to sync again. * Stopped it crashing when there is only one directory to be synced. * Made the version number managed by autoconf so I don't release it with inconsistant version numbers again! -- Russell Coker Thu, 9 Nov 2000 03:26:15 +0100 bonnie++ (1.00d) unstable; urgency=low * Use SA_RESETHAND instead of SA_ONESHOT for the benefit of Solaris. * Added a specific rule for bon_suid.cpp as it doesn't have a header file. * Added --prefix support to the ./configure , also made the default prefix be /usr/local instead of /usr . * Changed the autoconf support for checking for C++ slightly. It should now work better for some strange setups and work the same for everyone else (I hope). * Made the autoconf tests for semaphore headers work correctly, now it should compile on *BSD. * Added --disable-stripping option for ./configure if you don't want binaries stripped. * Added autoconf checking for bool, now it should compile on AIX using ICC without any special options! * Reverted zcav to the 1.00a version and then added the code for -u, -g, and the fix for large numbers of data points. The multi-threaded zcav code will go into 1.90 (the pre-2.00 tree). Bonnie++ versions < 1.90 will never again have threading code. * Made bon_csv2txt use every available character for the description. * Made it install man pages by default. * Added sun package support - "make -C sun"! -- Russell Coker Thu, 28 Sep 2000 16:22:15 +0100 bonnie++ (1.00c) unstable; urgency=low * Closes:#53545 Fixed in 0.99e, should have been closed in Debian BTS. * Closes:#53546 Fixed in 0.99e, should have been closed in Debian BTS. * Closes:#61925 Fixed in 1.00a. * Closes:#64995 It is in /usr/sbin because it can only sensibly be run by the administrator of the machine, otherwise it probably will give bad results and may impact the performance of the machine in question. Also it can now change UID/GID. The new way of dealing with the running as root issue is in the next change. * Made zcav and bonnie++ take -u and -g options to set the UID and GID to run as. For bonnie++ it is now mandatory to use the -u option when running as root. * Made bonnie++ not change it's command-line. * Documented the K and G suffixes for sizes on the command-line. * Now the CPU time field also displays as "+++" if the elapsed time is < 1s. * Fixed the machine-name broken-ness from 1.00b, and made the machine name default to the nodename as reported by utsname() (also the output of `hostname` or `uname -n`). * Now uses sysconf() to check the ram size, you can use -r to over-ride it. Also the default file size for IO tests will be twice the reported RAM size or 200M (whichever is larger). * Now Bonnie++ handles ^C from the keyboard, the XCPU and XFSZ (excessive CPU and excessive file size) flags and aborts the program cleanly removing all files when it receives them. Also ignores SIGHUP. * Added AC_PROG_CC to configure.in so that it can compile on systems with strange C compiler setups. -- Russell Coker Fri, 08 Sep 2000 08:22:47 +0100 bonnie++ (1.00b) unstable; urgency=low * Added more warnings to the compile and removed assertions. Made some trivial changes to the code (like changing variable names) to stop the warnings. * Fixed the memory management problem on bonnie++, these made it not work on IA64 (and stopped it working correctly on most platforms). Thanks to Electric Fence by Bruce Perens for the discovery of this. The worst part of it was introduced in testing this version, so it only hit me and my alpha-testers. * Fixed zcav for large numbers of data points. * Made zcav multi-threaded to test multiple hard drives at once. Changed the way it works totally. * Removed some dependencies on extensions to the C++ standard which are not supported in all compilers, also removed some trivial header file issues. These were found in testing on Tru64Unix. * Fixed a bug in bonnie++, it would go into an infinite loop when the file creation tests had a non-zero size. * Made bonnie++ work for block-reads that return partial blocks, now it will print an error and do another read for the rest. * Made Bonnie++ accept machine names up to 4095 bytes and not crash if the name is longer. Previously the limit was 20 bytes and it crashed when you exceeded it. * This version is fairly experimental but I'm releasing it now because I need wider testing of the new features. -- Russell Coker Fri, 25 Aug 2000 12:15:06 +0200 bonnie++ (1.00a) unstable; urgency=low * Added a 30 second startup delay when run as root. A user lost some data because of running it as root, if they had run it as a regular account they would be OK. I don't want this to happen again. * Zcav now displays an error if it can't read a single block. * Added some basic autoconf support which I will use to increase portability in future versions. * Now provides zcav. * Should compile with all old versions of gcc. * Fixed a warning on Alpha. -- Russell Coker Mon, 24 Apr 2000 23:34:02 +0100 bonnie++ (1.00) unstable; urgency=low * Now include ZCAV in the same package. ZCAV package should disappear. * License is now GPL. Tim Bray agrees to the GPL for his parts, the license conflict was stopping me from putting ZCAV into the archive. * ZCAV reads through a hard drive sequentially and reports the IO speeds for different zones of the drive. * Fixed a few minor issues with the documentation, and put the test programs in /usr/sbin as they aren't generally run by regular users. Also use man section 8. -- Russell Coker Sat, 01 Mar 2000 12:01:00 +0100 bonnie++ (0.99j) unstable; urgency=low * 0.99h core dumped when you didn't specify "-b" for file creation tests, fixed. -- Russell Coker Sun, 05 Mar 2000 11:16:42 +0100 bonnie++ (0.99h) unstable; urgency=low * Fixed a variety of bugs in the semaphore code which were introduced in 0.99g. * Fixed formatting of output. * Added "-b" option to sync all writes. * Changed the semaphore code to make it more easily hackable for BSD users, it won't compile as-is on BSD at the moment... -- Russell Coker Sun, 05 Mar 2000 11:16:42 +0100 bonnie++ (0.99g) unstable; urgency=low * Now use getopt() for checking command-line options. * Added new versions of fork and semaphore code, initially developed for postal. * Fixed the message that's displayed when bad command-line parameters are entered. * Version 1.[0-8]0 will use fork(). Version 1.90 and above will use POSIX threads and include the concurrant bonnie++ functionality I've been promising for so long. -- Russell Coker Wed, 23 Feb 2000 22:16:23 +0100 bonnie++ (0.99f) unstable; urgency=low * Added "-f" parameter to skip per-char tests and semaphore code to synchronise multiple instances of Bonnie++. Thanks to Christian Kagerhuber for the patch! * Added srand() after the fork so each child gets different random numbers. -- Russell Coker Wed, 12 Jan 2000 16:45:28 +1100 bonnie++ (0.99e) unstable; urgency=low * Fixed the operation of "-x" parameter (used to just cause crashes). * Made it cleanly exit under some error conditions where it used to crash. * Improved the bonnie++ man page. * Fixed some checking of command-line parameters. * Merged code from the OS/2 port, needs lots of testing... -- Russell Coker Wed, 12 Jan 2000 16:45:28 +1100 bonnie++ (0.99d) unstable; urgency=low * Added some more functionality. Tests hard and soft link creation. * Fixed CSV output of <100 seeks per second. -- Russell Coker Sun, 21 Nov 1999 22:37:42 +0200 bonnie++ (0.99c) unstable; urgency=low * Fix some bugs with big IO (fseek related) and include man pages. * Made it always print the CSV data. -- Russell Coker Sun, 21 Nov 1999 22:37:42 +0200 bonnie++ (0.99b) unstable; urgency=low * Initial Release as a Debian package. 0.99 Files are created mode 0600 not 0777. Fixed some bugs in 0.98 where the results from several tests were totally wrong. Now the random file code will take less CPU time when there are extremely large numbers of files. Changed the format of all the output files slightly. Notable change is that the percentages of CPU time are now rounded off to the nearest percent. This is because it's not that accurate anyway (results that are provably more than 1% wrong are not uncommon), and because I needed the extra 1 character per field. Also now it handles CPU time >100% properly. This is for SMP systems where more than 1 CPU is being used. Concurrant Bonnie++ will return many results significantly greater than 100% on OSs that work well with SMP. Added a csv2txt.pl program. The main aim of this is to display data well for 80 column braille displays for the blind. Added "-q" option for quiet mode (less output). Now the "-n" option works on a multiple of 1024. So "-n 10" means create 10240 files. This change is to allow the output to display in the same format and save space in display (who would want to test as a lower resolution than per 1024 files anyway). The -n option is now of the form "num[:max[:min]]" where max is the maximum size (default 0) and min is the minimum size (default 0). To simulate Squid use a max of 15000 and a min of 300. To simulate INN use a maximum of 4096 and a minimum of 512. 1.0 will be out soon! 0.98 Fixed a bug where the data size couldn't be an exact multiple of the size of each file (1 gig). Fixed a number of other minor bugs related to that and added more error checking as well. Changed the code to support up to 1000 files for the IO test, if each is a gig then you can test a tera-byte of data. Changing the code to have more than 1000 files wouldn't be that difficult to do. Use the new C++ type conversions. 0.97 I have stopped using cout/cerr and never plan to use them again. They caused me significant pain when trying to get it going on an ancient SGI system. Also changed the code structure a bit to make it cleaner. One advantage of this is that there is now a "-x" option to tell bonnie++ to run the same test a number of times (it's interesting to see the variance in the results). Now use fflush() after writing each set of results. This means that killing the program unexpectedly won't result in results being lost. Also fixes a strange bug related to printf() on Linux which I am still looking into. -- Russell Coker Wed, 13 Oct 1999 22:15:53 +0200 Local variables: mode: debian-changelog End: