1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-23 11:18:54 +00:00
freebsd/sys
Kirk McKusick a61ab64ac4 Directory layout preference improvements from Grigoriy Orlov <gluk@ptci.ru>.
His description of the problem and solution follow. My own tests show
speedups on typical filesystem intensive workloads of 5% to 12% which
is very impressive considering the small amount of code change involved.

------

  One day I noticed that some file operations run much faster on
small file systems then on big ones. I've looked at the ffs
algorithms, thought about them, and redesigned the dirpref algorithm.

  First I want to describe the results of my tests. These results are old
and I have improved the algorithm after these tests were done. Nevertheless
they show how big the perfomance speedup may be. I have done two file/directory
intensive tests on a two OpenBSD systems with old and new dirpref algorithm.
The first test is "tar -xzf ports.tar.gz", the second is "rm -rf ports".
The ports.tar.gz file is the ports collection from the OpenBSD 2.8 release.
It contains 6596 directories and 13868 files. The test systems are:

1. Celeron-450, 128Mb, two IDE drives, the system at wd0, file system for
   test is at wd1. Size of test file system is 8 Gb, number of cg=991,
   size of cg is 8m, block size = 8k, fragment size = 1k OpenBSD-current
   from Dec 2000 with BUFCACHEPERCENT=35

2. PIII-600, 128Mb, two IBM DTLA-307045 IDE drives at i815e, the system
   at wd0, file system for test is at wd1. Size of test file system is 40 Gb,
   number of cg=5324, size of cg is 8m, block size = 8k, fragment size = 1k
   OpenBSD-current from Dec 2000 with BUFCACHEPERCENT=50

You can get more info about the test systems and methods at:
http://www.ptci.ru/gluk/dirpref/old/dirpref.html

                              Test Results

             tar -xzf ports.tar.gz               rm -rf ports
  mode  old dirpref new dirpref speedup old dirprefnew dirpref speedup
                             First system
 normal     667         472      1.41       477        331       1.44
 async      285         144      1.98       130         14       9.29
 sync       768         616      1.25       477        334       1.43
 softdep    413         252      1.64       241         38       6.34
                             Second system
 normal     329         81       4.06       263.5       93.5     2.81
 async      302         25.7    11.75       112          2.26   49.56
 sync       281         57.0     4.93       263         90.5     2.9
 softdep    341         40.6     8.4        284          4.76   59.66

"old dirpref" and "new dirpref" columns give a test time in seconds.
speedup - speed increasement in times, ie. old dirpref / new dirpref.

------

Algorithm description

The old dirpref algorithm is described in comments:

/*
 * Find a cylinder to place a directory.
 *
 * The policy implemented by this algorithm is to select from
 * among those cylinder groups with above the average number of
 * free inodes, the one with the smallest number of directories.
 */

A new directory is allocated in a different cylinder groups than its
parent directory resulting in a directory tree that is spreaded across
all the cylinder groups. This spreading out results in a non-optimal
access to the directories and files. When we have a small filesystem
it is not a problem but when the filesystem is big then perfomance
degradation becomes very apparent.

What I mean by a big file system ?

  1. A big filesystem is a filesystem which occupy 20-30 or more percent
     of total drive space, i.e. first and last cylinder are physically
     located relatively far from each other.
  2. It has a relatively large number of cylinder groups, for example
     more cylinder groups than 50% of the buffers in the buffer cache.

The first results in long access times, while the second results in
many buffers being used by metadata operations. Such operations use
cylinder group blocks and on-disk inode blocks. The cylinder group
block (fs->fs_cblkno) contains struct cg, inode and block bit maps.
It is 2k in size for the default filesystem parameters. If new and
parent directories are located in different cylinder groups then the
system performs more input/output operations and uses more buffers.
On filesystems with many cylinder groups, lots of cache buffers are
used for metadata operations.

My solution for this problem is very simple. I allocate many directories
in one cylinder group. I also do some things, so that the new allocation
method does not cause excessive fragmentation and all directory inodes
will not be located at a location far from its file's inodes and data.
The algorithm is:
/*
 * Find a cylinder group to place a directory.
 *
 * The policy implemented by this algorithm is to allocate a
 * directory inode in the same cylinder group as its parent
 * directory, but also to reserve space for its files inodes
 * and data. Restrict the number of directories which may be
 * allocated one after another in the same cylinder group
 * without intervening allocation of files.
 *
 * If we allocate a first level directory then force allocation
 * in another cylinder group.
 */

  My early versions of dirpref give me a good results for a wide range of
file operations and different filesystem capacities except one case:
those applications that create their entire directory structure first
and only later fill this structure with files.

  My solution for such and similar cases is to limit a number of
directories which may be created one after another in the same cylinder
group without intervening file creations. For this purpose, I allocate
an array of counters at mount time. This array is linked to the superblock
fs->fs_contigdirs[cg]. Each time a directory is created the counter
increases and each time a file is created the counter decreases. A 60Gb
filesystem with 8mb/cg requires 10kb of memory for the counters array.

  The maxcontigdirs is a maximum number of directories which may be created
without an intervening file creation. I found in my tests that the best
performance occurs when I restrict the number of directories in one cylinder
group such that all its files may be located in the same cylinder group.
There may be some deterioration in performance if all the file inodes
are in the same cylinder group as its containing directory, but their
data partially resides in a different cylinder group. The maxcontigdirs
value is calculated to try to prevent this condition. Since there is
no way to know how many files and directories will be allocated later
I added two optimization parameters in superblock/tunefs. They are:

        int32_t  fs_avgfilesize;   /* expected average file size */
        int32_t  fs_avgfpdir;      /* expected # of files per directory */

These parameters have reasonable defaults but may be tweeked for special
uses of a filesystem. They are only necessary in rare cases like better
tuning a filesystem being used to store a squid cache.

I have been using this algorithm for about 3 months. I have done
a lot of testing on filesystems with different capacities, average
filesize, average number of files per directory, and so on. I think
this algorithm has no negative impact on filesystem perfomance. It
works better than the default one in all cases. The new dirpref
will greatly improve untarring/removing/coping of big directories,
decrease load on cvs servers and much more. The new dirpref doesn't
speedup a compilation process, but also doesn't slow it down.

Obtained from:	Grigoriy Orlov <gluk@ptci.ru>
2001-04-10 08:38:59 +00:00
..
alpha insert a magical second memory barrier prior to calling draina() in 2001-04-08 16:43:59 +00:00
amd64 - One can now specify the decimal pid of a process to trace as a parameter. 2001-04-09 21:43:45 +00:00
arm/include Correct disordering which is corresponding to bde's fix to 2001-02-17 14:51:11 +00:00
boot no longer needed now that we are able to build cdboot from sources again 2001-04-08 00:01:54 +00:00
cam If we have and error and are booting verbosely, don't be complaining 2001-04-04 18:24:35 +00:00
coda Send the remains (such as I have located) of "block major numbers" to 2001-03-26 12:41:29 +00:00
compat Add linux_sched_get_priority_max() and linux_sched_get_priority_min(): The 2001-04-01 06:37:40 +00:00
compile
conf Import kernel part of SMB/CIFS requester. 2001-04-10 07:59:06 +00:00
contrib/dev Send the remains (such as I have located) of "block major numbers" to 2001-03-26 12:41:29 +00:00
crypto Kernel crypto need binary key material, not symbolic ascii. 2001-03-10 13:02:58 +00:00
ddb Catch up to header include changes: 2001-03-28 09:17:56 +00:00
dev Add more diagnostic output for failure. 2001-04-10 05:29:26 +00:00
fs Import kernel part of SMB/CIFS requester. 2001-04-10 07:59:06 +00:00
geom fix a number of printf format string warnings inside DEBUG ifdefs 2001-03-29 15:05:08 +00:00
gnu Fixes to track snapshot copy-on-write checking in the specinfo 2001-03-07 07:09:55 +00:00
i4b Send the remains (such as I have located) of "block major numbers" to 2001-03-26 12:41:29 +00:00
i386 - One can now specify the decimal pid of a process to trace as a parameter. 2001-04-09 21:43:45 +00:00
ia64 Reduce the emasculation of bounds_check_with_label() by one line, so we 2001-03-29 20:26:12 +00:00
isa Catch up to header include changes: 2001-03-28 09:17:56 +00:00
isofs/cd9660 Add missed MODULE_VERSION() call, so loading of unicode conversion routine 2001-03-11 15:28:42 +00:00
kern Import kernel part of SMB/CIFS requester. 2001-04-10 07:59:06 +00:00
libkern Add function prototypes and base module for kernel side iconv library. 2001-04-09 09:39:29 +00:00
miscfs - Various style fixes. 2001-03-29 18:10:46 +00:00
modules We now depend on miibus_if.h. 2001-04-09 21:34:52 +00:00
msdosfs Grab the process lock while calling psignal and before calling psignal. 2001-03-07 03:37:06 +00:00
net Move the decision whether we want to request authentication from our 2001-04-08 20:29:09 +00:00
netatalk Mechanical change to use <sys/queue.h> macro API instead of 2001-02-04 13:13:25 +00:00
netatm Silence some warnings 2001-03-20 10:42:49 +00:00
netgraph Catch up to header include changes: 2001-03-28 09:17:56 +00:00
netinet fix security hole created by fragment cache 2001-04-06 15:52:28 +00:00
netinet6 - correct logic of per-address input packet counts for lo0 2001-04-05 19:45:02 +00:00
netipx Another round of the <sys/queue.h> FOREACH transmogriffer. 2001-02-04 16:08:18 +00:00
netkey
netnatm Change a couple of M_WAITOKs used in M_PREPEND() to M_TRYWAITs, which 2001-04-05 04:20:48 +00:00
netncp Move bindery authentication ncps to ncp_ncp.c file. ncp_login.c will stay 2001-03-22 10:38:16 +00:00
netns
netsmb Import kernel part of SMB/CIFS requester. 2001-04-10 07:59:06 +00:00
nfs o Rather than arbitrarily construct a credential in the nfs_statfs() 2001-04-05 06:12:38 +00:00
nfsclient o Rather than arbitrarily construct a credential in the nfs_statfs() 2001-04-05 06:12:38 +00:00
nfsserver Use a generic implementation of the Fowler/Noll/Vo hash (FNV hash). 2001-03-17 09:31:06 +00:00
ntfs Reviewed by: jlemon 2001-03-01 21:00:17 +00:00
nwfs Add dependancy on libmchain module. 2001-03-22 06:51:53 +00:00
pc98 Correct typo. 2001-04-01 07:15:16 +00:00
pccard Send the remains (such as I have located) of "block major numbers" to 2001-03-26 12:41:29 +00:00
pci Several things: 2001-04-09 21:54:15 +00:00
posix4 Lock need_resched with sched_lock. 2001-02-22 13:46:09 +00:00
powerpc Rework the witness code to work with sx locks as well as mutexes. 2001-03-28 09:03:24 +00:00
rpc Bring in a hybrid of SunSoft's transport-independent RPC (TI-RPC) and 2001-03-19 12:50:13 +00:00
svr4
sys Import kernel part of SMB/CIFS requester. 2001-04-10 07:59:06 +00:00
tools replace calls to non-existant bail() subroutine with calls to 2001-03-23 11:48:50 +00:00
ufs Directory layout preference improvements from Grigoriy Orlov <gluk@ptci.ru>. 2001-04-10 08:38:59 +00:00
vm Convert the allproc and proctree locks from lockmgr locks to sx locks. 2001-03-28 11:52:56 +00:00
Makefile