summaryrefslogtreecommitdiffstats
path: root/sbin
diff options
context:
space:
mode:
authormckusick <mckusick@FreeBSD.org>2001-04-10 08:38:59 +0000
committermckusick <mckusick@FreeBSD.org>2001-04-10 08:38:59 +0000
commit3931e94b1f7a6bb0b628197dc6652d90af54bae1 (patch)
treef12ad3eb5e978e73baa322d67393ab068631991c /sbin
parent0e306cda1d7c8450b9fb5b9a84458fb5b5a3110f (diff)
downloadFreeBSD-src-3931e94b1f7a6bb0b628197dc6652d90af54bae1.zip
FreeBSD-src-3931e94b1f7a6bb0b628197dc6652d90af54bae1.tar.gz
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>
Diffstat (limited to 'sbin')
-rw-r--r--sbin/newfs/mkfs.c13
-rw-r--r--sbin/newfs/newfs.c18
-rw-r--r--sbin/tunefs/tunefs.c57
3 files changed, 77 insertions, 11 deletions
diff --git a/sbin/newfs/mkfs.c b/sbin/newfs/mkfs.c
index 8386c1a..9ac5502 100644
--- a/sbin/newfs/mkfs.c
+++ b/sbin/newfs/mkfs.c
@@ -119,6 +119,8 @@ extern int maxbpg; /* maximum blocks per file in a cyl group */
extern int nrpos; /* # of distinguished rotational positions */
extern int bbsize; /* boot block size */
extern int sbsize; /* superblock size */
+extern int avgfilesize; /* expected average file size */
+extern int avgfilesperdir; /* expected number of files per directory */
extern u_long memleft; /* virtual memory available */
extern caddr_t membase; /* start address of memory based filesystem */
extern char * filename;
@@ -274,6 +276,17 @@ mkfs(pp, fsys, fi, fo)
if (sblock.fs_nsect <= 0)
printf("preposterous nsect %d\n", sblock.fs_nsect), exit(15);
/*
+ * collect and verify the filesystem density info
+ */
+ sblock.fs_avgfilesize = avgfilesize;
+ sblock.fs_avgfpdir = avgfilesperdir;
+ if (sblock.fs_avgfilesize <= 0)
+ printf("illegal expected average file size %d\n",
+ sblock.fs_avgfilesize), exit(14);
+ if (sblock.fs_avgfpdir <= 0)
+ printf("illegal expected number of files per directory %d\n",
+ sblock.fs_avgfpdir), exit(15);
+ /*
* collect and verify the block and fragment sizes
*/
sblock.fs_bsize = bsize;
diff --git a/sbin/newfs/newfs.c b/sbin/newfs/newfs.c
index e755993..8eed256 100644
--- a/sbin/newfs/newfs.c
+++ b/sbin/newfs/newfs.c
@@ -163,9 +163,7 @@ void fatal();
* The number of sectors are used to determine the size of a cyl-group.
* Kirk suggested one or two meg per "cylinder" so we say two.
*/
-
#define NTRACKS 1 /* number of heads */
-
#define NSECTORS 4096 /* number of sectors */
int mfs; /* run as the memory based filesystem */
@@ -199,6 +197,8 @@ int maxcontig = 0; /* max contiguous blocks to allocate */
int rotdelay = ROTDELAY; /* rotational delay between blocks */
int maxbpg; /* maximum blocks per file in a cyl group */
int nrpos = NRPOS; /* # of distinguished rotational positions */
+int avgfilesize = AVFILESIZ;/* expected average file size */
+int avgfilesperdir = AFPDIR;/* expected number of files per directory */
int bbsize = BBSIZE; /* boot block size */
int sbsize = SBSIZE; /* superblock size */
int mntflags = MNT_ASYNC; /* flags to be passed to mount */
@@ -256,8 +256,8 @@ main(argc, argv)
}
opstring = mfs ?
- "NF:T:Ua:b:c:d:e:f:i:m:o:s:" :
- "NOS:T:Ua:b:c:d:e:f:i:k:l:m:n:o:p:r:s:t:u:vx:";
+ "NF:T:Ua:b:c:d:e:f:g:h:i:m:o:s:" :
+ "NOS:T:Ua:b:c:d:e:f:g:h:i:k:l:m:n:o:p:r:s:t:u:vx:";
while ((ch = getopt(argc, argv, opstring)) != -1)
switch (ch) {
case 'N':
@@ -308,6 +308,14 @@ main(argc, argv)
if ((fsize = atoi(optarg)) <= 0)
fatal("%s: bad fragment size", optarg);
break;
+ case 'g':
+ if ((avgfilesize = atoi(optarg)) <= 0)
+ fatal("%s: bad average file size", optarg);
+ break;
+ case 'h':
+ if ((avgfilesperdir = atoi(optarg)) <= 0)
+ fatal("%s: bad average files per dir", optarg);
+ break;
case 'i':
if ((density = atoi(optarg)) <= 0)
fatal("%s: bad bytes per inode", optarg);
@@ -768,6 +776,8 @@ usage()
fprintf(stderr, "\t-d rotational delay between contiguous blocks\n");
fprintf(stderr, "\t-e maximum blocks per file in a cylinder group\n");
fprintf(stderr, "\t-f frag size\n");
+ fprintf(stderr, "\t-g average file size\n");
+ fprintf(stderr, "\t-h average files per directory\n");
fprintf(stderr, "\t-i number of bytes per inode\n");
fprintf(stderr, "\t-k sector 0 skew, per track\n");
fprintf(stderr, "\t-l hardware sector interleave\n");
diff --git a/sbin/tunefs/tunefs.c b/sbin/tunefs/tunefs.c
index 734e01d..1fda3f7 100644
--- a/sbin/tunefs/tunefs.c
+++ b/sbin/tunefs/tunefs.c
@@ -91,9 +91,10 @@ main(argc, argv)
char *special, *name;
struct stat st;
int Aflag = 0, active = 0;
- int aflag = 0, dflag = 0, eflag = 0, mflag = 0;
- int nflag = 0, oflag = 0, pflag = 0;
- int avalue = 0, dvalue = 0, evalue = 0, mvalue = 0, ovalue = 0;
+ int aflag = 0, dflag = 0, eflag = 0, fflag = 0, mflag = 0;
+ int nflag = 0, oflag = 0, pflag = 0, sflag = 0;
+ int avalue = 0, dvalue = 0, evalue = 0, fvalue = 0;
+ int mvalue = 0, ovalue = 0, svalue = 0;
char *nvalue = NULL;
struct fstab *fs;
char *chg[2], device[MAXPATHLEN];
@@ -104,7 +105,7 @@ main(argc, argv)
if (argc < 3)
usage();
found_arg = 0; /* at least one arg is required */
- while ((ch = getopt(argc, argv, "Aa:d:e:m:n:o:p")) != -1)
+ while ((ch = getopt(argc, argv, "Aa:d:e:f:m:n:o:ps:")) != -1)
switch (ch) {
case 'A':
found_arg = 1;
@@ -132,6 +133,14 @@ main(argc, argv)
errx(10, "%s must be >= 1 (was %s)", name, optarg);
eflag = 1;
break;
+ case 'f':
+ found_arg = 1;
+ name = "average file size";
+ fvalue = atoi(optarg);
+ if (fvalue < 1)
+ errx(10, "%s must be >= 1 (was %s)", name, optarg);
+ fflag = 1;
+ break;
case 'm':
found_arg = 1;
name = "minimum percentage of free space";
@@ -168,6 +177,14 @@ main(argc, argv)
found_arg = 1;
pflag = 1;
break;
+ case 's':
+ found_arg = 1;
+ name = "expected number of files per directory";
+ svalue = atoi(optarg);
+ if (svalue < 1)
+ errx(10, "%s must be >= 1 (was %s)", name, optarg);
+ sflag = 1;
+ break;
default:
usage();
}
@@ -239,6 +256,17 @@ again:
sblock.fs_maxbpg = evalue;
}
}
+ if (fflag) {
+ name = "average file size";
+ if (sblock.fs_avgfilesize == fvalue) {
+ warnx("%s remains unchanged as %d", name, fvalue);
+ }
+ else {
+ warnx("%s changes from %d to %d",
+ name, sblock.fs_avgfilesize, fvalue);
+ sblock.fs_avgfilesize = fvalue;
+ }
+ }
if (mflag) {
name = "minimum percentage of free space";
if (sblock.fs_minfree == mvalue) {
@@ -291,6 +319,17 @@ again:
warnx(OPTWARN, "space", "<", MINFREE);
}
}
+ if (sflag) {
+ name = "expected number of files per directory";
+ if (sblock.fs_avgfpdir == svalue) {
+ warnx("%s remains unchanged as %d", name, svalue);
+ }
+ else {
+ warnx("%s changes from %d to %d",
+ name, sblock.fs_avgfpdir, svalue);
+ sblock.fs_avgfpdir = svalue;
+ }
+ }
putsb(&sblock, special, Aflag);
if (active) {
@@ -307,9 +346,9 @@ void
usage()
{
fprintf(stderr, "%s\n%s\n%s\n",
-"usage: tunefs [-A] [-a maxcontig] [-d rotdelay] [-e maxbpg] [-m minfree]",
-" [-p] [-n enable | disable] [-o space | time]",
-" special | filesystem");
+"usage: tunefs [-A] [-a maxcontig] [-d rotdelay] [-e maxbpg] [-f avgfilesize]",
+" [-m minfree] [-p] [-n enable | disable] [-o space | time]",
+" [-s filesperdir] special | filesystem");
exit(2);
}
@@ -366,6 +405,10 @@ printfs()
sblock.fs_rotdelay);
warnx("maximum blocks per file in a cylinder group: (-e) %d",
sblock.fs_maxbpg);
+ warnx("average file size: (-f) %d",
+ sblock.fs_avgfilesize);
+ warnx("average number of files in a directory: (-s) %d",
+ sblock.fs_avgfpdir);
warnx("minimum percentage of free space: (-m) %d%%",
sblock.fs_minfree);
warnx("optimization preference: (-o) %s",
OpenPOWER on IntegriCloud