Overview¶
This is the README for ffs.py, a simulator of FFS allocation policies. Use it
to study FFS behavior under different file and directory creation scenarios.
The tool is invoked by specifying a command file with the -f flag, which consists of a series of file create, file delete, and directory create operations.
For example, run:
prompt> ./ffs.py -f in.example1 -c
to see the output from the first example in the chapter on how FFS based allocation works.
The file in.example1 consists of the following commands:
dir /a
dir /b
file /a/c 2
file /a/d 2
file /a/e 2
file /b/f 2
This tells the simulator to create two directories (/a and /b) and four files (/a/c, /a/d, /a/e, and /b/f). The root directory is created by default.
The output of the simulator is the location of the inodes and data blocks of all extant files and directories. For example, from the run above, we would end up seeing (with the -c flag on, to show you the results):
prompt> ./ffs.py -f in.example1 -c
num_groups: 10
inodes_per_group: 10
blocks_per_group: 30
free data blocks: 289 (of 300)
free inodes: 93 (of 100)
spread inodes? False
spread data? False
contig alloc: 1
0000000000 0000000000 1111111111 2222222222
0123456789 0123456789 0123456789 0123456789
group inodes data
0 /--------- /--------- ---------- ----------
1 acde------ accddee--- ---------- ----------
2 bf-------- bff------- ---------- ----------
3 ---------- ---------- ---------- ----------
4 ---------- ---------- ---------- ----------
5 ---------- ---------- ---------- ----------
6 ---------- ---------- ---------- ----------
7 ---------- ---------- ---------- ----------
8 ---------- ---------- ---------- ----------
9 ---------- ---------- ---------- ----------
prompt>
This first part of the output shows us various parameters of the simulation, from the number of FFS cylinder groups that are created, to some policy details. But the main part of the output is the actual allocation map:
0000000000 0000000000 1111111111 2222222222
0123456789 0123456789 0123456789 0123456789
group inodes data
0 /--------- /--------- ---------- ----------
1 acde------ accddee--- ---------- ----------
2 bf-------- bff------- ---------- ----------
3 ---------- ---------- ---------- ----------
4 ---------- ---------- ---------- ----------
5 ---------- ---------- ---------- ----------
6 ---------- ---------- ---------- ----------
7 ---------- ---------- ---------- ----------
8 ---------- ---------- ---------- ----------
9 ---------- ---------- ---------- ----------
For this instantiation, we have created a file system with 10 groups, each with 10 inodes and 30 data blocks. Each group just shows the inodes and data blocks, and how they are allocated. If they are free, a - is shown; otherwise, a different symbol is shown per file.
If you want to see a mapping of the symbols to file names, you should use the -M flag:
prompt> ./ffs.py -f in.example1 -c -M
You’ll then see a table at the bottom of the output shows the meanings of each symbol:
symbol inode# filename filetype
/ 0 / directory
a 10 /a directory
c 11 /a/c regular
d 12 /a/d regular
e 13 /a/e regular
b 20 /b directory
f 21 /b/f regular
Here, you can see the root directory is represented by the symbol /, the file /a by the symbol a, and so forth.
Looking at the output, you can thus see a number of interesting things: - The root inode is in the first slot of the Group 0’s piece of the inode table - The root data block is found in the first allocated data block (Group 0) - Directory /a was placed in Group 1, directory /b in Group 2 - The files (inodes and data) for each regular file are found in the same group as their parent inodes (as per FFS)
The rest of the options let you play around with FFS and some minor variants. They are:
prompt> ./ffs.py -h
Usage: ffs.py [options]
Options:
-h, --help show this help message and exit
-s SEED, --seed=SEED the random seed
-n NUM_GROUPS, --num_groups=NUM_GROUPS
number of block groups
-d BLOCKS_PER_GROUP, --datablocks_per_groups=BLOCKS_PER_GROUP
data blocks per group
-i INODES_PER_GROUP, --inodes_per_group=INODES_PER_GROUP
inodes per group
-L LARGE_FILE_EXCEPTION, --large_file_exception=LARGE_FILE_EXCEPTION
0:off, N>0:blocks in group before spreading file to
next group
-f INPUT_FILE, --input_file=INPUT_FILE
command file
-I, --spread_inodes Instead of putting file inodes in parent dir group,
spread them evenly around all groups
-D, --spread_data Instead of putting data near inode,
spread them evenly around all groups
-A ALLOCATE_FARAWAY, --allocate_faraway=ALLOCATE_FARAWAY
When picking a group, examine this many groups at a
time
-C CONTIG_ALLOCATION_POLICY,
--contig_allocation_policy=CONTIG_ALLOCATION_POLICY
number of contig free blocks needed to alloc
-T, --show_spans show file and directory spans
-M, --show_symbol_map
show symbol map
-B, --show_block_addresses
show block addresses alongside groups
-S, --do_per_file_stats
print out detailed inode stats
-v, --show_file_ops print out detailed per-op success/failure
-c, --compute compute answers for me
We’ll explore more of these options in the homework.