set project quota
Enabling project quota on an XFS filesystem (restrict files in vm1 directories to only using 1 gigabyte of space).
# rm -f /etc/projects /etc/projid
# mount -o prjquota /dev/sda4 /data
# xfs_quota -x -c 'project -s -p /data/vm1 1000' /data
# xfs_quota -x -c 'limit -p bhard=1g 1000' /data
project command
project [ -cCs [ -d depth ] [ -p path ] id | name ]
The -c, -C, and -s options allow the directory tree quota
mechanism to be maintained. -d allows to limit recursion
level when processing project directories and -p allows to
specify project paths at command line ( instead of
/etc/projects ). All options are discussed in detail below.
setup project
- Q1: implementation
当我们对某个目录调用xfs_quota设置project quota时，会遍历该目录下的所有文件（目录），然后对每个文件设置project id：
///xfsprogs
static int
setup_project(
const char *path,
const struct stat *stat,
int flag,
struct FTW *data)
{
...
if ((fd = open(path, O_RDONLY|O_NOCTTY)) == -1) {
exitcode = 1;
fprintf(stderr, _("%s: cannot open %s: %s\n"),
progname, path, strerror(errno));
return 0;
} else if (xfsctl(path, fd, XFS_IOC_FSGETXATTR, &fsx) < 0) {
exitcode = 1;
fprintf(stderr, _("%s: cannot get flags on %s: %s\n"),
progname, path, strerror(errno));
close(fd);
return 0;
}
fsx.fsx_projid = prid;
fsx.fsx_xflags |= XFS_XFLAG_PROJINHERIT;
if (xfsctl(path, fd, XFS_IOC_FSSETXATTR, &fsx) < 0) { ///see xfs_file_ioctl
exitcode = 1;
fprintf(stderr, _("%s: cannot set project on %s: %s\n"),
progname, path, strerror(errno));
}
close(fd);
In kernel:
xfs_file_ioctl -> xfs_ioc_fssetxattr -> xfs_ioctl_setattr
- Q2: skipping special file
# xfs_quota -x -c 'project -s vm44' /data
Setting up project vm44 (path /data/docker/overlay/a698018e8a59c5c440e72a11958eb6aebdbd1f1872e657b8da91a799e8f2ea5a/)...
xfs_quota: skipping special file /data/docker/overlay/a698018e8a59c5c440e72a11958eb6aebdbd1f1872e657b8da91a799e8f2ea5a/upper/etc/mtab
XFS quota在setup时，会略过一些特殊文件，比如符号链接、块设备、字符设备、FIFO、socket：
//xfsprogs
static int
setup_project(
const char *path,
const struct stat *stat,
int flag,
struct FTW *data)
{
...
if (EXCLUDED_FILE_TYPES(stat->st_mode)) {
fprintf(stderr, _("%s: skipping special file %s\n"), progname, path);
return 0;
}
- Q3: Don’t cross mount points
xfs_quota不会跨文件系统设置quota:
///xfsprogs
static void
project_operations(
char *project,
char *dir,
int type)
{
switch (type) {
case CHECK_PROJECT:
printf(_("Checking project %s (path %s)...\n"), project, dir);
nftw(dir, check_project, 100, FTW_PHYS|FTW_MOUNT);
break;
case SETUP_PROJECT:
printf(_("Setting up project %s (path %s)...\n"), project, dir);
nftw(dir, setup_project, 100, FTW_PHYS|FTW_MOUNT);
break;
case CLEAR_PROJECT:
printf(_("Clearing project %s (path %s)...\n"), project, dir);
nftw(dir, clear_project, 100, FTW_PHYS|FTW_MOUNT);
break;
}
}
FTW_MOUNT and FTW_PHYS:
FTW_MOUNT
If set, stay within the same filesystem (i.e., do not cross
mount points).
FTW_PHYS
If set, do not follow symbolic links. (This is what you
want.) If not set, symbolic links are followed, but no file
is reported twice.
limit command
limit [ -g | -p | -u ] bsoft=N | bhard=N | isoft=N | ihard=N |
rtbsoft=N | rtbhard=N -d | id | name
Set quota block limits (bhard/bsoft), inode count limits
(ihard/isoft) and/or realtime block limits (rtbhard/rtbsoft).
The -d option (defaults) can be used to set the default value
that will be used, otherwise a specific user/group/project
name or numeric identifier must be specified.
- implementation
//xfsprogs
static void
set_limits(
__uint32_t id,
uint type,
uint mask,
char *dev,
__uint64_t *bsoft,
__uint64_t *bhard,
__uint64_t *isoft,
__uint64_t *ihard,
__uint64_t *rtbsoft,
__uint64_t *rtbhard)
{
fs_disk_quota_t d;
memset(&d, 0, sizeof(d));
d.d_version = FS_DQUOT_VERSION;
d.d_id = id;
d.d_flags = type;
d.d_fieldmask = mask;
d.d_blk_hardlimit = *bhard;
d.d_blk_softlimit = *bsoft;
d.d_ino_hardlimit = *ihard;
d.d_ino_softlimit = *isoft;
d.d_rtb_hardlimit = *rtbhard;
d.d_rtb_softlimit = *rtbsoft;
if (xfsquotactl(XFS_SETQLIM, dev, type, id, (void *)&d) < 0) {
exitcode = 1;
fprintf(stderr, _("%s: cannot set limits: %s\n"),
progname, strerror(errno));
}
}
In kernel:
quotactl -> do_quotactl -> xfs_fs_set_xquota
/* Copy parameters and call proper function */
static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id,
void __user *addr)
{
...
case Q_XSETQLIM: {
struct fs_disk_quota fdq;
if (copy_from_user(&fdq, addr, sizeof(fdq)))
return -EFAULT;
return sb->s_qcop->set_xquota(sb, type, id, &fdq);///xfs_fs_set_xquota
}