►
From YouTube: CephFS Code Walkthrough: rstats
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
A
A
A
A
Entries
in
the
directory
now
subsequent
entries
statistics
for
entries
deep
more
than
one
level
deep,
are
accumulated
in
the
structure
called
nest
info
now.
Both
of
these
structs
are
embedded
embedded
in
a
container
struct
called
f
node,
underscore
t,
which
is
maintained
per
del
frag
or
per
leader
object.
A
The
m
time
or
c
time
of
any
of
the
files
or
directories
are
collected
together,
whichever
is
higher
or
greater,
is
maintained
as
the
rc
time
and
propagated
to
the
root
rc
r
starts
propagation
is
to
the
root,
is
lazy
and
is
not
instantaneous.
It
doesn't
happen
in
a
blink
of
eye,
since
all
the
cedar
objects
in
the
path
up
to
the
root
have
to
be
updated
or
the
nest
info
and
frag
info
structures
have
to
be
updated
along
the
path.
A
This
involves
a
lot
of
locking-
and
this
is
a
hard
problem
to
solve,
with
multiple
active
mdss
pin,
to
pin
directly
spin
to
different
mdss
with
mds
failures
and
restarts
and
updates
to
snapshot
directories.
This
is
a
really
hard
problem.
Map
rc
time
updates
to
snapshot.
Directories
is
currently
buggy
and
is
being
looked
into.