►
From YouTube: Ceph Crimson/SeaStore Meeting 2023-01-11
Description
Join us weekly for the Ceph Crimson/Seastore meeting: https://ceph.io/en/community/meetups
Ceph website: https://ceph.io
Ceph blog: https://ceph.io/en/news/blog/
Contribute to Ceph: https://ceph.io/en/developers/contrib...
What is Ceph: https://ceph.io/en/discover/
A
All
right,
let's
get
started
so
the
first
item
for
me
is
last
week:
I
sent
OS
8
builds
we're
not
working
and
therefore
neither
were
the
Crimson
builds.
We
were
making
some
progress
on
that
and
then
the
CPU
VPN
and
a
bunch
of
other
lab
stuff
went
down
as
well.
It's
the
same
root
cause
as
the
issues
that
came
up
last
October,
the
gluster
cluster,
underlying
the
lab
infrastructure,
corrupted
some
of
the
VM
disks
to
be
clear.
This
is
a
gluster,
not
stuff.
A
There
is
ongoing
work
to
this,
this
infrastructure
off
of
off
of
that.
But
it's
not
done
yet
so
that's
backup,
but
we
don't
have
to
build
fix
just
yet.
There
are
two
options
here:
the
only
user
for
Java
which
is
what's
breaking
the
build,
is
the
CFS
jni
bindings,
but
they
are
not
actually
that
important.
So
if
we
don't
get
them
fixed
in
the
next
24
hours,
we're
just
going
to
disable
the
cephs
Java
stuff
completely.
That
will
fix
it,
but
Dan
thinks
he
has
a
fix.
A
So
it
may
already
be
working
we'll
we'll
see
anyway,.
B
A
Really
sorry
that
it's
taken
this
long,
hopefully
it'll
be
fixed
by
tomorrow,
other
than
that
I
spent
some
time
getting
caught
up
in
reviews
how's
it
going.
C
A
Okay,
hush.
D
Hey
hello,
so,
as
I
mentioned,
because
of
the
builds
not
being
available,
I
was
a
bit
blocked
there
and
the
DSL
lab
that
that
I'm,
currently
using
for
testing
Crimson
I,
had
set
it
up
for
a
seven
node
cluster
which
had
which
I
had
to
move
to
a
four
node
twister
with
which
addressed
that
Upstream
tracker.
So
once
that
was
done,
unfortunately,
there
were
no
Reef
build
images
of
crimson,
so
I
had
to
go
ahead
with
the
Quincy.
D
Well,
initially,
I
wasn't
sure
if
that
was
going
to
reproduce
the
same
issue
that
I
had
faced
previously,
but
yeah
I
do
I
am
definitely
able
to
reproduce
the
same
problem
here.
I
will
update
the
Upstream
tracker,
but
with
the
information
that
every
requested,
Sam
and
also
Nixon
had
connected
with
me
two
days
ago.
So
I'll
speak
with
him
as
well
with
the
information
that
he
has
asked.
So,
okay,
that's
about
the
update
on
that
Upstream
tracker.
D
Apart
from
that,
there
hasn't
been
a
lot
of
testing
going
hands
on
this
thing
with
Crimson
because
of
all
the
blockers.
Last
week,
so
I
have
been
working
on
the
tutology
Suites.
The
I
haven't
run
them
before
so
I'm.
Getting
up
to
speed
there,
so
I
can
run
the
Upstream
suits
that
are
available
and
understand
what
exactly
is
being
tested
there.
D
Apart
from
that,
there
has
been
a
plant
internalization
for
the
QE
team
on
Crimson
itself,
so
I'm
working
on
some
documentation
there,
so
I
can
deliver
a
session
internally
for
the
QE
team.
That
is
that's
pretty
much.
It.
A
Okay,
thanks
Kevin's
up.
E
C
E
Kevin
from
early
narrow,
so
so
this
is
the
first
time
for
me
to
join
the
Crimson
meeting,
so
I
I
just
want
to
catch
up
with
what
currently
Crimson
is
developing
and
doing
so,
and
hopefully
I
can
help
some
things
in
the
future.
So
currently
I'm
working
on
some
bug
fix
on
the
staff,
I'm
64
issues,
so
I,
actually
I
I
will
build
some
United
text
failures
on
yeah,
okay,.
B
Modify
the
vertical
CSO
according
to
Missy's
comments
and
I
will
give
the
address
that
is
on.
You
found
the
article
OST
and
next
to
it
we
could
meet
him.
We
can
discuss
on
the
the
mean
data
structure
and
the
main
floor.
F
This
week,
I
get
several
little
PR's
for
another
round
of
review
and
I.
Think
I
I
found
the
root
cause
for
the
for
the
memory
leak
that
we
that
I
mentioned
last
week.
Here's
the
results.
It
seems
the
route
causes
that,
since
star
allows
each
CPU
Accord
to
free
pointers
within
itself
within
its
own
64
gigabytes
memory
space.
F
If
the
memory
spec,
if
the
point
is
located
in
memory
spaces,
then
out
outside
this
range
and
it
the
current
CPU
core
will
have
to
notify
another
CPU
core
to
to
release
to
free
it.
But
if
there's
no
CPU
core
available,
then
it'll
be
leaked,
so
I
think
that's
that's
the
case.
Actually
I
think,
as
the
sister
Community
has
already
fixed
this
bug,
if
HW
lock
is
enabled
on
on
the
system,
but
otherwise
it.
F
It
still
have
this
issue
so
I'm
trying
to
right
now,
I'm
testing
the
fix
for
the
for
this
memory
leakage
and
leave
work
if
it
works,
I
think
I'll
submit
it
to
the
sister
community.
That's
all
for
me
this
week
for
next
week.
Okay,
next
week,
I
will
go
on
resume
the
work
on
the
LBA
three-pointer
PR.
That's
all
for
me,
cool.