►
Description
Presented by: Nizamudeen A
Event: https://ceph.io/en/community/events/2022/ceph-virtual/
A
Hello,
everyone
I,
am
nizamudi
I'm,
going
to
talk
briefly
about
operating
safe
from
the
SF
dashboard
fast
present
and
diffuser
a
bit
of
History
lesson
here
so
I'll
be
covering
this
in
a
two
part,
slides
two-part
session.
So
part
one
will
cover
the
introduction
y,
a
dashboard
previous
attempts
in
the
dashboards,
like
the
open
ethic,
one,
the
dashboard
version,
one
a
little
bit
on
the
architecture,
the
key
features
on
qnc
and
the
roadmap
to
reef,
and
hopefully
a
demo
as
well.
A
No
I
am
currently
a
software
engineer
at
Red,
Hat
I
joined
the
safe
dashboard
team
in
around
2020,
so
either
on
the
time
of
Corona
and
I've
been
working
with
them
ever
since.
Before
that
I
used
to
work
with
the
Rook
operator
and
it's
Downstream,
the
Downstream
with
the
OCS
operator
and
all
I'm,
also
the
current
component
lead
of
the
safe
dashboard
I've.
A
A
dashboard
for
safe,
except
dashboard
is
like
a
fully
fledged
management.
Ui,
where
you
know
all
kinds
of
Advanced
self-management
operations
can
be
done
like
it
is
like
a
management
plus
monitoring,
UI
Tool.
It
has
all
these
Advanced
set
of
features,
and
it
is
not
just
limited
to
the
monitoring
part
and
why
just
why?
The
dashboard
part
right-
and
why
did
we
remove
the
dashboard
from
here,
because
in
luminous,
when
we
actually
created
the
SF
manager
module
and
the
dashboard
was
introduced,
it
was
actually
a
dashboard.
A
There
was
not
a
lot
of
management
going
on
in
the
at
the
time,
so
we
kept
the
name
dashboard
and
when
the
dashboard
version
2
was
introduced,
it
was
a
lot
of
trouble
to
change
the
dashboard
to
you
know
some
thing
else,
and
otherwise,
like
a
lot
of
references
in
the
documentation
in
the
code,
so
we
just
kept
the
name
there.
But,
however,
the
dashboard
name
doesn't
do
justice
to
the
actual
extent
of
the
safe
dashboard,
because
it
is
actually
a
fully
fixed
management.
Web
UI
tool
so
coordinates
the
dashboard
right.
A
You
know
distributed
systems,
as
we
all
know.
It's
like
very
complex
and
very
complicated
thing.
Like
you
know
the
installing,
as
you
can,
as
you
know,
the
installing
SF
is.
It
is
not
just
about
installing
this
stuff
right.
It
is
about
installing
all
the
components
of
this
app,
also
like
the
RBD,
it
redoes,
the
Azure
wsffs.
Everything
else
right-
and
these
are
all
this-
is
not
just
a
monolithic
quiz
and
it's
all
completely
independent.
A
So
we
need
to
install
them,
we
need
to
monitor
them
and
and
we
need
to
continually
manage
building,
so
it
is
better
to
do
with
the
UI
and
it
also
involved
configuration
of
all
these
things
right,
like
you
know,
in
the
main
branch
at
least,
the
configuration
settings
like
in
involves
around
1919
configuration
settings
which
applies
for
all
these
different
components,
and
you
know
if
we
dump
them.
It
is
not
actually
exactly
pretty
nice
to
look
at,
but
if
you
go
into
the
safe
dashboard
UI
it
is
listed
in
a
table.
A
We
can
search
for
the
configuration
settings
we
can
list
them.
We
can
you
know,
filter
them
apply,
apply
filters
to
the
configuration.
It
is
easier
to
see
the
configurations,
the
settings
in
the
dashboard
and
if
we
talk
about
operating
the
every
different
ecosystems
in
the
safe
world,
it's
like
you
know,
every
component
has
its
own
serialized
apis
and
we
need
to
be
consistent
about
around
them
and
maintaining
these
cluster
is
actually
you
know
a
little
bit
easier
with
the
you
know.
A
The
management
UI
troubleshooting
enforces
you
know
looking
into
the
logs
and
you
know,
adjusting
the
settings
or
monitoring
networking
all
these
kinds
of
stuff,
and
actually,
when
it
comes
to
troubleshooting,
we
actually
introduced
a
feature
into
the
safe
dashboard
recently,
which
is
like,
which
is
like
a
centralized
logging,
whether
user
can
go
into
a
particular
area
and
into
in
the
dashboard
and
view
all
the
you
know,
container
logs
in
a
in
a
single
location
and
safe
CLI
doesn't
provide
a
unified
user
experience.
It's
it's
particular
right.
A
There
is
like
a
different
CLI
for
HW
rbds
of
ADM
and
in
the
dashboard,
it's
all
unified,
the
user
experience
with
with
a
single
dashboard.
You
can
go
into
all
these
areas
and
you
can
manage
you
can
monitor
whatever
you
want,
then,
comparing
the
CLI
with
the
you
know
the
the
the
the
dashboard
like
in
the
protocols
like
this
SSH
versus
the
https.
It
is
more
easier
to
you
know,
use
the
HTTP
protocol
instead
of
the
SSH
and
it's
the
configuration
is
easier.
A
A
A
Then
there
is
a
safe
Dash
which
is
a
community
project
and
then
the
ink
scope
one
and
then
there
is
a
self
manager.
Then
in
2016
the
SF
manager
got
introduced
into
this
F
and
with
the
help
of
that,
you
know
the
safe
dashboard.
The
first
version
of
this
F
dashboard
got
introduced
into
into
the
manager
as
a
manager
module
inside
the
chef
and
in
2018
DSF
dashboard
version
2,
which
is
like
a
fully
featured
management.
Ui
tool
came
into
the
manager
module
of
this
F
uck
2019.
A
So
we
have
like
a
two
one
or
two
guided
you
Workforce
in
the
UI
itself,
and
then
we
improve
the
serviceability
as
well
at
the
centralized
loading
we
will
come
as
part
of
that
and
we
are
continuing
to
improve
the
usability
experience
overall
on
the
dashboard
right.
You
know
in
the
coming
like
right
now
as
well
as
well.
A
These
are
all
the
common
dashboards
that
were
around
in
the
you
know
in
the
safe
Community
like
we
have
all
these
dashboard
in
the
from
the
previous
eras,
and
when
we
look
at
the
safe
dashboard,
it
is
powered
by
the
Safe
Community
itself,
so
the
maintenance
or
the
The
Continuous
Improvement
and
everything
is
powered
by
the
Safe
Community.
It's
it's
still
active
and
it
is
going
to
be
active.
So
it
started
around
2018
and
it
is
still
going
very
strongly
and
we
have
SF
management.
A
You
know
very
Advanced
set
of
self-management
operations
going
on
inside
because
we
actually
directly
consume
the
safe
manager,
the
modules
apis
from
the
SF.
You
know
from
our
backend
itself,
so
it's
it's
much
easier
to
communicate
with
them
without
any
lags
in
between,
and
we
also
embedded
the
graphina
inside
the
dashboard
so
that
you
know
user,
don't
need
to
go
into
different
dashboards
to
view
the
graph
Dynamics
or
the
Prometheus,
Matrix
or
and
the
grafana
charts,
and
so
everything
is
inside
the
chef
dashboard
as
well.
A
This
is
the
calamari
dashboard,
the
one
of
the
first
dashboard
that
came
for
this
if
the
open
attic
one
developed
by
suse.
So
this
has
like
a
you
know
this
stuff
dashboard.
The
first
version
of
this
app
dashboard.
Has
you
know
the
second
version
of
the
safe
dashboard
has
routes
into
the
openactic
project.
So
if
you
look
at
the
earlier
version
of
the
self
dashboard,
it
kind
of
looks
exactly
or
somewhat
similar
to
this
openactic
dashboards.
A
This
F-1,
which
is
the
which
is
a
community
or
open
source
project,
I,
think
it
is
still
active
but
somewhat
active,
but
not
that
active.
Then
we
have
the
safe
dashboard
version.
One
mostly,
you
know
monitoring
activities
only
I,
don't
think
there
are
a
lot
of
management
stuffs
going
on
over.
Here
we
can
see
the
logs
and
all
those
stuffs
in
different
print
UI,
but
I
like
the
dark
theme
Here-
and
this
is
the
self
dashboard
version
to
us-
we
see
in
the
you
know,
frequency
and
the
main
brands
of
the
stuff.
A
As
you
can
see,
there
are
a
lot
of
changes
over
here.
I
can
I'll
show
a
demo
in
the
end
so
that
this
will
you
can
and
then
I
will
try
to
explain
mostly
everything-
and
this
is
the
proposed
and
agreed
design
for
the
safe
dashboard
version
3,
which
will
be
coming
soon
in
hopefully
in
the
reef
release
fingers
crossed.
A
So
the
works
are
ongoing,
mostly
done.
We
are
patching
up
some
of
the
charts
over
here
and
with
that
we
will
be
able
to
release
this
in
in
the
oncoming
release,
so
how
safe
users
monitor
the
separate
most
of
these
have
users.
This
is
a
survey
results
from
the
deaf
user
survey
result
like
you
know.
The
2022
happened
in
2022,
so
most
of
the
users
use
the
more
Prometheus
to
keep
track
of
the
monitoring
activities
in
the
cluster.
A
A
The
safe
dashboard
version
2..
So,
as
I
said
previously,
safe
dashboard
has
safe
dashboard.
The
version
2
has
roots
into
the
open
ethic
project
and
when
it
combined
with
the
safe
manager
and
this
dashboard
version
2,
we
got
the
safe
dashboard
version
2,
as
we
are
seeing
right
now.
So
since
luminous
theft
offers
theft
itself
offers
what's
needed
to
develop
a
fully
featured
and
smoothly
integrated
dashboard.
So
this
manager
provides
an
in
memory
cache
for
all
core
data
versus
highly
efficient
interface.
A
To
save
you
know,
based
on
python
course,
and
the
safe
manual
safe,
monitor
will
also
provide
a
highly
available
persistence
for
the
small
data
sets,
so
it
makes
easier
for
us,
you
know,
and
it
makes
it
efficient.
This
dashboard
version
2,
combined
with
the
set
manager
it
makes
everything
efficient.
A
A
There
is
SF
manager,
module
API
and
when
the
front
end
of
the
dashboard
requests
some
data,
it
goes
through
the
rest
API,
which
is
set
in
Python
and
communicate
with
the
module
API
and
that
fetches
the
data
from
this
F
cluster,
of
course-
and
you
know,
within
a
cached
manner,
it
provides
the
information
that
needs
to
needs
to
populate
the
front
end
of
the
dashboard.
Again
there
is,
you
know
there
is
a
promise
exporter
which
exports
all
the
pro
from
metrics
and
populates
to
the
grafana.
A
You
know
from
popularity
to
the
Prometheus
and
it
again
encoded
by
grafana
and
then
the
graph
finer
dashboards
are
again
embedded
into
the
into
the
safe
dashboard
itself,
and
between
that
there
is
like
you
know.
Recently
we
started
consuming
this
aphidium
for
some
of
the
you
know
day,
two
management
applications,
so
it
makes
the
communication
easier
for
us
to
directly
go
to
the
core
than
you
know,
consuming
the
apis
directly,
so
consequency
as
you
can
see,
we
have
achieved
a
lot
of
things.
A
Like
you
know,
most
of
the
most
of
the
components
are
like
you
know,
are
closely
in
priority
with
the
CLI,
so
we
have
in
the
configuration
Civic.
As
you
can
see,
we
we
covered
almost
95
percent
of
percentage
of
it
in
osds.
We
had
a
like.
You
know,
somewhat
somewhat
a
little
work
to
do
in
the
OST
spot.
The
pools
are
almost
covered
in
the
RBD.
There
were
like
some
missing
features
which
are
available
in
the
CLI,
but
not
in
the
dashboard
as
a
w.
A
It's
like
always
playing
catch
up
with
the
HW
Fox
because
they
introduce
new
features
constantly
and
we
are
always
trying
to
catch
up
with
the
RSW
folks.
So
it's
ever
growing
the
otherw
area
ffs.
We
have
all
the
monitoring
available
monitoring
functionality
available
in
the
surface,
but
not
just
the
management
part
and
we
are
planning
to
take.
A
You
know
actions
over
there
in
the
coming
releases,
not
maybe
the
reef,
but
you
know
after
the
reef
we'll
be
focusing
on
this
ffs
for
the
reef
release
we'll
be
focusing
on
the
rjw
and
trying
to
introduce
the
features
like
the
multi-site
or
things
like
that
and
the
orchestration
we
are,
you
know,
started
improving
the
orchestration
part
from
the
fluency
releases
and
we
made
some
significant
progress
over
there.
If
you
compare
it
with
the
reef,
we
improved
the
osds
area
by
introducing
some
OSD
creation
wizard
into
the
UI.
A
The
RBD
is
almost
complete
by
introducing
some
missing
features
like
the
RBD
snapshot,
mirroring
and
promotion
and
demotion,
for
you
know
ability
to
to
do
that
from
the
UI
and
in
the
rjw
we
introduced
some
rgw
server
set
encryption
so
and
over
here
you
can
also
see
some
of
the
other
features
that
we
went.
You
know
delivered
into
the
Reef
releases
like
the
centralized
logging.
We
also
made
the
dashboard
compatible
with
the
you
know:
accessibility,
wcag
level,
A2
level,
double
a,
and
we
also
you
know,
introduce
the
safe
auth
management.
A
We
will
be
introducing
them
them
in
the
oncoming
Reef
release.
Oh,
who
goes
beyond
the
reef,
so
I'll
talk
about
the
low
code
initiative.
Briefly,
in
the
end,
so
I
can
explain
what
this
local
initiative
means,
so
the
replacing
grafana
with
built-in
search
so
right
now
the
grafana
is
populated
into
the
dashboard
using
a
iframe
which
is
good,
but
not
that
you
know
I
can
see
like
you
know.
So
we
are.
We
are
planning
to
replace
them
with
built-in
sets
and
directly
consume
the
Prometheus
Matrix.
A
A
Okay,
I'll
try
to
show
a
demo
and
after
the
call
I
will
be
after
the
entire
session.
I
will
be
planning
to
take
some
questions
as
well,
but
I'll
try
to
show
the
demo.
Let
me
switch
to
my
dashboard
I
hope
it
is
visible.
So
this
is
the
login
page
of
the
SF
dashboard
when
you
go
into
the
dashboard
using
the
username
and
login.
The
first
thing
you
will
begin
with
is
the
you
know
the
as
please
expand
your
cluster
first
notification.
A
If
you
just
bootstrap
your
cluster
and
there
is
nothing
else
added
to
the
cluster,
this
will
be
the
first
thing
that
is
shown
and
right
now,
I
have
some
something
some
hosts
are
added
into
the
cluster,
but
I
anyway.
Trigate
this
to
happen,
so
you
can
anyway
add
or
host.
If
you
want
a
in
here-
and
you
can
also
create
more
osds,
you
can
also
add
more
services
and
in
the
end
you
can
review
everything
else
and
you
can
click
on
the
expand
cluster
and
it
will
expand
the
cluster.
A
So
basically,
all
the
host
and
osgs
you
added
will
be
added
once
you
once
the
expansion
is
completed,
and
this
is
the
dashboard
you
can
see
the
cluster
status
over
here
with
a
little
right
now.
I
have
some
warning.
So
if
you
click
on
the
warning,
there's
like
a
little
pop-up,
which
shows
why
the
cluster
is
in
warning
stage.
I
can
go
to
the
logs
and
see
why
you
know
why
it
is
in
one
next
state
and
things
like
that.
A
A
A
Let
me
add
that,
so
in
the
host
management,
we
have
the
everything
we
can
manage
the
host.
Basically,
we
can
create
post,
we
can
edit
the
host,
we
can
remove
a
host
and
we
can
add,
you
know
we
can
put
a
host
in
the
maintenance
and
we
can
you
know
drain
her
host
completely.
You
can
remove
them
and
we
there's
also
a
performance
details.
A
This
is
the
iframe
component
that
I
said
we
are
directly
embedding
the
graphina
here
there
you
can
see
the
the
performance
of
the
individual
host
and
also
the
overall
performance
of
the
all
hosts
like
I,
have
like
three
OSD
hosts
and
things
like
that
can
be
seen
in
the
overall
performance.
We
have
the
physical
disk
page,
which
lists
all
the
available
disk.
A
The
monitor
space,
for
you
know
showing
all
the
monitors
Services
plates
to
create
and
manage
the
services.
You
can
also
see
the
diamonds
and
do
the
data
management
actions
like
you
know,
in
a
stop
and
restart
or
redeploy
the
services
or
diamonds.
Here
when
you
expand
them,
you
can
see
that
in
the
OSD
section
again
we
listed
all
the
osds
and
it
has
like
a
very
Advanced
set
of
features
in
the
osds,
like
you
know:
scrub,
deep
scrub,
rebate,
more.
B
A
Out
marking
everything
that
you
are
doing
in
dcla,
it
can
be
done
in
the
management
in
this
this
UI
we
have
the
cluster
wide
configuration
the
overall
performance
again
that
populates
from
the
grafana
again,
the
individual
OSD
performance
details,
the
devices,
the
performance
counters.
Everything
is
listed
in
the
UI,
as
I
said,
the
configuration
UI,
which
is
filtered
by
you,
know
by
default
basic
the
crush
map,
the
manager
modules.
All
the
management
modules
are
listed
here.
This
is
the
suffix
users,
which
is
which
is
currently
a
work
in
progress
thing
right
now.
A
It
says
it
will
be
easier.
This
way,
at
least
with
the
for
this,
you
need
to
enable
the
create
services
like
the
blocky
and
promptail,
and
we
have
the
monitoring
and
alerting
going
on
Section
going
on
over
here,
I'll
be
quickly
going
over
every
every
Pages.
Like
you
know
the
course
management
going,
Force
management
is
listed
here,
create
and
delete,
and
edit
edit
tools,
the
overall
performance
of
the
polls
block
for
the
RBD
images,
and
you
know.
B
A
As
I
said,
the
file
system
ffs
is
mostly
monitoring.
There
is
not
a
lot
of
management
going
on
over
here,
but
you
can
see
still
see
out
of
the
monitoring
related
activities.
Here
we
can
create
a
snapshot
for
the
directories.
You
can
evict
clients
and
you
know
the
empty
Express
counters
and
all
those
details
how
it
will
be
here.
The
object
Gateway
includes
the
payments,
the
users
and
the
buckets
and
in
the
buckets
we
can
now
configure
the
encryption
of
the
packets
as
well,
and
here
we
have
the
notifications.
A
All
the
notifications
will
be
shown
here,
including
the
alerts
as
well
as,
and
you
know,
default
SF
notifications.
On
the
safe
dashboard
notifications.
We
have
an
option
to
report
an
issue
from
the
dashboard.
There
is
a
way
to
configure
the
Telemetry
from
here.
There's
like
a
two-step
process.
You
click
on.
A
You
know
next
and
if
you
accept
the
you
know,
settings
from
the
Telemetry
update
from
here
and
it
will
update
you
know,
it
will
basically
accept
the
Telemetry
and
configure
the
Telemetry
for
you
guys,
yeah,
that's
mostly
it
from
from
managing
perspective.
You
know
if
there
is
any
questions,
please
ask
in
the
comments
and
I
will
try
to
address
them
in
the
end.
A
A
So
as
a
user,
the
dashboard
is
enabled
by
default
with
this
fadm.
If
it's
not,
you
can
enable
the
dashboard
by
you
know
enabling
the
dashboard
module
and
basically,
following
all
these
steps
like
creating
the
self-cent
certificate,
creating
the
dashboard
user-
and
you
will
be,
you
know,
enabled
with
a
setup
with
a
safe
dashboard.
A
So
this
is
the
login
page
of
the
self
dashboard.
So
if
you
have
any
query
any
questions
like
you
can
always
reach
us
all
this.
If
users
mailing
list
and
also
we
are
also
reachable
on
the
IRC
or
FTC
Channel,
like
the
dev
dashboard
and
as
I
shown
earlier,
like
you
know,
we
have
a
report
on
issue
section
in
the
dashboard.
A
So
if
you
enable
the
feedback
module
and
if
you
click
on
the
report
and
issue
from
here
and
add
this
F
tracker
API
key-
you
know
here
and
basically,
if
you
explain
the
issue
and
click
on
submit,
it
will
basically
create
a
Tracker,
safe
tracker.
For
with
that
issue,
that'll
be
easier.
You
know.
If
you
want
to
do
that
and
as
a
documenter
we
show
the
documentation
of
the
dashboard
in
this
URL.
You
can
report
issue,
you
can
also
edit
and
image
the
changes
by
yourself.
A
If
you
click
on
this
edit
on
GitHub
icon
or
you
can
also
report
a
documentation
bug
and
as
a
translator,
you
can
help
like
you
know.
Right
now
we
have
support
around
17
languages
and
the
dashboard
transfix
dashboard
is
listed.
You
know
you
can
you
can
basically
view
them
in
in
this
URL
I'll
share
the
slides
as
well
in
mailing
list,
and
you
can
go
over
them.
A
So
as
a
developer,
there
is
like
a
mailing
list
for
developers
at
safe.io.
We
have
an
IRC,
SF
dashboard
and
we
have
developer
guides,
for
you
know
the
safe
developer
type
guide
and
as
well
as
the
safe
dashboard
developer
guide.
We
have
the
backend
setup
in
Python,
3.6
different
and
mostly
written
in
angular,
12
and
bootstrap
4,
which
will
change.
We
will
be
using
the
angular,
13
and
bootstrap
5.
So
that's
the
new
features
in
the
angular
and
bootstrap
version
will
be
available
to
the
dashboard.
A
So
what
is
the
local
initiative
right?
So
generally,
the
UI
component.
We
are
planning
to
generate
the
UI
component
from
the
back
end.
Only
so
no
typescript
or
no
angular
skills
are
needed.
So
that's
the
end
goal
that
we
are
trying
to
achieve.
So
you
know
front
end,
will
just
call
the
API
like
here
as
you
can
see,
from
the
API
dot
cluster.user
and
it
will
populate
the
entire
table
or
in
the
entire
form
and
basically,
in
the
back
end
we
will
be
providing
some
self
descriptive
python.
A
You
know
classes
and
everything
and
that
will
you
know,
generate
different
UI
table.
So
that
is
the
Encore
we
are
trying
to
achieve.
So
no
UI
code
will
be
there
so
that
everyone
without
a
UI
skills
can
contribute
to
the
dashboard.
So
that's
the
right
now.
The
some
of
the
features
are
like
depth
order
management.
We
already,
you
know,
implemented
using
this
approach
and
since
March
2018
we
have
around
20
2700
full
request.
A
4200
commits-
and
you
know
one
like
21
000
lines
of
code
added
to
the
dashboard,
and
this
is
the
all
the
brains
behind
the
dashboard
and
all
the
hands
behind
the
dashboard,
and
we
have
people
from
all
over.
You
know
more
than
25
countries
from
four
different
organizations:
red
hat
suse
and
Fujitsu,
and
people
from
the
outreachy
Google
summer
of
code
and
people
from
10
different
countries.
Basically,
we
we
have
stand-ups
around
daily
11
30
a.m
to
each
time
and
Upstream
sink
on
Tuesdays
face
to
face
happened
three
pandemic
times.
A
We
don't
officially
support
the
mobile
devices.
We
only
support.
You
know,
resolutions
up
to
1920
into
full-less
Data
Solutions,
but
I've
seen
in
mailing
list
some
people
trying
to
do
that.
I
I
personally
tried
on
my
mobile
as
well.
We
can
do
some
things
over
there.
You
know
if
it's
simple,
like
you
know,
tracking
the
self-health
or
status
like
that
I
mean
it's
you
know
is
we
can
do
that
from
the
mobile
as
well.
So
anyway,
the
UI
will
come
in
the
link,
but
it
won't
be
nice,
it's
just.
A
A
B
A
A
A
For
one
thing,
I
would
recommend
you
know
the
new
users
that
are
coming
to
the
safe
world
to
use
the
dashboard,
because
it's
more
easier
to
configure
the
dashboard
using
the
dashboard
I
mean
with
the
software
DM.
We
can
bootstrap
the
cluster
right
and
once
that
is
done,
if
you
go
to
this
dashboard,
we
can
just
configure
the
entire
cluster
using
the
guided
workflow
itself.
So
it
will
be
much
easier
for
them
to
be
honest
and
yeah
I
I
hope
that
covers
the
questions
you
asked.