►
From YouTube: GitLab 13.7 Kickoff - Verify:Runner
Description
Kickoff video for GitLab Runner 13.7
https://gitlab.com/gitlab-org/gitlab-runner/-/issues/27188
https://about.gitlab.com/direction/verify/runner/
A
Hey
everyone
darren
eastman
growth
manager
for
gitlab
runner,
hey
so
today,
I'm
just
going
to
cover
what's
on
deck,
for
gitlab
runner
for
13.7,
so
my
screen,
I'm
currently
sharing
the
397
runner
team
iteration
plan,
which
this
is
a
mechanism
for
us
to
kind
of
show
you
what's
kind
of
our
key
focus
areas
as
we
get
into
the
food
net.
Seven
development
iteration
milestone.
A
As
always.
This
is
sort
of
a
quick
sort
of
recap
of
quick
aside.
If
you're
interested
in
the
broader
vision,
the
broader
long-term
strategy
for
gitlab
runner,
I
suggest
you
take
a
look
at
our
category
direction,
page
for
the
runner,
which
is
at
above.github.com
forward
slash
direction,
power,
slash,
verify
forward,
slash
runner,
and
here
you
can
see
we
were
talking
about
some
of
the
visionary
items,
some
of
the
long-term
picture
items
that
we
have
in
mind
for
evolving,
get
lab
runner
and
things
related
and
get
that
on
the
product.
A
But
for
today
we're
focused
on
what's
on
deck,
specifically
for
13.7.
So
the
first
thing
I
want
to
kind
of
focus
on
are
the
items
that
we
are
working
on.
So
a
few
key
items
that
we're
working
on
for
the
runner
core,
which
is
the
core,
get
lab
runner
binary.
A
What
we're
working
on
right
now
with
food
note
7?
Is
this
mvc,
so
just
kind
of
just
a
quick
step
back
and
sort
of
say
grounding
where
we're
at
so.
Over
the
past
year
we've
been
working
alongside
the
red
hat
engineering
team
on
an
operator
that
will
deploy
gitlab
run
on
openshift
and
we'll
work,
also,
of
course
in
kubernetes.
A
So
the
first
version
of
of
the
operator
was
done
roughly
in
the
june
time
frame,
however,
was
very
restrictive
due
to
any
id
restrictions,
and
we
can
get
into
some
of
the
details
on
that
on
a
later
video
or
you
can
you
can
ping
me
directly
in
the
issue
if
you
have
any
questions
about
what
that
means,
but
what
we're
working
on
in
397
is
we're
working
on
finalizing,
adding
deductible
files
with
the
required
permissions
to
run
with
our
new
id
on
openshift.
A
With
this
work
complete,
we
can
officially
announce
a
beta
version
of
the
github
run
on
autoship
on
openshift
and
one
that's
fully
supported
by
gitlab.
I
think
again,
it's
kind
of
one
of
the
iterations
on
our
way
to
offering
a
fully
supported
first-class
experience
of
get
lab
runner
on
openshift
in
ga
by
the
january
2021
time
frame
now
specific
to
397
and
the
work
that
we
have
on
deck.
A
We
do
have
a
few
engineering
challenges
that
will
potentially
implement
or
introduce
some
delivery
risk
to
us
getting
this
over
the
over
the
fence
and
13.7,
and
this
is
due
to
some
core
os
dependencies
in
our
ci
workflows.
So
hopefully
we
can
work
through
this
and
get
the
the
beta
version
of
the
runner
for
right
at
open
shift
out
and
the
net
seven
the
next
issue.
I
want
to
call
your
attention
to
that.
A
We're
working
on
for
runner
core
in
philippines,
seven
is
to
remove
usage
of
umass
zero,
zero
issued
or
to
remove
usage
of
new
masters
or
to
be
specific
to
be
secure
like
default
issue,
and
let
me
see
if
I
can
explain
a
bit
what
that
means
and
what
this
is
all
about.
So
at
a
high
level
today,
in
our
code
base,
there
is
a
command
umass,
zero,
zero,
zero.
I
believe
it
will
see
what
the
command
is
called
and
that
command
runs
before
the
cloning
process.
A
Inside
of
the
github
on
a
helper
image
right,
this
command
was
introduced
to
basically
solve
a
permission
problem
for
container
images
that
do
not
have
the
root
user.
Yes,
and
we
do
recognize
that
having
not
using
a
root
user
for
your
container
images
is
a
security
best
practice.
A
However,
this
command
as
implemented
today
changes
the
permissions
on
the
chrome
directories,
so
directories
that
are
referenced
in
our
gitlab
runner.
Environment
variables,
sci
underscore
project
underscore
dir
or
ci
underscore
bills
underscore
dirt.
Those
directories
are
able
to
be
read,
written
or
executed
by
any
user
because
of
the
implementation
of
the
current
humanities:
zero
zero
command
if
the
containing
image
there.
So
the
proposal
on
day
for
15.7
is
to
implement
logic
that
before
running
the
user
script
we'll
check
to
see
if
the
container
image
will
run
as
a
rule
or
not.
A
If
the
container
image
does
not
known
as
root,
then
we'll
retrieve
the
uid
or
good
of
the
user
and
change
the
owner
of
the
corresponding
directory.
So
now,
on
the
surface,
I'm
just
saying
backup
really
fast
for
those
folks
that
are
interested
in
the
actual
details
of
that
proposal.
We're
implementing
it's
it's
covered
here
and
there's
a
nice
little
workflow
that
the
engineers
added
to
this
issue
that
describes
the
logic
that
we're
in
that
we'll
be
implementing
in
397..
A
So
on
the
surface,
this
doesn't
seem
like
a
complex
or
very
interesting
new
feature,
but
actually
in
actuality.
A
This
is
a
somewhat
of
a
complex
change
and
so
there's
a
significant
amount
of
testing
that
we
have
to
do
to
ensure
that
we're
not
introducing
any
regressions
to
this
core
functionality
of
doruna
when
we
introduce
this
just
change
to
the
to
the
codebase,
and
so
you
can
see
here
in
the
issue
under
the
implementation
section,
there
are
some
additional
details
that
are
called
down
here
that
are
meant
to
help
us
mitigate
the
risk
of
introducing
a
regression
again,
one
that
would
be
detrimental
to
the
core
functionality
of
run
if
we
could
actually
amend
the
regression
here.
A
A
The
third
thing
we
are
looking
at
in
397
that
I
want
to
call
your
attention
to
for,
for
ronald
call.
Is
this
issue
around
creating
a
pull
policy
that
can
fall
back
on
local
cash,
just
gonna,
just
kind
of
step
back
and
give
a
bit
more
context
about
this
particular
issue?
A
A
A
And
so
this
is
what
this
feature
proposal
is
really
about
again
in
introducing
a
new
pro
policy
that
says
hey
if,
during
the
connection
to
the
target
container
registry,
that
connection
fails
and
failed,
then
fall
back
to
a
local,
locally
cached
copy
of
the
container
image.
If
that
one
is
available,
so
at
a
high
level,
I
just
wanted
to
say
that
there
continues
to
be
internally
here.
A
We
get
live
a
lot
of
in-depth
technical
discussions
with
all
of
the
teams
that
are
involved
on
this
as
to
as
it
relates
to
the
effectiveness
of
this
feature,
as
well
as
the
potential
security
implications
of
introducing
such
a
feature,
so
that
you
know
so
we'll
continue.
Looking
at
this.
A
As
we
get
into
seven
and
and
if
you're
interested
in
this
feature
and
the
progress
as
we
go
through
the
thinner,
seven
iteration
cycle,
I
do
suggest
that
you
follow
along
here
with
this
features
two
six,
five,
five,
five,
eight
or
add
any
comments
to
it.
So
it's
a
key
capability
that
we
have
heard
from
a
few
customers
that
they
think
will
be
super
helpful
in
helping
them
with
mitigating
the
risk
of
those
connection
failures.
A
We
have
a
significant
number
of
customers
already
participating
in
the
closed
beta
of
the
macpos
build
out,
and
we
certainly
understand-
and
I've
heard
from
a
number
of
customers
that
their
needs
for
mobile
development
has
accelerated
in
the
past
year.
A
lot
more
firms
are
developing
only
to
accelerate
development
of
features
and
functionalities
for
their
mobile
applications.
So
we
understand
that
folks
that
are
really
needing
us
to
to
add
this
capability
to
our
gitlab.com,
offering
so
I
said
right
now
we're
in
closed
beta.
A
We
have
a
number
of
folks
testing
out
the
current
solution
in
our
closed
beta
as
part
of
the
closed
beta
and
what
we're
working
on
and
to
enable
us
to
move
to
the
open
beta,
which
right
now
is
roughly
targeted
at
the
to
do
to
be
ready
by
filming
that
one
zero
is
this
work
on
the
auto
scaler
and
again
with
the
auto
scale.
A
Right
so
with
397,
we'll
be
working
on
getting
making
the
iterative
steps
required
to
get
the
auto
scaler
done
so
that
we
can
transition
to
the
mac,
os,
open,
beta
and
finally,
in
terms
of
the
other
bucket
in
terms
of
the
three
pillars
as
it
relates
to
the
running
product
development
strategy.
Under
the
enterprise
management
bucket,
we
don't
have
anything
planned
to
ship
and
fit
in
that
seven.
A
Our
cable
for
fitness,
seven
is
making
progress
on
the
ui
design
around
an
automated
recent
reset
of
instant,
wide
mono
registration
tokens
and
just
to
step
back
and
get
a
step
back.
A
bit
running.
Enterprise
management
is
really
our
bucket.
That
captures
things
related
to
to
managing
runner's
enterprise
scale,
as
well
as
features
and
capabilities
in
the
gitlab
ui,
as
it
relates
to
the
running
ui
as
it
relates
to
the
user
experience
in
the
ui
and
also
management
capabilities
in
the
ui.
A
And
so
we
recognize
that
a
core
component
is
simplifying,
not
only
the
user
experience
in
the
ui,
but
also
ensuring
that
we
can
manage
the
experience
and
enterprise
scale
so
again
for
307.
Nothing
planned
to
ship
that
around
the
enterprise
management
theme.
But
we
want
to
make
progress
on
some
ui
designs
again.
Iterative
improvements
and
those
designs
will,
as
it's
part,
of
our
long-term
strategy
in
terms
of
simplifying
the
enterprise
management,
experience
for
gitlab
runner
and
plus
making
it
easy
for
you
for
folks
to
manage,
runs
at
scale.
A
So,
roughly
speaking,
it's
kind
of
to
recap:
fitness
7,
runacore,
we're
working
on
the
red
hat,
openshift,
nvc
or
beta,
and
that's
critical
for
us
and
looking
forward
to
getting
that
out
the
door
and
food
not
seven,
so
that
we
can
just
transition,
hopefully
to
a
ga
product
as
early
as
january
of
next
year,
we're
working
on
a
key
security
feature
for
the
poor
owner
product,
which
is
the
the
umass
zero
zero
zero
feature
that
I
talked
about
earlier
and
then
the
other
thing
that
is
key
for
us
as
well
and
critical.
A
Is
this.
This
new
poll
policy
that
I
talked
about
talked
about
earlier
again
related
to
entry,
introducing
some
resiliency
to
the
way
docker
images
are
pulled
and,
of
course
those
images
are
critical
to
you
to
your
ci
workflows.
A
So
that's
the
focus
series
for
runner.
The
key
focus
here
for
the
gitlab
one
or
two
things,
not
seven.
If
you
have
any
questions
about
things
in
our
broader
strategic
strategic
direction,
forget
lamb,
runner
or
things
that
are
currently
in
around
our
backlog,
on
our
kanban
board
and
simply
curious
in
terms
of
where
those
things
stand
and
how
quickly
you
might
get
to
them
and
so
on
feel
free
to
print
me
as
always
directly
on
the
issue
or
drop
me
an
email,
the
eastman
gets
live.com
anyway.
A
This
is
going
to
be
our
last
kickoff
before
the
the
us
thanksgiving
holidays,
I'll
be
back
again
in
december
for
another
kickoff.
So
for
those
folks
in
the
us
and
happy
thanksgiving
and
look
forward
to
seeing
you
soon.