►
From YouTube: GitLab 14.10 Kickoff - Create:Gitaly
Description
This is the kickoff video for the Create:Gitaly stage for the GitLab 14.10 release.
Planning Issue Link - https://gitlab.com/gitlab-org/create-stage/-/issues/12988
A
A
A
A
We
have
a
couple
things
we're
going
to
send
through
the
problem:
validation,
solution,
validation
cycle
in
the
upcoming
releases.
One
of
those
is
get
for
large
files.
This
is
an
initiative.
We
have
to
figure
out
how
we
can
better
support
those
customers
and
those
users
who
have
very
large
files.
Git
is
not
an
ideal
solution
for
that.
Currently,
it's
very
difficult
to
manage
large
files.
Therefore,
we're
looking
for
opportunities
to
improve
that
experience
for
our
users.
A
In
addition,
we
have
a
few
things
here:
around
solution,
validation.
We
want
to
come
up
with
a
path
forward
on
these
issues.
Here,
probably
the
one
to
highlight
here
is
the
clarifying
the
upgrade
process
for
gili
cluster.
This
is
an
often
requested
thing
from
our
users.
We
want
to
make
sure
that
we
have
not
only
a
recommended
solution-
that's
well
documented,
but
that
recommended
solution
has
been
tested
and
verified
by
our
team
that
it's
the
right
approach
and
it
reduces
as
much
risk
as
possible
during
upgrades
I'm
moving
into
the
features.
A
We
have
a
few
themes
here
that
we're
talking
through
one
of
the
major
themes
in
this
release,
and
it
has
been
for
a
while-
is
repository
level
incremental
backups.
I
have
linked
to
the
epic
here.
I
will
also
link
this
issue
in
the
video
notes
below
one
of
the
most
important
things
we're
looking
to
do
is
have
incremental
backups
use
the
standard,
backup.rake
task
that
we
already
have
as
part
of
gitlab.
A
This
is
the
mvc
of
incremental
backups.
We
expect
to
iterate
on
this
continuously
moving
forward
for
multiple
releases
to
make
this
really
a
stable
part
of
our
product,
we're
looking
to
release
something
in
an
mvc
state
here
in
1410
such
that
we
can
start
getting
user
feedback
and
allow
users
to
experience
the
benefits
of
incremental
backups
over
the
very
long
backup
task.
When
you
have
a
very
large
repository.
A
The
next
theme
that
I
want
to
touch
on
here
is
the
improving
reliability
and
scalability
on
the
replication
queue.
The
replication
queue
is
responsible
for
replicating
your
data
from
one
gilly
cluster
node
to
another.
So
we
have
a
carryover
issue
from
the
previous
release
from
14.9,
where
we're
working
to
reduce
and
avoid
duplicate
jobs
being
queued
up
for
replication
to
reduce
bandwidth
and
reduce
load
on
the
servers.
A
The
final
thing
that
I
want
to
touch
on
here
is:
we
also
have
a
theme
for
moving
repository
maintenance
ownership
into
italy.
This
has
been
ongoing
for
a
while.
This
is
something
that
you
may
not
notice
immediately
in
through
use,
but
we're
restructuring
a
lot
of
our
rpc's
and
we're
removing
those
rpcs
that
are
no
longer
necessary
and
we're
trying
to
kind
of
standardize
on
a
consistent
set
of
rpcs,
allowing
the
best
interface
possible
and
allowing
the
best
testing
possible
by
reducing
the
number
of
rpcs.
A
A
If
you
have
any
questions
upon
any
of
these
things,
please
feel
free
to
reach
out
to
us
via
comment
on
an
issue
or
directly
via
email
to
myself
or
andrus
the
engineering
manager
or
anybody
on
the
team.
We'd
love
to
hear
from
you
and
hear
your
feelings
on
this,
as
always
I'll
link
this
below.
If
there's
any
questions
or
comments,
please
again
feel
free
to
reach
out.
Thank
you
very
much
have
a
great
day.