►
From YouTube: 2023 08 29 Jenkins Infra Meeting
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
So
Stefan
go
go,
go
for
you,
for
maybe
you
will
have
to
restore
the
builds
and
merge
and
deploy
Etc,
or
we
can
add
the
micro
disk
to
update
CLI,
ready
to
roll
for
infra
and
next
release.
Item
list
ongoing,
so
I,
assume,
Kevin
and
homework
and
all
others
will
take
care
of
the
last
item.
Especially
change
like
is
that
correct.
B
A
So
that's
all
for
the
weekly,
so
no
problem,
which
means
the
correct
the
corrections
we
did
after
last
week,
weekly
and
LTS
around
the
gdka
17
used
for
agents
and
windows,
packaging
and
stuff.
This
or
this
correction
looks
good.
So
we
can
proceed
on
the
next
steps.
A
A
B
A
B
Candidate
will
be
6
September.
A
Perfect
thanks
Mark
do
we
have
secreted
visery
and
none
so
good
and
next
major
event.
We
have
thumb
I,
don't
believe.
B
Devops
World
Tour
is
coming
to
I'll.
Be
writing
a
blog
post
today
about
it
actually
devops
World
Tour
New
York
Chicago,
Santa,
Clara,
Singapore
London.
A
I,
don't
have
other
major
events
good
for
you,
folks,
okay,
so
let's
start,
let's
start
with
the
task,
we're
able
to
finish
during
that
milestone.
We
had
a
request
from
a
plug-in
Miner
that
they
closed
since
so
that
request
was
about
using.
They
asked
the
infrastructure
to
install
visuals
vs
2020
to
SDK
to
benefit
from
a
more
recent
version
of
msbill.exe
on
the
agent
template
for
Windows
on
CI
Jenkins,
so
I
wasn't
sure
so
I
asked
for
help.
We
had
an
answer
from
the
deck.
A
B
B
There's
we
certainly
have
shipped
binaries
in
the
past
that
were
generated
exe
files
for
Windows
the
windows
service
wrapper,
for
instance,
has
done
that.
So
so
there's
not
there's
not
anything
that
says
we
can't
ship
them
so
long
as
they're
based
on
open
source
we're
allowed
to
ship
them.
We
don't
have
anything
forbidding
that,
but
in
this
case
I
don't
see
any
reason
why
we
would
create
the
added
liability
for
our
infrastructure
team
of
installing
Visual
Studio
2022
on
our
Windows
machines
that
Visual
Studio
2022
is
enormous
and
it's
it's
a
difficult
thing
to
track.
B
B
A
A
Yes,
I
think
so:
I'm
not
I
looked
quickly
at
their
plugin
and
I.
Don't
believe
that
what
they
need
to
build
is
already
the
plugin
itself.
It
seems
like
they
are
building
different
things
and
I
think
they
should
stick
with
app.
They
are
that
they
use
with
their
own
account
for
that
part
right,
because
our
infrastructure
is
made
for
building
plugins
that
are
just
shipped
to
Jenkins
users
and
I
had
the
feeling
that
we
were
be.
We
were
being
used
as
a
freemium
service,
which
is
not
the
case.
A
B
A
B
A
So
yeah,
so
we
can
proceed,
but
thanks
for
the
answer
and
thanks
Vedic
for
helping
with
that
area,
another
topic
which
is
closed
was
able
to
run
a
batch
command
on
Old
artifactory
users.
They
all
have
now
the
proper
configuration
so
I've
added
some
numbers.
So
we
know
the
order
of
magnitude
of
running
some
scripts
against
artifactory.
A
After
discussing
with
Stefan,
Danielle
and
Mark,
my
proposal
is
the
following:
no
more
action
required,
but
if
the
problem
comes
again,
meaning
if
artifactory
has
new
surprising
features
they
deployed
without
being
written
on
their
change
log,
then
we
will
have
to
add
the
monitoring
job
on
trusted
CI,
which
has
the
admin
token,
and
that
will
check
for
the
user
correct
setup.
It's
not
required
here,
because
it
was
a
one-time
problem.
But
if
the
problem
happened
again,
then
we
will
we
go
that
direction.
A
Yeah,
we
add
users
with
problem.
I
cannot
login
into
Jenkins
artifactory
anymore,
so
some
user
answer
the
question
and
fix
the
problem.
Some
other
didn't.
A
A
Last
week
we
had
an
issue
with
window
download
page,
not
working
during
the
releases
we
weren't
able
to
catch
that
the
windows
packaging
failed
for
the
weekly,
and
it
also
we
only
cooked
that
when
it
was
for
the
LTS
we
realized.
Oh,
both
packages
are
missing,
so
we
were
able
to
fix
the
problem
and
now
I
receive
alerts
again
on
page
or
Duty,
because
I
didn't
I
should
have
so
I.
Think
it's
because
I
changed
my
email
settings
and
we
should
be
able
to
track
that
problem
in
the
future
by
also
covering
more
exhaustively.
A
A
Thanks
everyone
involved
in
that
one
so
funny.
If
anyone
when
you
are
running
a
Windows
inbound
agent
container
image
be
careful
with
the
Java
options,
every
Dash
X
column,
something
options
must
be
wrapped
within
literal
double
quotes
on
Windows,
because
column
is
a
special
character
in
both
Powershell
and
bat.
So
that
mean
you
need
simple
codes
and
then
double
quotes.
Wrapping
each
of
these
options
on
your
command
line
set
or
you
can
use
double
quotes
and
use
backticks
backslash
in
order
to
to
have
double
code
being
literal
yeah.
A
A
A
A
I
think
that's
all
I
think
we
can
look
now
on
the
other
issue.
Let
me
look
at
the
nodes.
Yep
forecloses
not
planned,
so
now
work
in
progress,
I'm
taking
so
the
two
most
important
one
updates
Jenkins
CI
dot,
IO
the
update
Center.
So
we
have
a
mirror
system
working
on
Azure
that
updates
the
jenkins.io,
at
least
without
mirrors
working
now,
but
we
have
the
radioactor
system
ready
to
use
I
took
over
after
we
went
in
holidays,
so
it
was
able,
during
the
past
Milestone
to
set
up
an
internal
everything
system.
A
The
idea
is
that
we
can
add
cloudflare
bucket,
where
we
don't
pay
outbound
bandwidth
to
serve
HTTP
data
and
when
we
had
this
to
the
mirror,
we
need
to
synchronize.
So
we
are
now
synchronizing
by
scanning
a
local
internal
AirSync
while
we
have
the
remote
system.
So
the
next
step
now
for
me
is
to
find
the
proper
AirSync
URL,
but
the
service
is
reachable
and
the
mirror
bit
is
trying
to
scan
so
one
step
forward.
I
had
to
roll
back
some
of
the
changes
that
everyone
that
I
did.
A
We
don't
need
that
everything
server
to
be
publicly
available.
So
I've
wrote
back
this
one.
It's
only
internal
and
the
next
step
will
be
first
deploying
the
AirSync
and
Apache
copy
on
digitalocean
public
cluster.
So
we
should
be
able
to
have
two
mirror
one
cloudflare,
which
is
on
the
US
East
and
one
on
digital
Lucerne,
where
the
bank,
the
bandwidth
cost,
is
really
cheap.
A
We
need
two
mirror
to
get
started.
Then
next
step
will
be
once
I've
solved.
The
sync
boards
having
a
scanning
issue
and
integrating
the
synchronization
from
trusted
CI,
which
updates
plugin
index
every
15
minutes
to
copy
the
new
index
and
data
to
both
mirrors
not
only
to
the
current
real
life
system,
so
it
will
be
copied
on
three
locations
and
to
trigger
every
15
minute
after
successful
copy
a
mirror
scan
once
this
will
be
done,
so
that
should
be
in
one
or
two
day.
A
A
A
A
A
It
looks
like
there
is
a
specific
setting
in
that
builds
that
Force,
the
usage
of
one
or
the
other
I
don't
see
any
obvious
reason
why
it's
only
trying
to
use
Maven
Central,
but
there
is
a
maven
Dash
settings
that
XML
file
at
the
root
of
the
repository,
which
is
a
smell
for
me.
I'm
sure
there
is
a
specific
Maven
option
somewhere
that
make
it
be
behave
differently.
Why
did
it
start
to
fail,
though?
I
don't
know
that
one
is
weird.
A
We
have
a
current
virtual
repository
named
Albus
ldesk
Dash
3599
on
artifactory
that
we
can
use
to
to
simulate
right
now.
Mark
I
wasn't
able
to
review
to
to
fail
a
build
with
the
maven
HPI
plugin
from
scratch
on
my
machine,
but
I
use
the
settings
XML
for
seeing
every
connection
to
go
there,
at
least
for
the
repo
Jenkins
or
to
switch
back
to
to
Central.
A
The
new
virtual
repo
without
central
I
started
to
see
warnings
and
then
errors
that
say,
oh
I
see
the
dependency
I've
marked
the
dependency
has
been
downloaded
from
former
repo,
and
now
it's
on
the
new
one,
that's
the
same
checksum,
but
it's
weird
I
believe
the
error
we
had
during
the
burnout
could
be
related
to
dependency
already
present
somewhere
and
we
should
change
the
metadata.
A
A
Repo
didn't
add
any
time,
so
we
can
that
when
we
keep
it
for
the
next
the
next
week,
Stefan
your
turn
irm
64
for
application
on
the.
C
I
did
I
did
finish
today,
the
unit
test
and
so
for
right
now,
my
pull
request
is
working
as
before.
Building
all
the
images
with
multi-platform
compatibility
in
one
in
one
run,
Next
Step
will
be
probably
in
another
pull
request
to
deal
with
the
the
difference
between
Main
and
tag
right
now
we
use
main
to
build
then
to
create
a
tag
and
then
that
tag
trigger
a
new
pipeline
that
is
dealing
with
the
deploy
we.
C
C
I
think
it's
still
draft
I
need
to
change
that,
but
it
should
be
ready
for
review
now
nice
and
then
I
will
be
able
to
build
the
the
images
for
Wiki,
for
example,
which
was
the
one
I
tried
a
lot
and
and
to
migrate
those
services
on
the
RM
that
now,
that
should
be
easier.
A
Cool
next
step
is
Wiki,
we'll
have
to
wait
two
things
we
discovered
on
that
part.
The
first
one
is.
We
should
be
able
to
do
the
same
work,
but
it's
not
mandatory
for
the
rm64
part,
but
in
the
pipeline
Library
we
will
be
able
to
replace
the
tons
of
make
file
shell
and
groovy
code
groovy
code
being
the
worst
here,
because
not
reproducible
to
use
the
same
technique
as
the
one
we
had
for
the
official
rankings
agent
images.
A
A
So
same
idea
is
what
Stefan
said
same
convention.
We
should
set
up
using
a
Docker
compost,
yaml
file,
but
that's
not
mandatory
here,
and
that
will
cover
the
Windows
case,
and
second
thing
is
that
we
will
have
to
to
update
stock
inbound
agents
a
Windows
image.
We
will
have
to
migrate
them
to
a
bake
file,
a
proper
bake
file,
but
your
experimentation
shows
that
the
holder
paths
are
working,
so
you
don't
have
to
focus
on
this
one
right.
Now
that
mean
on
the
end,
less
pipeline
code,
which
is
a
good
thing.
A
A
So
the
idea
is
that,
in
order
to
reach
the
SSH
machine
of
trusted
CI
joint
against
IO,
you
will
have
to
only
use
the
VPN
meaning
no
need
for
public
IP
anymore
to
reach
that
machine.
That
will
be
an
internal
IP
and
only
only
that
people
with
the
proper
permissions
on
the
VPN
and
the
full
VPN
access
will
be
able
to
reach,
which
is
clearly
easier
than
having
a
low
list
of
public
IP
from
from
people.
A
Next,
one
on
the
lists,
LF
status
page
redirect,
may
be
cached
for
too
long
no
feedback
Mark
is
going
in
the
discussion
with
CDF
and
LF
on
that
topic
today.
So
we
will
have
maybe
some
news,
but
that's
not
important.
A
Next,
one
ath
will
commonly
begun
and
responsive,
so
that's
pipeline
code.
That
will
be
the
next
step
after
Stefan
new
finish
on
on
the
python
Library.
It's
a
few
Bible
line
of
code,
but
that
can
take
some
time
to
try.
But
yeah
is
that,
okay
to
keep
it
on
the
Milestone,
or
do
you
want
to
put
it
on
backlog?
A
A
Here,
command.
C
Prompts
I
started
this
one
locally
on
my
on
my
computer
with
backgrounds,
I
think
I
matched
where
to
to
delete.
I
am
struggling
with
the
escaping
between
all
all
the
tools,
Erb
and
interpolation
from
there
and
there
until
the
prawns
so
work
in
progress.
Okay,.
A
A
A
Third
CIA
Jenkins
the
virtual
machine
used
by
the
Jenkins
security
team.
That's
a
specific
and
hidden
and
restricted
controller
that
virtual
machine
has
been
migrated
to
a
new
VM
in
a
new
network
with
agents
that
are
more
powerful
but
cheaper,
yeah.
That's
true!
Thanks
AMD
for
the
Epic
CPUs
on
data
centers.
A
So
it
has
been
migrated
yesterday
and
for
the
upcoming
Milestone.
We
should
be
able
to
remove
the
former
resources.
The
plan
is
to
do
it
tomorrow.
Two
days
after
the
migration,
we
have
a
backup
of
the
genkinson
that
are
inside
our
production,
backup,
volt
and
the
team
is
able
to
access
using
the
new
VPN
by
the
way
I've
stopped
the
VPN,
the
former
VPN
machine
yesterday.
Just
to
be
sure,
we
don't
forget
about
anything
So.
The
plan
is
tomorrow
to
delete
the
former
VPN
former
search
CI
and
former
Legacy
network
resources
on
Azure.
C
C
B
A
Update
it
now
because
I'm
still
working
on
it,
I'm
gonna
commit
my
change,
so
you
will
be
able,
but
after
our
meeting
I.
A
A
good
thing
for
you
to
ask
when
you
wear
a
hole
here,
so
the
next
step
for
the
next
I
will
add
it
on
the
links,
but
we'll
have
the
old
VPN
machine
which
cost
a
lot
of
money,
because
it's
a
big
machine
because
it
requires
multiple
network
interfaces
and
we
will
have
removed
the
remnant
of
Legacy.
So
three
issues
that
should
be
completed
on
cleaner
part
for
the
upcoming
milestone,
I'm,
adding
them
right
now.
A
You
can
add
it
up
back
to
backlog
and
finally
Stefan.
We
also
have
high
availability
of
replicated
Services.
We
need
to
work
on
the
anti-affinity,
given
the
eye
load
of
tasks
that
you
have
on
your
backlog.
Is
that
okay,
if
I
take
it.
A
No
more
triage
for
this
one,
okay,
I
think!
That's
all
for
the
issues
we
are
at
work
in
progress,
so
I
started
working
on
this
one
on
my
side
by
reading
again
documentation
of
kubernetes,
because
I
I
tend
to
forgot
everything
each
time
so
I
need
documentation
to
remind
oh,
oh,
that
is
that
it's
working
like
this
okay
and
that's
all
for
me,
done
gotta
remove
tomorrow.
Do
you
have
other
topics
that
you
worked
on
and
I
could
have
forgotten.
C
A
No
resources
to
be,
let
me
add
it
to
be
removed
so
tomorrow
so
new
subject.
Let
me
add
a
note
new
subjects:
gdker
21
change
from
nightly
bills
to
weekly.
A
A
I
I
don't
speak
about
implementation
details
that
might
have
tons
of
if
and
then
else,
which
is
not
a
problem,
because
we
already
did
the
same
kind
of
thing
for
the
other
ddk,
but
in
terms
of
configuration
that
should
be
clearly
simplified.
Someone
setting
or
maintaining
one
of
our
controller
would
not
have
to
specify
a
nighty
date
version
for
each
of
the
gdka
architecture.
For
each
agent
to
link
agreed.
Don't
forget,
Packer
image.
A
A
We
had
an
issue
from
Alex:
remove
account
request
field
from
your
login
page
I'm
gonna.
Add
this
one
also
to
a
new
milestone.
When
someone
log
into
will
shoes,
then
Kim
Sayo.
There
is
an
unexpected
message,
as
you
can
see,
let's
say
to
request
an
account.
Please
contact
us,
a
message
with
links
and
these
elements
that
link
and
that
message
shouldn't
be
there
or
should
point
to
account
change.
A
Do
we
have
new
other
topics,
I
think
that
was
really
already
a
lot.
No
new
topics.