https://flyte.org
Join Slack
After upgrading flyte to 1.9.1, when I click the relaunch button for a workflow execution in the UI,...
# flyte-support
n
After upgrading flyte to 1.9.1, when I click the relaunch button for a workflow execution in the UI, it'll pop up the form and then instantly go white screen.
Just a bunch of
Copy code
react-dom.production.min.js:125 TypeError: Cannot destructure property 'type' of 'undefined' as it is undefined.
    at m (main.7d9178e441dc822c8094.js:1:334745)
    at t.UnionInput (main.7d9178e441dc822c8094.js:1:267049)
    at we (react-dom.production.min.js:84:293)
    at zj (react-dom.production.min.js:226:496)
    at Th (react-dom.production.min.js:152:223)
    at tj (react-dom.production.min.js:152:152)
    at Te (react-dom.production.min.js:146:151)
    at react-dom.production.min.js:61:68
    at unstable_runWithPriority (react.production.min.js:25:260)
    at Da (react-dom.production.min.js:60:280)
all the network requests look ok. they are all status 200
t
@late-eye-50215?
have you seen this?
l
No we haven't seen that; could you tell me more about the inputs to this launch form? cc: @some-raincoat-58190
👀 1
n
it has optional[list[str]]
Copy code
x: str,
    y: str,
    z: Optional[List[str]],
    a: Optional[List[str]],
    c: int = 20,
    b: bool = False,
    d: str = "g",
    e: str = "h",
    f: bool = False,
    g: bool = False,
l
Okay that is helpful - thanks! We'll see if we can figure out what's going on 👍
🙏 1
n
the specific execution has one of the list args populated and the other empty
were you able to reproduce it? or is it something just off about my config
bump on this since it's preventing us from upgrading to 1.9.x
t
were the optional args filled in or not?
n
one was filled in and one wasn't
t
and if you try in a different browser you get the same behavior right?
n
uh, I only tried with chrome.
let me reupgrade my cluster and try some more minimal examples like you said
t
thanks
i’m testing with
Copy code
import typing

from flytekit import task, workflow


@task
def repro_1(a: typing.Optional[typing.List[str]], b: typing.Optional[typing.List[str]]):
    print(a)
    print(b)


@workflow
def run_repro_1(a: typing.Optional[typing.List[str]], b: typing.Optional[typing.List[str]]):
    repro_1(a=a, b=b)
not able to repro, i can re-launch an execution without issue
might just need a hard refresh?
n
haha maybe. i coudln't even load the launch form for the workflow
hmm, i tried chrome incognito, closing all chrome and then opening, and also safari. couldn't open the launch form
Copy code
@task
def my_mappable_task(a: typing.Optional[typing.List[int]]):
    print(a)

@workflow
def my_wf(x: typing.Optional[typing.List[int]]):
    my_mappable_task(a=x)
t
and others are also seeing this problem?
n
yes, my teammate just white screen'ed also
t
and you’re running single-binary or split components?
n
split components
t
can you give us the image versions of flyteconsole and admin?
n
Copy code
<http://cr.flyte.org/flyteorg/flyteconsole-release:v1.9.1|cr.flyte.org/flyteorg/flyteconsole-release:v1.9.1>
<http://cr.flyte.org/flyteorg/flyteadmin-release:v1.9.1|cr.flyte.org/flyteorg/flyteadmin-release:v1.9.1>
says my helm chart
any specific pod logs I should try to look for?
t
pod logs maybe not, but more console logs from chrome debugger maybe.
n
Copy code
react-dom.production.min.js:125 Error: empty string
    at Function.d [as fromString] (vendor.d3268180b8c0d189c23e.js:2:21291)
    at Object.toLiteral (main.7d9178e441dc822c8094.js:1:299325)
    at Object.typeDefinitionToDefaultValue (main.7d9178e441dc822c8094.js:1:294199)
    at main.7d9178e441dc822c8094.js:1:266673
    at Array.reduce (<anonymous>)
    at main.7d9178e441dc822c8094.js:1:266522
    at t.UnionInput (main.7d9178e441dc822c8094.js:1:266721)
    at we (react-dom.production.min.js:84:293)
    at zj (react-dom.production.min.js:226:496)
    at Th (react-dom.production.min.js:152:223)
Me @ react-dom.production.min.js:125
main.7d9178e441dc822c8094.js:1 Error: empty string
    at Function.d [as fromString] (vendor.d3268180b8c0d189c23e.js:2:21291)
    at Object.toLiteral (main.7d9178e441dc822c8094.js:1:299325)
    at Object.typeDefinitionToDefaultValue (main.7d9178e441dc822c8094.js:1:294199)
    at main.7d9178e441dc822c8094.js:1:266673
    at Array.reduce (<anonymous>)
    at main.7d9178e441dc822c8094.js:1:266522
    at t.UnionInput (main.7d9178e441dc822c8094.js:1:266721)
    at we (react-dom.production.min.js:84:293)
    at zj (react-dom.production.min.js:226:496)
    at Th (react-dom.production.min.js:152:223)
t
@late-eye-50215 could you try to deduce on your end? and @numerous-hamburger-7178 could you submit an issue for this is you haven’t already.
n
just made an issue
t
fyi we are also running 1.9.1 (but using the single binary chart)
n
oh nice. this was fixed!
i should bump to v1.9.6? is that a real release? https://github.com/flyteorg/flyte/releases. I don't see it in here tho
t
no stick with 1.9.1 for now.
we will release 1.10 soon ish
n
I never upgraded to 1.9.1 bc of this, so I can just wait for 1.10 then.
t
oh did we never submit a patch for this?
😞
1.10 it is then.
n
https://github.com/flyteorg/flyteconsole/releases/tag/v1.9.6 there's this but I don't think this tag shows up in docker or helm
t
the image should be there though.
so if you want to just upgrade console to 1.9.6 (if you’re running flyte-core) that’s doable
if not yeah you will have to wait.
n
Copy code
➜  x git:(master) ✗ docker pull <http://cr.flyte.org/flyteorg/flyteconsole-release:v1.9.6|cr.flyte.org/flyteorg/flyteconsole-release:v1.9.6>                                                                                           
Error response from daemon: manifest unknown
➜  x git:(master) ✗ docker pull <http://cr.flyte.org/flyteorg/flyteconsole-release:v1.9.1|cr.flyte.org/flyteorg/flyteconsole-release:v1.9.1>
v1.9.1: Pulling from flyteorg/flyteconsole-release
t
Copy code
docker pull <http://cr.flyte.org/flyteorg/flyteconsole:v1.9.6|cr.flyte.org/flyteorg/flyteconsole:v1.9.6>
n
oo thx
Open in Slack
PreviousNext
Powered by

Flyte

Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Powered by