Your Spark job requested 8 vcores. However, the workspace has a 0 core limit. Try reducing the numbers of vcores requested or increasing your vcore quota.

Jon-Paul Boyd 26

Hi,

In Azure Synapse, having an issue executing a Spark notebook. I have checked, using monitor, that no other jobs or notebooks are running.

What I find curious, is that apparently the Synapse workspace should be allocated 50 vcores as default (?), yet the error message from the notebook indicates the workspace has a 0 core limit. How can this possibly be (0 core limit)?

Many thanks for any support as right now I'm unable to progress.

DwaneEDDC 31 Reputation points

2022-04-02T16:38:39.997+00:00

I am running into the same issue. Started on April 1.
Jon-Paul Boyd 26 Reputation points

2022-04-02T22:34:54.933+00:00

Hey. Unresolved here. Raised a support ticket to increase vcore quota (to 50, which is the default anyway, just in case there was some "internal" issue).

I still get the same error, even when I create a new spark pool, and I dont have any other active spark notebooks/jobs/sessions that I can see.

Please do let me know if you are able to resolve, working to resolve same here and will feedback, really frustrating.
Alwyn du Plessis 6 Reputation points

2022-04-03T07:56:01.02+00:00

Getting the exact same problem from 1 April, and also can't seem to solve it
Jon-Paul Boyd 26 Reputation points

2022-04-03T13:29:52.337+00:00

Out of interest, in what region is your Synapse workspace?
DwaneEDDC 31 Reputation points

2022-04-03T15:06:54.597+00:00

West US 2. We tried to raise the quota as well, which did not resolve the issue.
DwaneEDDC 31 Reputation points

2022-04-03T15:26:38.967+00:00

I also tried South Central US and got the same failure.
DwaneEDDC 31 Reputation points

2022-04-03T15:27:11.727+00:00

I think this is also impacting our dataflows, which run spark under the hood. This is the error I get on the Dataflow:

> Internal Server Error:Failed to submit job on job cluster. Integration Runtime <runtime id> ActivityId: <activity id>.
Jon-Paul Boyd 26 Reputation points

2022-04-03T17:14:50.383+00:00

You created a new Synapse Workspace in South Central US and tried to execute another Spark notebook, which failed with same "0 core limit error"?
Jon-Paul Boyd 26 Reputation points

2022-04-03T17:15:27.637+00:00

Would seem the same issue as no compute to execute them. Have you raised a support request?
Jon-Paul Boyd 26 Reputation points

2022-04-03T17:16:23.463+00:00

The quote was raised through a service request? Really unfortunate it didn't resolve, running out of ideas on hot to get it fixed and progress my project
DwaneEDDC 31 Reputation points

2022-04-03T17:21:05.793+00:00

Correct
DwaneEDDC 31 Reputation points

2022-04-03T17:26:55.74+00:00

Yes. We've raised a support request. I'll keep you posted.
Jon-Paul Boyd 26 Reputation points

2022-04-03T17:30:20.48+00:00

Ouch, but thanks for feeding this back, as I was reluctant to set up a new Synapse workspace but was going to give it a go to see if problem persisted
Alwyn du Plessis 6 Reputation points

2022-04-04T06:46:53.36+00:00

It is West Europe
Jon-Paul Boyd 26 Reputation points

2022-04-04T07:04:01.693+00:00

Same here, West Europe, although I note @DwaneEDDC had deployed to both West US 2 and South Central US, and had same issue with both deployments
Jon-Paul Boyd 26 Reputation points

2022-04-04T07:15:37.91+00:00

Hi, curious if under your Subscription > Service Health, you have a Data Factory service issue (mitigated) starting 30/03/2022 (as I have)? Wondering if it's related.
Jon-Paul Boyd 26 Reputation points

2022-04-04T07:15:51.793+00:00

Hi, curious if under your Subscription > Service Health, you have a Data Factory service issue (mitigated) starting 30/03/2022 (as I have)? Wondering if it's related.
Merger All 21 Reputation points

2022-04-04T12:52:51.627+00:00

same here, simple notebook as:

%%pyspark
df = spark.read.load('abfss://wwi-02@asadatalakex69cuek.dfs.core.windows.net/sale-poc/sale-20170501.csv', format='csv'
, header=True
)
display(df.limit(10))

run to error of:

InvalidHttpRequestToLivy: Your Spark job requested 12 vcores. However, the workspace has a 0 core limit. Try reducing the numbers of vcores requested or increasing your vcore quota. HTTP status code: 400.

It was OK few days before.
Jon-Paul Boyd 26 Reputation points

2022-04-04T13:00:54.64+00:00

Thanks @Merger All for providing your feedback as well, and can see you have exactly the same issue. a simple print("Hello World") in the spark notebook fails. I have an open service request, will link this to feedback multiple cases with same issue. Have you opened a service request?
Alwyn du Plessis 6 Reputation points

2022-04-04T13:23:45.73+00:00

The complexity of the notebook code does not seem to have an effect. It could literally be a 1+1 calculation or an ML model, neither make a difference as the issue is with the quota.
Jon-Paul Boyd 26 Reputation points

2022-04-04T13:24:59.403+00:00

^^^ This
John M 6 Reputation points

2022-04-04T15:54:31.943+00:00

Bumping for visibility.

Same issue as listed here, as of April 1st 2022 dataflows and PySpark notebooks fail to run or enter debug mode citing 0 available vcores. Subscriprion page also shows that all vCPU quotas are 0/0. This change in subscription was never requested.
Lucas Gomes 6 Reputation points

2022-04-04T18:03:17.097+00:00

Same problem here.

It also started on April 1st 2022. Dataflows and PySpark notebooks fail to run or enter debug mode citing 0 available vcores.

Region of Synapse: East US.
Jon-Paul Boyd 26 Reputation points

2022-04-04T19:52:51.453+00:00

It would appear my "workspace has a 0 core limit" issue is not unique, with several folks here reporting the same originating around 1st April.

It would really help if everyone can raise a service request to report through the official channel, perhaps include this link, and please feedback any resolution you get.

Here's hoping we can get our Synapse workspaces with Spark pools back up and running soon, after 4 days downtime.

Thank-you
Derek de Oliveira 11 Reputation points

2022-04-04T22:53:00.777+00:00

MSFT starting to acknowledge the issue here:
https://app.azure.com/h/2NS0-P80/985c4a

Although the issues lists that it is "mitigated", we're still unable to run spark notebooks atm - also waiting on the capacity team to provide capacity so we can get our pipelines going
Merger All 21 Reputation points

2022-04-05T00:54:50.943+00:00

They said quota increased, but issue still there
Jon-Paul Boyd 26 Reputation points

2022-04-05T07:51:23.14+00:00

@John M Your notebooks running OK now. It seems the issue has been resolved for a few of us, currently root cause behind vcore limit being 0 is unknown.
Jon-Paul Boyd 26 Reputation points

2022-04-05T07:51:42.983+00:00

You might want to try again, seems resolved for a few of us.
Struan Cooper 1 Reputation point

2022-11-07T14:42:39.387+00:00

I had the same problem, I recreated in another region (US East) instead of West Europe.
Mir Ali 5 Reputation points

2023-08-21T14:04:36.4433333+00:00

August 21, 2023

Having the same issue, I can't even request a quota raise since I'm on a student account.

4 answers

Merger All 21 Reputation points

2022-04-05T01:55:27.023+00:00

It's a Azure backend issue.
I request to increase vCore quota per workspace, they feedback in Email of request sumitted successful.
I tried with no improve during 10 hours afterwards.
And just now, the issue gone. the notebook can run on my sparkpool successfully.
Please sign in to rate this answer.

3 people found this answer helpful.
Derek de Oliveira 11 Reputation points

2022-04-05T02:33:21.28+00:00

Our instance only started to work after they handled our capacity request - this solved Spark pools; however, Data Flows were still not working. We did eventually get this to work by turning on Data Flow Debug mode and then literally turning it back off again. When in doubt....reboot!

Alwyn du Plessis 6 Reputation points

2022-04-05T05:18:06.19+00:00

Thanks @Merger All , we have done the same. Whether it was do to this or whether they have otherwise solved the issue, things are working again on our side as the vcore issue is no longer present

Jon-Paul Boyd 26 Reputation points

2022-04-05T07:49:18.48+00:00

Thanks for the feedback. I've just tried running a Spark notebook again, which works.

This is without any update on support request that workspace vcore limit has been update.

If it's OK I'd like to leave this Q open as I'd like to follow up with MS on root cause and share here.

Jon-Paul Boyd 26 Reputation points

2022-04-05T07:50:10.713+00:00

Same here, all appears OK now without update to "reset" vcore limit. Will follow up with MS on root cause and share here

Merger All 21 Reputation points

2022-04-05T11:32:14.137+00:00

It's up to you. Just do not expect much deep details.
Sign in to comment
Gerben Dam 6 Reputation points

2022-10-31T10:36:03.27+00:00

We have the same issue in all new synapse workspaces on the spark pools, West Europe region. I can not run a single notebook. Just raised a support request. Very annoying!
Please sign in to rate this answer.

1 person found this answer helpful.

0 comments No comments
Sign in to comment
Ranjith Abraham 1 Reputation point

2022-08-17T10:26:35.69+00:00

Although I contacted Microsoft Support and got 64 vcores to my workspace, even after 12 hours I cannot run a basic Spark pool cell and the same error comes "InvalidHttpRequestToLivy: Your Spark job requested 28 vcores. However, the workspace has a 0 core limit. Try reducing the numbers of vcores requested or increasing your vcore quota. HTTP status code: 400."
Please sign in to rate this answer.

0 comments No comments
Sign in to comment
Carl Brown 0 Reputation points

2023-02-28T17:07:44.7333333+00:00

February 28, 2023. And still the same issues. "Looks like this error is most likely a platform issue." - Microsoft Support.

They told me to issue a ticket at msft.it/60145i6Bw . But when I do, it says I have to pay $29.99/month to join their support access. I've read all of their online material. These notebooks work sometimes, and sometimes they don't. None are scheduled are used in pipelines. We only have 3 notebooks. I'm not putting that stuff in production if it's not dependable. Makes absolutely no sense to me.

If Microsoft wants us to use their Synapse Workspace in team settings, where there's a mix of data engineers and predictive scientists, something has to get better here.
Please sign in to rate this answer.
AnnuKumari-MSFT 31,061 Reputation points Microsoft Employee

2023-03-02T09:58:38.12+00:00

Hi Carl Brown ,

Since this is an old thread, requesting you to create a new post so that we can actively engage on that. Thankyou

Nazmul Bhuiyan 0 Reputation points

2023-03-20T04:42:03.11+00:00

Is this issue resolved?

Luke Herbst 0 Reputation points

2023-07-11T02:58:54.3666667+00:00

doesn't seem resolved. Couldn't run print('hello') got the same error
Sign in to comment