-
Notifications
You must be signed in to change notification settings - Fork 296
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Max retries exceeded caused by SSLError(SSLZeroReturnError #1356
Comments
This is not an auth issue, but possibly an SSL issue. There is not enough data in this stack trace to determine what the issue might be. It seems that the connection is timing out when trying to reach these endpoints. Did you notice this issue on multiple endpoints? Are the errors isolated to Google? Are you able to establish a TLS session to the URLs that are having issues with something like OpenSSL's tls client? |
Hi @lautaropacellaDR , I wanted to mention that we ran into what seems like a highly related issue. We think it stems from how the PubSub PublisherClient handles batched message publishing and authentication, and went to some trouble to demonstrate how to reproduce it here: googleapis/python-pubsub#977 Happily, we also found a couple of workarounds that we document in the reproduction script. I hope that they can be useful to you. Warmly, |
Hi @acolavin! thanks for that, I did do a quick check but I think that would solve issues specifically when using pubsub. Hello @clundin25, thanks for taking the time to check on this.
I'm very sorry @clundin25, but I'm not sure how to do this |
I have a similar issue on the different context using dbt-osmosis with BigQuery. I am doubting the issue is caused by conflicts between multpile threads in applications and the Google Cloud client. I am keep looking into it. |
I am getting the same error is there any news on this ? |
Can you share your stack trace @Hsgngr ? |
Unfortunately only stack trace I have is this because its coming from the Cloud Run. @clundin25 although my problem is not about google-auth-library The problem happens on my cloud Run as well. Probably this issue is related to that.
requests==2.31.0 |
Hi @clundin25 - I am having a similar SSL Error that I have a full stack trace for (just anonymized my project and dataset name, can provide privately if needed):
When attempting to debug locally, this error trace is then shown:
|
Hi Strimboli, is this also Cloud Run that tries to connect with BigQuery or local ? |
Hi @Hsgngr - it is running in a kubernetes job, not local. |
Any solutions to this? I've been on it for weeks, any help will be greatly appreciated pls! I am also getting the same error on Cloud Run: This happens when I call the Bigquery client for the second time on Cloud Run or even when I call Google Sheets for the second time and subsequently. It works perfectly when I run my app locally. Traceback from google sheet call:
|
I have the same issue when I call gspread on Cloud Run. It runs fine on local. |
Having this same issue now. Running the auth and storage Python libraries and getting the same error arround Max retries exceeded and SSL error. As other commentators have said it works fine on local and the issue is intermittent, i.e. works sometimes on Cloud Run and not others, with no indication or code changes between revisions. |
So I have turned on no-cpu-throttling on my cloud run and the problem went away. I think it is due to the fact that it needs stable network environment to make external network calls. However turning on no-cpu-throttling incurred more cost. I am still figuring out a way to mitigate this without spending too much. |
I'm experiencing the same issue with pods inside our GKE cluster writing to bigquery. The strange part is it happens with some pods and not all. |
The only way I was able to solve this was to create a cloud function that specifically writes to Google sheet and big query, hence my app on cloud run calls cloud function using a httpx client. |
We've made some code changes (so far, it's working; I'll know more next week). The issue was we were writing too much data to a big query. One of our JSON payloads contained megabytes of data. |
Same issue here. Pretty much identical setup and trace as the OP. |
As frustrating as this is, my issue just seems to have resolved itself with no changes on our end. It hasn't come back now for a couple of weeks so I believe it's stable, although it would be great to know what is causing this. |
LOL same error still occurs to me. |
I am having the same error while trying to connect to ClickHouse. |
I've encountered the The issue persists despite using the following library versions:
This seems to be a similar issue to what others have experienced, as discussed in this thread. Has anyone found a reliable solution or workaround for Cloud Run environments? |
Also getting this error - for both BigQuery and Cloud Storage calls from Cloud Run. |
I have the same error, I use this version: |
+1 |
I'm going to close this, as I believe this issue is unrelated to the code maintained in this repository. From what I can tell this is a networking issue for large payloads. If anyone can reproduce this consistently, please share and we can try to track down a solution. Otherwise, the best bet is to open a support case (https://cloud.google.com/support/) and collaborate with the support team to collect a trace of what happened on the server side. The client side logs are opaque and I believe there is little we can do without an investigation on the server end. |
This is because by default cloudrun only allocate CPU during requests, it means that in background task there is no CPU allocated, and This can be change by adding ref: https://cloud.google.com/run/docs/configuring/cpu-allocation |
Hi, Im getting this error since mid June when before it was working well.
Max retries exceeded with url: /entities/x?fields=y (Caused by SSLError(SSLZeroReturnError(6, 'TLS/SSL connection has been closed (EOF) (_ssl.c:1129)'))
I'm getting this error for all the external sources I try to request from this slack bot. Being storage, bigquery, cloud runs and cloud functions. I want to emphasise that this had only started happening in mid June, and before it was working well.
When I run the same repo locally, it works perfectly. This is why I think it could be something related to auth?
Sorry if not, I'm pretty new to this.
Any help would be much appreciated!
Environment details
Running on Cloud Run. Fast API and Slack SDK packages.
Python Version: 3.11.4
Requirements.txt:
Full error report:
Thanks for your time!
The text was updated successfully, but these errors were encountered: