-
-
Notifications
You must be signed in to change notification settings - Fork 63
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
HTTP Error: 502, "Bad gateway", "/graphql" #83
Comments
Is that fairly large repo publicly accessible? |
Unfortunately not. Hence the offer to provide more debug info if you can tell me what you need. |
I have the same issue on a public repo: https://github.com/NixOS/nixpkgs. The log message with
|
And after reading the other open issues related to this problem and trying to fetch notifications, I get a different error:
The traceback is very large and includes my GitHub notifications all of the repos I follow but here it is anyway, somehow weirdly formatted though... https://clbin.com/bPu84 That's from when I had about 50 notifications waiting for me, a mix of issues and pull-requests. When I acknowledged them and tried again with 0 notifications, it went OK, at least something... Edit: The last note is wrong - even with 0 pending notifications, it still throws an error. It seemed to work to me because it took a very long time for the debugger to open... |
I tried out the same with the newly added "only pull topics since date", but the error is the same even when pulling topics since today. That's assuming I'm doing it correctly, writing '2019-01-21' at the prompt for "since".
|
My best idea on how to deal with this is to just automatically to try again a few times. The problem is on the Github's side. This is what the error means.
Github provides some additional information in the response:
My guess is that it is a timeout due to temporary overload. I have seen this error myself before, but when I did then simply trying again always worked. And then:
Could you please collect a few such error ids and then send an email to Github asking them what has gone wrong for them. But I fear that will not lead anywhere. Expected response "we give requests n time units to succeed and if it takes longer then we send a 502". There actually was one situation where we could reproduce this issue more consistently: fetching the topics associated with the notifications that were fetched earlier. We solved that by fetching at most 100 topics at a time. So if you run into this problem, then determine whether it happens when fetching notifications or when fetching other things. If it is the former then fetching fewer topics-for-notifications at a time might help. |
Can reproduce on |
I can also get this consistently on the xonsh repo. Fetching notifications works but |
The workaround seems to be: try again ten seconds later. For our purpose this isn't actually a large repository. It's smaller than the magit repository for example, which never caused my any troubles. But I wasn't able to fetch the data for this repository on the first try either. It failed pretty early on and my guess is that it took to long to wake something up, resulting in an internal timeout on within github. On the second try that thing was already awake. |
I see; thanks very much for the explanation! Upon trying that repo again, you were right, it did work. |
Just to add to the conversation I'm also experiencing this with Nixpkgs. |
This doesn't happen very often and when it does, then it isn't our fault. While we could investigate whether it might possibly be an good idea to simply retry when this error occurs I don't think it would be worth the effort. |
Naive question, but is there any way to perform |
That's huuuuuuge. No, an "incremental initial pull" is not supported. If you were to succeed downloading everything, then you would just have another issue: forge would be way to slow in every day use. Of course theoretically it would be possible to address all that, but I am not enough people, so forge will likely never be improved to support such humongous repositories. Sorry. |
No worries. Thanks for the response anyway and thanks for all the work you've already done on this! |
This happened to me with https://github.com/zephyrproject-rtos/zephyr. And I did retried 10 times and voila it finally worked! |
In case it is helpful to someone, I got this error when fetching all topics. I had set up Forge a few months in the past but didn't really use it so I was picking it up again. When I tried to fetch one topic, it told me it couldn't until I added the repo. I did so and as part of that it fetched all items without any problems. |
I'm still getting an error similar to #20 on a fairly large repo (running
20190107.1955
). I can pull notifications, and that works okay, but trying to doF-y
results in:Unlike above however, even if I restart the magit buffer there are no pull requests or issues listed.
Let me know if there's more debug info that could help.
Thanks!
The text was updated successfully, but these errors were encountered: