Git import times out - CGAL

Asked by Thomas Karl Pietrowski on 2020-02-09


I need some help doing a first import of a (large?) codebase.

So far it is always timing out when sending data to launchpad. I can imagine that only the first import takes that much time.
All the following imports shouldn't take that much time (I hope).

Can you help somehow? Or is there a way to download the files directly from Github (depth=1) and build packages without importing?

- Thomas

Question information

English Edit question
Launchpad itself Edit question
No assignee Edit question
Last query:
Last reply:
Colin Watson (cjwatson) said : #1

I'm not sure it's actually a timeout. The most recent import there is, but the ones before it are all:

2020-02-02 17:01:21 INFO To https://<redacted>
2020-02-02 17:01:21 INFO * [new branch] master -> master
2020-02-02 17:03:45 INFO remote: Traceback (most recent call last):
2020-02-02 17:03:45 INFO remote: File "hooks/pre-receive", line 148, in <module>
2020-02-02 17:03:45 INFO remote: rule_lines = check_ref_permissions(sock, rpc_key, ref_paths)
2020-02-02 17:03:45 INFO remote: File "hooks/pre-receive", line 130, in check_ref_permissions
2020-02-02 17:03:45 INFO remote: {'key': rpc_key, 'paths': ref_paths})
2020-02-02 17:03:45 INFO remote: File "hooks/pre-receive", line 120, in rpc_invoke
2020-02-02 17:03:45 INFO remote: res = json.loads(netstring_recv(sock))
2020-02-02 17:03:45 INFO remote: File "hooks/pre-receive", line 101, in netstring_recv
2020-02-02 17:03:45 INFO remote: c = sock.recv(1)
2020-02-02 17:03:45 INFO remote: socket.error: [Errno 104] Connection reset by peer
2020-02-02 17:03:45 INFO To https://<redacted>
2020-02-02 17:03:45 INFO ! [remote rejected] releases/CGAL-4.14-branch -> releases/CGAL-4.14-branch (pre-receive hook declined)
2020-02-02 17:03:45 INFO ! [remote rejected] releases/CGAL-5.0-branch -> releases/CGAL-5.0-branch (pre-receive hook declined)
[lots more of the same]

There's something a bit more complicated going on here that we'll need to investigate.

Thanks, Colin!

Especially for the quick reply on this! :)

Colin Watson (cjwatson) said : #3

I think this is probably a bug caused by the per-ref permissions check being relatively slow when lots of refs are being pushed all at once. Could you please use the "Create bug report" link on this question to convert it into a bug, and we'll see if there's something we can do to optimise this?

Sure that would be a pleasure to do!

Can you help with this problem?

Provide an answer of your own, or ask Thomas Karl Pietrowski for more information if necessary.

To post a message you must log in.