On Mon, Jun 20, 2022 at 6:48 AM Tom Lane <t...@sss.pgh.pa.us> wrote: > Adrian Klaver <adrian.kla...@aklaver.com> writes: > > On 6/19/22 20:33, Wen Yi wrote: > >> |[beginnerc@fedora Research]$ git clone > >> ||https://git.postgresql.org/git/postgresql.git > >> <https://git.postgresql.org/git/postgresql.git>| > >> |Cloning into 'postgresql'...| > >> |remote: Enumerating objects: 30747, done.| > >> |remote: Counting objects: 100% (30747/30747), done.| > >> |remote: Compressing objects: 100% (13431/13431), done.| > >> |error: RPC failed; curl 92 HTTP/2 stream 3 was not closed cleanly > >> before end of the underlying stream| > > > I have to believe it is related to bleeding edge OS Fedora 36 in > > combination with latest Git 2.36.1. > > No, I think it's more about this: > > >> I check the download speed,that is about 220kb/s. > > I've seen this failure multiple times on very slow machines. > I think there's some sort of connection timeout somewhere in > the git.postgresql.org infrastructure, causing a "git clone" > that takes more than a couple of minutes to fail. I've > complained about it before, but we've not isolated the cause. >
The last time we debugged it pretty much the only conclusion we managed to come to I think was that it is *not* in the git.postgresql.org infrastructure. I tried it from *many* different locations and it worked fine from all of them, even when artificially slowing it down to something much slower. So yes, there is *something*, but it's not with in the pg.org infrastructure. One thing we got to work that time, I think, was to run: git config --global http.version HTTP/1.1 -- Magnus Hagander Me: https://www.hagander.net/ <http://www.hagander.net/> Work: https://www.redpill-linpro.com/ <http://www.redpill-linpro.com/>