Re: Premature timeout on git.postgresql.org?

On Sun, Feb 20, 2022 at 3:52 PM Bruce Momjian <bruce@momjian.us> wrote:
>
> On Sat, Feb 19, 2022 at 01:36:58PM -0500, Tom Lane wrote:
> > Trying to check out from the git mirror on a machine with a
> > slow internet connection, I repeatably get
> >
> > $ time git clone https://git.postgresql.org/git/postgresql.git pgsql
> > Cloning into 'pgsql'...
> > remote: Enumerating objects: 14523, done.
> > remote: Counting objects: 100% (14523/14523), done.
> > remote: Compressing objects: 100% (9482/9482), done.
> > error: RPC failed; curl 92 HTTP/2 stream 3 was not closed cleanly before end of the underlying stream
> > fetch-pack: unexpected disconnect while reading sideband packet
> > fatal: early EOF
> > fatal: fetch-pack: invalid index-pack output
> >       364.36 real         0.02 user         0.08 sys
> >
> > It gets almost but not quite to the end of the data-transfer stage
> > before failing.  I think I've seen this symptom before, and IIRC
> > we traced it to a poorly-chosen timeout somewhere in the PG web
> > infrastructure.  I can't find a record of that in the archives,
> > though; the closest thing is [1] which is a somewhat different
> > symptom.

Hmm. That's interesting. Do you see it repeatable now? Or is it
possible it was a temporary network flakiness?

How long (ish) does it take from the start of the clone until it fails?


> I did a git clone from https://git.postgresql.org/git/postgresql.git on
> Friday and just now and was only seeing 1.15 MiB/s, which seemed low on
> my 100Mb connection.

I get about 50MB/500Mbit per second when I clone from it. The git
server is physically located in Amsterdam, maybe you have a limited
transit performance there (or it has to you).

Can you, as well, try again now and see if you see the same level of
performance now?

-- 
 Magnus Hagander
 Me: https://www.hagander.net/
 Work: https://www.redpill-linpro.com/