Re: Cannot clone development sources
Alquama Salim <[email protected]>
| Newsgroups | gmane.comp.gnu.octave.general |
|---|---|
| Message-ID | <CAENwqN_8zRSfM1MUQv+B74O9pOeUS9+r22pdg_p96TKEWj0CnQ@mail.gmail.com> |
It worked. thanks a lot On Fri, Dec 18, 2020 at 3:28 AM Bob Proulx <[email protected]> wrote: > Torsten Lilge wrote: > > Alquama Salim wrote: > > > i am a noobie interested in Octave. I wished to see the code > > > underlying but was unable to clone it. i used the command > > > hg clone https://www.octave.org/hg/octave > > > but it shows > > > abort: HTTP Error 502: Bad Gateway > > > please someone help me out cloning the whole thing in my system. > > Please try again. The server was getting overloaded. > > > > Thanks in advance > > > > This seems to be a problem on the server. The web-interface > > http://hg.savannah.gnu.org/hgweb/octave/ responds with the same error. > > Accessing the repo via ssh is possible. > > Since last night we have been having external influences overloading > the git server that is shared with hg on the same system. The > Savannah Hackers team knows about the problem and now have some > mitigation in effect to prevent the worst of it. So hopefully it is > no longer negatively impacting everyone now. > > If anyone had this problem please try the action again. It should be > okay now. Mostly. Still perhaps getting transient overloads. > > In detail the problem is that an external host is starting a download > of the emacs repository and the immediately either disconnecting or > being disconnected. That is leaving a git-pack-objects process > running on the rather large repository. It doesn't know that the > parent process has disconnected and exited until it eventually writes > to the output and finds the file descriptor closed. At which time it > would exit itself. But it is a large repository and that takes some > many minutes to happen. In the meantime each of those processes is > hogging about 800 MB of ram with an active working set as it is > spinning through the repository files. And then this is repeated > again and again dozens of times leaving dozens of large processes > running, hogging all available memory, and eventually driving the > system into memory thrash. It's an I/O bandwidth overload of the file > system storage backlogging the running processes. > > Bob >