Hi,
I am trying to do a One-time SVN repository import with SubGit ,but the process fails after sometime.
command: ./subgit configure --layout auto --svn-url svn+ssh://u0146162@svn.dev-scholarone.com/svn/mc/projects/v3 /home/jmishra/Downloads/test_repo_new
This is a huge repository.
Please find the log attached.
Please advise.
Regards,
Jai
Hello Jai
the problem appears to be on the SVN server side: initially, SubGit was able to connect to the SVN server, read its history and started analyzing its history to generate the mapping configuration, but after about 7 minutes the connection was reset:
[2019-05-28 11:26:38.624][subgit-configure][18] com.trilead.ssh2.transport.TransportManager : Receive thread: error in receiveLoop: Connection reset
I’d recommend investigating ssh on the SVN server side as it looks like the server closes the connection after some time and that causes the problem.
Thanks Ildar.
I’ll take a look at it and get back if need any help.
Regards,
jai
@ildarhm Hi,
we have fixed this timeout error but now the process gets stuck at 90%(since 3 hrs) please see below
Peg location detected: r65084 trunk
Fetching SVN history… Done.
Growing trees… Done.
Project origin detected: r1 trunk
Building branches layouts… Done.
Combing beards…
df beard 58585 of 65083 combed | 90% [||||||||| ]
Is it because of its huge size? how can we ensure that the subgit is still working on it and its not dead.
is there any way to limit the number of tags we fetch during ‘configure’?
My command is: ./subgit configure --layout auto --svn-url svn+ssh://u0146162@svn.dev-scholarone.com/svn/mc/projects/v3 /home/jmishra/Downloads/test_repo_new
Regards,
Jai
Hi Jai
the repository size doesn’t matter that much at the configuration stage as its history size and changes in particular revisions. There might be many changes in revision 58586 and that may cause the process stumbling. The command output is basically the main place to track its progress; you can also check the configuration log, SubGit may write to it during the layout creating and also the process cpu and ram consumption, but I’m afraid, there’s no more progress tracks but that.
At the time of configuration there is no ways to exclude somethings from the process: in fact, the ‘configure’ command with ‘layout auto’ is designed to scan SVN repository history and create a suitable mapping configuration for it, so it’s natural for this process not to exclude anything. Tags or branches may be excluded from the subsequent import, however.
A few hours don’t seem unreasonable for the huge repositories with long and robust history, so it would worth to wait a bit to see if it progresses.