svn checkout fails on large projects with large files

[Bjørn Konestabo]

Trying to check out a large project with several large files usually fails with the error message.

svn: REPORT of '/svn/!svn/vcc/default': Could not read response body: Secure connection truncated

Web searches seem to indicate that this is a server side issue.

Are there things to do on my end to prevent this?

[Chris Clarke · Accepted solution]

Atlassian have informed us (via another ticket) that due to a 15 minute hard timeout on the ngnix web server that sits in front of Fisheye SVN in OnDemand, checkouts with large files may never complete. This timeout is enforced even if the connection is submitting data (it's not an idle timeout)

OnDemand is different to JIRA Studio hosted in this respect, which didn't utilise Ngnix, so checkouts under JIRA Studio for the same repos wouldn't be affected in this way. I'm amazed no-one from Atlassian has been along and answered this one for you.

Chris

[riddepk]

Hy i've got the same problem , i used TortoiseSVN and RapidSVN  still:

Error while performing action:REPORT of '/..../.../!svn/me' could not read response body: An existing connection was forcibly closed by the remote host.

[Stephen Wells]

We are experiencing the problem right now trying to download a large project. The download ran for about 40 minutes and then stopped with the following.

svn: E175002: SSL peer shut down incorrectly

svn: E175002: REPORT request failed on '/svn/!svn/vcc/default'

is there any work around for this? I am working on a migration plan but don't expect to begin implementation until this summer.

[Gareth Hughes]

I'm seeing this right now on a large checkout from our repo on a jira studio instance.

I'm going to migrate our source to the server soon and am configuring Hudson to test builds but checkouts are failing with

org.tmatesoft.svn.core.

SVNException: svn: SSL peer shut down incorrectly

svn: REPORT request failed on '/svn/!svn/vcc/default' and an svn checkout on the command line of my own machine failed with Secure connection truncated

[Bjørn Konestabo]

After a brief downtime, atlassians servers are back up, and the problems seems to be gone.