[jira] [Commented] (SVN-3264) Large directory checkouts fail, connection forcibly closed

"Daniel Sahlberg (Jira)" <[email protected]>
Newsgroups gmane.comp.version-control.subversion.issues
Message-ID <[email protected]>
    [ https://issues.apache.org/jira/browse/SVN-3264?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17385799#comment-17385799 ] 

Daniel Sahlberg commented on SVN-3264:
--------------------------------------

One of C. Michael Pilato's bullet points (too much time spent on "installing" a working file) seems to have been addressed by a bunch of commits by kotkov starting at 1886490, for svn checkout/update, and later on also svn export.

> Large directory checkouts fail, connection forcibly closed
> ----------------------------------------------------------
>
>                 Key: SVN-3264
>                 URL: https://issues.apache.org/jira/browse/SVN-3264
>             Project: Subversion
>          Issue Type: Bug
>          Components: mod_dav_svn
>    Affects Versions: 1.5.x
>         Environment: Windows 2000
>            Reporter: Subversion Importer
>            Priority: Major
>             Fix For: unscheduled
>
>
> {noformat:nopanel=true}
> Large directory checkouts (with files hundreds of megabytes) fail on our
> relatively slow workstations. They function correctly on faster servers.
> The client side either reports:
>   REPORT of '/source/svn/Common/!svn/vcc/default': Could not read response body:
> An existing connection was forcibly closed by the remote host.
> or:
>   REPORT of '/source/svn/Common/!svn/vcc/default': Could not read chunk size: An
> existing connection was forcibly closed by the remote host.  
> Here /source/svn/Common is the URL path to our repository.
> Apache reports:
>   Provider encountered an error while streaming a REPORT response.  [500, #0]
>   A failure occurred while driving the update report editor [500, #190004]
> What seems to happen is:
>   * We initiate a checkout
>   * For each directory:
>     - All files are transfered through the network to <DIR>\.svn\tmp\text-base\
>     - Once all files have arrived, they are copied from
> <DIR>\.svn\tmp\text-base\* to <DIR>\
> The copy from <DIR>\.svn\tmp\text-base\ to <DIR>\ seems to take too long and we
> encounter some sort of timeout - at least, that is what seems to happen. The
> copy takes about five minutes, in one reproducible case.
> Server: Apache 2.2.8 or 2.2.9 with Subversion 1.5.0 or 1.5.1 on Linux RHEL ES 4
> update 5 (Apache and Subversion compiled from source)
> Client: Either TortoiseSVN 1.5.2, Build 13595 against Subversion 1.5.1, or
> Collabnet svn 1.5.1 (CLI) on Windows 2000.
> Things we have tried:
>   * Setting the Apache Timeout to 3600
>   * Disabling all authentication by removing the authz_svn and authn_bugzilla
> modules
>   * Setting the Apache LimitXMLRequestBody to 0 (unlimited)
>   * Setting the Apache DavMinTimeout to 3600
> A workaround is available: if the checkout fails, you can resume it by doing an
> update. But this is far from ideal.
> {noformat}
> Original issue reported by *frodol*



--
This message was sent by Atlassian Jira
(v8.3.4#803005)
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.