> -----Original Message-----
> From: Geoff Rowell [mailto:geoff.rowell_at_gmail.com]
> Sent: vrijdag 30 oktober 2009 3:16
> To: 'Neels Janosch Hofmeyr'
> Cc: dev_at_subversion.tigris.org
> Subject: RE: mem leak?
> Neels Janosch Hofmeyr wrote:
> > here's a report from crazy_imp on #svn:
> > With 1.6.6, a checkout of freebsd was interrupted by a network
> (takes longer than a day o_O). After that, `svn cleanup' takes some
> hours to
> complete, and after that, `svn update' grows to about 3 gig memory
> > Don't know what to make of it, maybe someone else would like to
> > Some chat snippets...
> > <crazy_imp> svn, Version 1.6.6 (r40053)
> > <crazy_imp> neels: all i've done is: svn co
> freebsd - due reconnects, i need to run svn cleanup and svn up again.
> > <neels> so your checkout was interrupted, and you had to cleanup,
> then you
> updated to try and finish the checkout. right?
> > <crazy_imp> yes
> > <crazy_imp> asked yesterday why svn does not reconnect to the server
> keeps on going after a reconnect, nobody answered the question (think
> should be possible, the client still knows it's state and which stuff
> needs to fetch)
> > <borg-queen> crazy_imp: are you really sure you want to checkout the
> > <crazy_imp> borg-queen: yes :)
> > <borg-queen> crazy_imp: then maybe a checkout "--depth immediates"
> updates of individual subdirs may work better
> > (etc., 2009-10-30 at about 13:00 CET)
> I'm also still (since 1.6.5) seeing a memory leak using SVN 1.6.6 on
> 2003 R2 over http protocol. Requires an Apache restart every few days.
Which httpd version do you use?
Do you use compression in your httpd config?
(There is a known issue on using compression)
Received on 2009-10-30 15:32:55 CET