[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

RE: SVNADMIN Recover and Verify show good but repository still hosed???

From: Karan, Cem \(Civ, ARL/CISD\) <CKaran_at_arl.army.mil>
Date: 2004-11-10 16:32:20 CET

I'm running an OS X box and had terrible problems with my repositories
getting hosed on a regular basis (like every 3-4 days), all because of
BDB. My problems were sometimes similar to what you had; the
repositories passed recover and verify (including the BDB toolset using
the catastrophic settings), but the repo was still wedged. Someone on
the list suggested that I switch to the FSFS type repositories, that
they don't go south as easily (although they are slower, and may cause a
timeout under apache). I did that and reloaded from a recent dumpfile.
Maybe you can switch to an FSFS based repository and that will help??

Disclaimer: I've JUST started running the FSFS based repository; I don't
know if it is going to get hosed as well. If you don't hear any
complaints from me in the coming months, then it probably worked for
me...

Cem Karan

> -----Original Message-----
> From: m christensen [mailto:dfs@xmission.com]
> Sent: Tuesday, November 09, 2004 5:07 PM
> Cc: users@subversion.tigris.org
> Subject: Re: SVNADMIN Recover and Verify show good but
> repository still hosed???
>
>
> No real revelation as to what happened or why.
> A dump and reload ran without any errors and a full rebuild
> of a local
> working copy ran without a hitch.
>
> Marc
>
> > I ran a dump, it appears to have worked, or at least I saw
> no errors.
> > Maybe I'll try to rebuild from that.
> >
> > Nick_Gianakas@sybari.com wrote:
> >
> >> I experienced this same problem last week. Upon getting the 500
> >> Internal Server Error, I stopped all SVN access (DAV) and ran
> >> verify. It reported the last revision (which was
> correct)--no errors
> >> or warnings.
> >> However, if I tried svnadmin hotcopy, I got a bunch of errors that
> >> the DB (BDB) needed recovery. I ran recover followed by
> verify and
> >> everything came back w/out problems. Re-running hotcopy
> resulted in
> >> the same set of warnings/errors as the first time.
> >>
> >> Some notes of interest:
> >> - While the repo was being backed up (svnadmin hotcopy
> --cleanlogs ),
> >> a user tried accessing it (update). That was the first
> time this 500
> >> error came up. It then snowballed until the repo was completely
> >> inaccessible.
> >> - My system setup had not changed in weeks.
> >> - Setup:
> >> - SVN 1.1.0
> >> - All repos are of type BDB
> >> - Apache 2.0.52 (WebDAV)
> >> - x86; Gentoo Linux
> >>
> >> W/out much help from Google and limited time, I dumped the repo and
> >> reloaded it. So far, no problems.
> >>
> >> Regards,
> >> Nick G.
> >>
> >>
> >
> >
> ---------------------------------------------------------------------
> > To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
> > For additional commands, e-mail: users-help@subversion.tigris.org
> >
> >
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Nov 10 16:33:49 2004

This is an archived mail posted to the Subversion Users mailing list.