You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Bob Breckenridge <bb...@hotmail.com> on 2008/04/15 18:10:52 UTC
Failing to get source code
ome users reported yesterday that they have been having problems with SVN for the past month. Apparently the problem was similar or the same as the fatal one we encountered yesterday, which I've listed below:
Error Berkeley DB error for filesystem 'C:/svn/repos/db' while beginning Berkeley DB transaction:
Error DB_RUNRECOVERY: Fatal error, run database recovery
Error bdb: PANIC: fatal region error detected; run recovery
I mentioned this problem to our IT Department, which suggested restarting the Windows machine, which I did but that did not help.
After consulting the Repository Administration section of the SVN Book, I found this suggestion:
1. Make sure that there are no processes accessing (or attempting to access) the repository.For networked repositories, this means shutting down the Apache HTTP Server or svnservedaemon, too.2. Become the user who owns and manages the repository. This is important, as recovering arepository while running as the wrong user can tweak the permissions of the repository'sfiles in such a way that your repository will still be inaccessible even after it is “unwedged”.3. Run the command svnadmin recover /path/to/repos. You should see output like this:Repository lock acquired.Please wait; recovering the repository may take some time...Recovery completed.The latest repos revision is 19.This command may take many minutes to complete.4. Restart the server process.
So I:
1. Shutdown the SVN Server running on our Windows machine.
2. Logged into the Windows machine as myself, since I'm an administrator and wasn't aware of any other user who owns and manages the repository. I am the person to whom the keys were given, when the original architect of our build system left the company a couple of months ago, so if there was some knowledge transfer that should have happened in regard to an admin user, I missed that.
3. I ran the command specified above. The response was similar to that specified above but happened remarkably quickly, so we suspected that it didn't actually do anything.
4. I restarted the server and the same problem occurred (specified at the beginning).
Then, one of our senior Developers suggested running the following command:
svnadmin recover -c -v -h C:/svn/repos/db
The error we got from the server, immediately after issuing this command indicated that the -c option was not valid.
And now the error we get, when attempting to SVN Update is:
svn: Berkeley DB error for filesystem 'C:/svn/repos/db' while opening environment:
svn: Invalid argument
svn: bdb: DB_REGISTER limits each process to a single open DB_ENV handle
Any advice on this would help. I don't even find the options for svnadmin recover...
_________________________________________________________________
Pack up or back up–use SkyDrive to transfer files or keep extra copies. Learn how.
http://www.windowslive.com/skydrive/overview.html?ocid=TXT_TAGLM_WL_Refresh_skydrive_packup_042008
RE: Failing to get source code
Posted by Bob Breckenridge <bb...@hotmail.com>.
We ended up restoring from a backup the day before the failures started happening. It looks like an upgrade of Subversion from 2 to 5, switching from BerkeleyDB to FSFS and using Apache 2.2 might be on the horizon for us. Is this the most robust configuration you would advise?Thanks,
Bob
_________________________________________________________________
Get in touch in an instant. Get Windows Live Messenger now.
http://www.windowslive.com/messenger/overview.html?ocid=TXT_TAGLM_WL_Refresh_getintouch_042008
Re: Failing to get source code
Posted by Ryan Schmidt <su...@ryandesign.com>.
On Apr 15, 2008, at 3:10 PM, Bob Breckenridge wrote:
> I don't know what my senior developer thought the -c -v -h options
> would do. I asked for the web page where he got that command but
> he did not respond. It looks like we'll just restore from backup
> and see what happens.
Don't forget to reply to the list too.
Restore from backups if you can. Retain the broken repositories in
case someone else comes up with a solution for saving them. I don't
have experience with BDB repositories myself. Unless there's a
specific reason why you want BDB repositories, you might also want to
consider migrating to FSFS repositories, as they cannot break in this
way.
http://subversion.tigris.org/faq.html#bdb-fsfs-convert
> > CC: users@subversion.tigris.org
> > From: subversion-2008b@ryandesign.com
> > Subject: Re: Failing to get source code
> > Date: Tue, 15 Apr 2008 15:05:35 -0500
> > To: bbrecken@hotmail.com
> >
> > On Apr 15, 2008, at 1:10 PM, Bob Breckenridge wrote:
> > > ome users reported yesterday that they have been having problems
> > > with SVN for the past month. Apparently the problem was similar or
> > > the same as the fatal one we encountered yesterday, which I've
> > > listed below:
> > > Error Berkeley DB error for filesystem 'C:/svn/repos/db' while
> > > beginning Berkeley DB transaction:
> > > Error DB_RUNRECOVERY: Fatal error, run database recovery
> > > Error bdb: PANIC: fatal region error detected; run recovery
> > >
> > > I mentioned this problem to our IT Department, which suggested
> > > restarting the Windows machine, which I did but that did not help.
> > >
> > > After consulting the Repository Administration section of the SVN
> > > Book, I found this suggestion:
> > > 1. Make sure that there are no processes accessing (or attempting
> > > to access) the repository.
> > > For networked repositories, this means shutting down the Apache
> > > HTTP Server or svnserve
> > > daemon, too.
> > > 2. Become the user who owns and manages the repository. This is
> > > important, as recovering a
> > > repository while running as the wrong user can tweak the
> > > permissions of the repository's
> > > files in such a way that your repository will still be
> inaccessible
> > > even after it is “unwedged”.
> > > 3. Run the command svnadmin recover /path/to/repos. You should see
> > > output like this:
> > > Repository lock acquired.
> > > Please wait; recovering the repository may take some time...
> > > Recovery completed.
> > > The latest repos revision is 19.
> > > This command may take many minutes to complete.
> > > 4. Restart the server process.
> > >
> > > So I:
> > > 1. Shutdown the SVN Server running on our Windows machine.
> > > 2. Logged into the Windows machine as myself, since I'm an
> > > administrator and wasn't aware of any other user who owns and
> > > manages the repository. I am the person to whom the keys were
> > > given, when the original architect of our build system left the
> > > company a couple of months ago, so if there was some knowledge
> > > transfer that should have happened in regard to an admin user, I
> > > missed that.
> > > 3. I ran the command specified above. The response was similar to
> > > that specified above but happened remarkably quickly, so we
> > > suspected that it didn't actually do anything.
> > > 4. I restarted the server and the same problem occurred (specified
> > > at the beginning).
> > >
> > > Then, one of our senior Developers suggested running the following
> > > command:
> > > svnadmin recover -c -v -h C:/svn/repos/db
> > >
> > > The error we got from the server, immediately after issuing this
> > > command indicated that the -c option was not valid.
> > >
> > > And now the error we get, when attempting to SVN Update is:
> > > svn: Berkeley DB error for filesystem 'C:/svn/repos/db' while
> > > opening environment:
> > > svn: Invalid argument
> > > svn: bdb: DB_REGISTER limits each process to a single open DB_ENV
> > > handle
> > >
> > > Any advice on this would help. I don't even find the options for
> > > svnadmin recover...
> >
> > The only option for svnadmin recover that I see in the book or via
> > "svnadmin help recover" is "--wait".
> >
> > http://svnbook.red-bean.com/en/1.4/svn.ref.svnadmin.c.recover.html
> >
> > What did your senior developer believe -c -v -h would do?
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Failing to get source code
Posted by Ryan Schmidt <su...@ryandesign.com>.
On Apr 15, 2008, at 1:10 PM, Bob Breckenridge wrote:
> ome users reported yesterday that they have been having problems
> with SVN for the past month. Apparently the problem was similar or
> the same as the fatal one we encountered yesterday, which I've
> listed below:
> Error Berkeley DB error for filesystem 'C:/svn/repos/db' while
> beginning Berkeley DB transaction:
> Error DB_RUNRECOVERY: Fatal error, run database recovery
> Error bdb: PANIC: fatal region error detected; run recovery
>
> I mentioned this problem to our IT Department, which suggested
> restarting the Windows machine, which I did but that did not help.
>
> After consulting the Repository Administration section of the SVN
> Book, I found this suggestion:
> 1. Make sure that there are no processes accessing (or attempting
> to access) the repository.
> For networked repositories, this means shutting down the Apache
> HTTP Server or svnserve
> daemon, too.
> 2. Become the user who owns and manages the repository. This is
> important, as recovering a
> repository while running as the wrong user can tweak the
> permissions of the repository's
> files in such a way that your repository will still be inaccessible
> even after it is “unwedged”.
> 3. Run the command svnadmin recover /path/to/repos. You should see
> output like this:
> Repository lock acquired.
> Please wait; recovering the repository may take some time...
> Recovery completed.
> The latest repos revision is 19.
> This command may take many minutes to complete.
> 4. Restart the server process.
>
> So I:
> 1. Shutdown the SVN Server running on our Windows machine.
> 2. Logged into the Windows machine as myself, since I'm an
> administrator and wasn't aware of any other user who owns and
> manages the repository. I am the person to whom the keys were
> given, when the original architect of our build system left the
> company a couple of months ago, so if there was some knowledge
> transfer that should have happened in regard to an admin user, I
> missed that.
> 3. I ran the command specified above. The response was similar to
> that specified above but happened remarkably quickly, so we
> suspected that it didn't actually do anything.
> 4. I restarted the server and the same problem occurred (specified
> at the beginning).
>
> Then, one of our senior Developers suggested running the following
> command:
> svnadmin recover -c -v -h C:/svn/repos/db
>
> The error we got from the server, immediately after issuing this
> command indicated that the -c option was not valid.
>
> And now the error we get, when attempting to SVN Update is:
> svn: Berkeley DB error for filesystem 'C:/svn/repos/db' while
> opening environment:
> svn: Invalid argument
> svn: bdb: DB_REGISTER limits each process to a single open DB_ENV
> handle
>
> Any advice on this would help. I don't even find the options for
> svnadmin recover...
The only option for svnadmin recover that I see in the book or via
"svnadmin help recover" is "--wait".
http://svnbook.red-bean.com/en/1.4/svn.ref.svnadmin.c.recover.html
What did your senior developer believe -c -v -h would do?
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org