[Mulgara-dev] Very large backup file

Paul Gearon gearon at ieee.org
Wed Jun 6 17:43:15 UTC 2007


Sorry your message took so long to show up.  It got flagged as having
too large an attachement, but I let it through to the list yesterday.
I don't know why it didn't show up until today.

First off, when did you get this checkout of Mulgara?  This error log
looks like a problem that got solved earlier this year.  I'm sorry
that we can't just give you a build number to work on (that should be
working with the upcoming release).  Subversion info (svn info) should
help you here.

As for your backup, you can look at it using a streaming gzip
decompression program, such as zcat or zless.  I'm sure you can get
these native for Windows, but cygwin also comes with them.

The backup file is quite straightforward.  It starts out with a header
like this:
XA BACKUP V6
Tue Sep 28 17:22:08 EST 2004
RDFNODES

Then it lists all the URIs and literals, along with the internal
numbers they get mapped to.  An example is:
4 <http://mulgara.org/mulgara#Model>
6 <http://mulgara.org/mulgara#preallocatedNode>
7 <http://mulgara.org/mulgara#preallocatedNodeModel>
5 <http://mulgara.org/mulgara#preallocatedNodes>
67702 <http://www.cogsci.princeton.edu/~wn/concept#100001740>
67707 <http://www.cogsci.princeton.edu/~wn/concept#100002086>
312302 "\"He fell down to his knees"
298047 "\"Her old vigor returned\""
304538 "\"I could hear rifles explode\""
.
.
.
266166 "zymosis"
53577 "zymotic"
165906 "zymurgy"

Finally, the triples (well.... the quads) are all listed, in terms of
the numbers described above:
TRIPLES
2 3 4 2
5 6 2 7
5 6 3 7
5 6 4 7
.
.
.


If you can stream through your backup file and it looks something like
what I just described, then it was working correctly.  If not, then
could you please let me know, as this is something we would really
need to address!


Regards,
Paul Gearon

On 5/29/07, Ben Hysell <BenH at viewpointusa.com> wrote:
>
>
>
>
> Hope everyone had a great holiday weekend, it appears while we were out of
> the office Mulgara also took a little bit of a holiday, and sadly we had to
> restart her after three weeks of continued up time.
>
>
>
> We run backups on our production system every four hours with a script that
> runs the backup, through the ItqlInterpreterBeanService (we are all Windows
> and C#), if the backup fails try again five times then fail.  (I don't
> believe this is a SOAP issue because we have seen this same thing happen
> when attempting to kick off backups through the web ui).
>
>
>
>
>
> Our 12:30 pm backup kicked off yesterday at 12:30 and has been running for
> 20 hours.  The backup file itself has been monotonically increasing in size
> since 12:30 yesterday.  Our typical backup size is currently around 200
> megs, when I finally stopped it the file had reached 1.1 gig.
>
>
>
> Attached is the log output from midnight on the 28th till when I shut it
> down this morning on the 29th.  The first attempt to create backup
> backup_2007-05-28_12-30.txt.gz failed and we tried to create
> backup_2007-05-28_12-52.txt.gz which is the backup that ran all night and
> grew to 1.1 gig.  I just attempted to open up the 1.1 gig backup, however
> gzip told me I had run out of room on my hard drive, I did have 14 gig
> free...
>
>
>
> Also, as an aside, our server1 directory is roughly 27 gigs right now,
> normally after a restore the directory is only 6 gigs.
>
>
>
> Thoughts, ideas?
>
>
>
> Thanks
>
>
>
> -ben
> _______________________________________________
> Mulgara-dev mailing list
> Mulgara-dev at mulgara.org
> http://mulgara.org/mailman/listinfo/mulgara-dev
>
>
>



More information about the Mulgara-dev mailing list