We've started marking redirects within each one of the archive
snapshots. Starting on 7/28/09 each history and article snapshot will
entry so that everyone can easily identify which articles are in fact
This came as a request from Erik Zachte to further improve on our stats
collection and has allowed us to surface more user contributor stats
that are not filled with articles lacking significant content.
Tomasz Finc wrote:
> Tomasz Finc wrote:
>> Brion Vibber wrote:
>>> Tomasz Finc wrote:
>>>> Looks like we aren't getting in the replacement drives until mon/tues
>>>> of next week so the array will continue to be in degraded state until
>>>> then. Thankfully it's still under warranty so the turn around wont be
>>>> too bad. Tentatively putting the work to happen on Tuesday now.
>>> We were able to put in new disks today, but the raid array didn't fully
>>> recover. We got lots of I/O errors, and have been unable to run JFS
>>> recovery successfully so far.
>>> In the meantime we're running http://download.wikimedia.org/ off the
>>> copy of the last couple of dumps that had been copied to another server.
>>> The dump _files_ are there but currently the index is not.
>>> We're not 100% sure whether we'll be able to recover the earlier dumps
>>> or not, but of course more will be made soon enough. :)
>>> Some additional files such as the MediaWiki release download and DVD ISO
>>> downloads are still in process of being restored.
>>> -- brion
>> Thanks for the update Brion. I'll be checking in with Rob tomorrow to
>> see how ready the new set of drives are and if we are set to start
>> generating the snapshots anew.
>> Xmldatadumps-admin-l mailing list
> Sadly Fred and Rob took a look at the JFS storage and were not able to
> salvage any of the existing file system. We've gone ahead and started
> clean with the archives I made last week as the seeds.
> There will be one more day of testing tomorrow for drive removal and I
> expect to have the system back up and running by the end of the week. It
> should take about a week after to get a full cycle of all wikis.
Everything has been looking really good so far and I'm finally
comfortable in starting the snapshots back up. The only bit left to do
is to test by pulling a drive but that will have to wait till we have
RobH on site again.
Were currently running at five snapshot processes and if nothing weird
happens I'll dial it up to eight.
cc'ing xmldatadumpms-l on this.
Phil Adams wrote:
> hi tomasz,
> phil (philadams) here from #wikimedia-tech earlier today.
> i'm interested in looking at user behaviour on wikipedia, so i figured
> that the en wiki stub-meta-history would be a good place to start. i
> grabbed and uncompressed the 2009 07/02 version, and started just
> exploring it a little. i had a few questions:
> * is this dump supposed to contain ALL revisions to each en wiki page
> (articles and user pages in particular)? i ask b/c when i look at the
> revision history for (say) AmericanSamoa, the meta dump shows only 5
> or 6 revisions for that page, spread across time from 2001 to 2007.
> the en wiki history page online
> shows far more edits. what am i missing?
The XML files available on download are snapshots in time of our data
set. When each snapshot runs, the stub step gets a consistent view of
our database at that exact time. Any new revisions will only be
available in the next run.
AmericanSamoa is showing up just like it should in the snapshot because
its a redirect. If you take a look at
then you will notice that it's only had a handful of edits when compared to
note the space between the two words.
> * is there any sort of ordering to the history dump? it appears
> nominally alphabetic, although isn't strictly alphabetic.
The ordering is by page id.
> * if i have misunderstood the purpose of the meta dumps, but still
> wanted the same information, is my best recourse simply to d/l the
> entire en wiki dump? does that contain complete revision histories for
> all pages?
The only difference between a stub and the full history is the full page
content. If you don't need the content then they are effectively the same.
Why was the newst copy of enwiki with the full history removed from
the downloads site? I checked around and was only able to find one
place with it:
You'll want the "enwiki-20080103-pages-meta-history.xml.7z" file,
which is about 17GB. There is another file that is 130GB, but that is
the SAME thing, just compressed with bz2 insteaf of 7z, making it
larger, so don't get that one.
Tomasz, I am willing to volunteer my services as a programmer to help
with this problem in making full history enwiki dumps, if it is
possible (I can't donate hardware/money). What are the issues which
are causing it to be so slow and what methods are you employing to
I know that LiveJournal has some sort of live backup system using
MySQL and Perl, but couldn't find any details on their presentations.
You might be able to ask one of their developers for help, on their LJ
blog. Can Wikimedia afford a snapshot server? It doesn't need to be as
fast as the others.
In the long run, whatever this system is, it will probably need to be
integrated into some sort of backup, because it would be a huge pain
if something happened at the data center and you needed to restore
from the partial quasi-backups in the current systems.
How does the current dump method work? Are they incremental in the
sense that they build up on previous dumps, instead of re-dumping all
of the data?
For future dumps, we might have to resort to some form of snapshot
server that is fed all updates either from memcaches or mysqls. This
allows for a live backup to be performed, so it's useful for not just
Is it possible to suspend any individual slaves temporarily during off
peak hours to flush the database to disk and then copy the database
files to another computer? If not, we may still be able to use a
"stale" database files copied to another computer, as long as we only
use data from it that is at least a few days old, so we know that it's
been flushed to disk (not sure how mysql flushes the data...).
Of course, this may all be totally off, since I don't know a lot about
the current configuration and issues, so I'll take whatever input you
have to help work on something better.