Last modified: 2014-07-03 10:44:10 UTC
Asaf Bartov: why is Hebrew Wikipedia without figures in http://stats.wikimedia.org/EN/TablesArticlesBytesPerArticle.htm ? (I do know Hebrew is a two-bytes-per-character language, but so is Arabic...)
All hewiki metrics where page content is involved, and hence full archive dump is needed, (links, avg page size, word count) suffered the same fate. Which indicates the job didn't complete normally (and it has run only once for all wikis on full archive dumps on new server for last 4 years). And indeed the log says: 22:09:58 Read xml dump file '/mnt/data/xmldatadumps/public/hewiki/20140328/hewiki-20140328-pages-meta-history.xml.7z' ***** String '<mediawiki' not found at start of file '/mnt/data/xmldatadumps/public/hewiki/20140328/hewiki-20140328-pages-meta-history.xml.7z'. Incomplete or corrupt file? Filesize 1770782752 bytes. File age -0.0 days. ***** Execution aborted. So I am rerunning with newest dump and it seems to run fine so far. I will also restart the cycle with exclusion of enwiki, which made the job stall (out of server resouces?) I'll have to investigate enwiki issue, but that will probably be after London. To be sure, this is the extra run on different server than normal wikistats run, on a slower than monthly schedule, only for page content related metrics. -- I had not looked at the report for long time. BTW I'm more puzzled by metrics on that same page for Kannada, Maltese, Khmer, Assamese, etc where trend totally stands out from the rest. But I won't have time to look into it now, four weeks till London, and some serious coding to do for my editor migration study.