I was wondering why tools returned weird results, turned out that Wikidata Query (http://wdq.wmflabs.org/) hasn't been updating for a week. According to http://wdq.wmflabs.org/stats the last update is from 2015-02-08T09:14:51Z. This is a week ago.
Update: http://wdq.wmflabs.org/stats now mentions the time "2015-03-02T00:00:00Z". But most queries I try time out with "502 Bad Gateway". Could somebody please indicate the status and priority of this task? Thanks. --Fred Johansen
I haven't had time for a detailed look yet. Yuvi updated WDQ to the latest dump, which is where the timestamp comes from. He says both VMs are running 1 CPU at 100%, but don't reply to queries. That sounds like they try to update to the live state via the API, but got stuck in a loop.
A quick patch could be to disable incremental updates; then, the VMs would be stuck at the current timestamp, but at least answer to queries. Then we'd have to figure out where the updating fails. Not sure when I'll have time.
Please fix the WDQ queries, I need to debug and re-run my Python script. I did not know that such read-only API requires write updates and is unreliable.
Hoi,
Work has been done to make it better.. This morning it worked for some time
for me.
DO understand that it is not official software. It is exceptional that this
software has become so important. It is great to see that Labs Ops does
what it can. It however has limited options. I am ever so pleased that
Magnus and Yuvi work well together and I am frustrated that there is always
something else why it just does not work this time..
Yes, one issue has been resolved, there is apparantly more.
Thanks,
GerardM