Ralf,

The streak has to be calculated somehow as in a standard BOINC database (which I doubt IBM tampered with...) the "work done today" numbers are only held in the host_app_version table and the host table, not the user table, so the obvious way to do it was to tack it on to the existing WCG profile stats run. As for reasons for not running said stats, it may or may not be connected...

That's the only reason for my "whimsy" about the streak being the issue here - I obviously didn't apply enough smileys to it!

And I wonder if either Adri or you read this post from a month or so ago about internal and external statistics and the likely outcomes when they [try to] turn the internal stats stuff back on again; it was in response to a pair of your postings, but neither of you commented further (which might have been a blessing!) :-)

Cheers - Al.

[quote

nivrip, you're getting a steady stream of all project wu's with no issues? You must have tapped into those "Yorkshire Domes" in your neck of the woods?

.

A quick search did find the documentation of BOINC's Database Purging Utility (db_purge), and this says:

Meaning, for the WCG-specific stats it should be possible to parse the db_purge archive files and this way nothing is really lost.

While it would definitely be nice to see something besides zeros in my current production stats, WCG being back to 21+ hours deferral due to lack of work is at least to me a much more pressing problem than knowing if I've hit another badge.

Well, that's because

a) it is not a post from a technical person at WCG/Krembil given some insights as to what is happening (or not).

b) is actually not explaining even from an outside view what is happening, or not.

Something either has been thoroughly been borked during the transition, or the actual reason might be as simple as not having the server powered back up that runs the twice daily internal WCG stats..
And that run is independent, apparently not related to the external (BOINC) stats file creation, which runs (ran) at different times. And in the past, I don't recall ever that there were problems with the internal stats generation, beside if there was a (un)planned outage due to database upgrade/maintenance and knreed and/or uplinger had to kick the process that created those external stats file back into gear. Again, never was there an issue with the internal stats. And as you can see, I am with the WCG project not just since yesterday and I do have returned quite an amount of work for the various project during this time.That combined with the fact that I am working in IT, doing sysadmin/database/programming work for a few decades by now, I think I can make observations and deductions as to what may or may not be at least the source of a problem.

The real big issue here is that Krembil is absolutely lacking any transparency about what is happening.. I have been sticking with WCG for all those years because in the past, this was the most well run DC project out there. Very rarely having real issues, and if something happened, which according to Mr.Murphy is always a possibility, things were sorted rather quickly, usually in a matter of a couple of days, if not hours.
Rosetta's Baker Lab at WU has gone though problems for example as well, but never have I seen such a mess (to put it simply) as currently with WCG at Krembil.There are single dude projects run from a garage that are better organized and reliable.
The download issues "seemed" to be solved a few days ago, now it's all back as bad as it ever was. No discernible progress.Likewise with the stats issue.

What some people forget is the overall time frame we are dealing with here.
The move was announced 1 year ago! (And no, I don't feel like celebrating).
It was 5 months before the system was shutdown at IBM and the physical transition to Krembil was supposed to start.
It was another 4 months until the first "Test WUs" were being send out.
And now they are "testing" for another. 3 months, with nothing really to show for.

And Krembil is not even able to provide regular updates, with some real progress (or problems) info, just that eventually someone will look into an issue, they do have it on their radar...
And no, I am not the only one complaining, maybe just the most vocal. Not to mention all those newcomers, people that have a real interest in the science behind the WCG projects, that are being put off and leave before you even saw anything from them...

Ralf

Ralf:Oh no, not again! - I do recall a problem with the internal stats generation.
Right.
I'm afraid there was.
Three years ago, something unusual happened with the internal WCG stats.

Adri