Thank you for all the posts with information. It is nice to see progress. Some WUs going well and getting validated, and others with errors (this is what testing is for).

I'm having a bit of FOMO as I didn't get any of the stress testing WUs, I'll finish up the few WUs of the backup project, in hope of getting some WCG WUs soon.

Hopefully with the increase in communication and eventually a new status page, we won't have to crowdsource this information. I'll have to start playing one of the games on the forum, as I do enjoy seeing you all post.

edit: no more FOMO. I got some testing MCM WUs

Regarding validating without credit...

I wonder if they are running all six (?) validators yet (which might explain why only 17 of my 1400+ returned results so far have validated!) and, if so, whether the way they've tried to integrate validation and assimilation is causing a delay (in classic BOINC, credit is assigned by the validator wrapper as part of finishing off for a specific WU)

I seem to have got credit for two short tasks and 6 long tasks (and the long task credits look about right...); however, I notice that the claimed credit for any tasks that haven't got granted credit yet sometimes show 202.5 claimed and on other occasions show a claim that is much closer to what I'd expect for the amount of CPU time used!

It also seems that there are a lot of tasks with one returned result flagged Valid whilst the other is still flagged Pending Validation (as noted earlier by Adri) -- that might just be a feature of their new method, but then again...

So (as noted by Ralf) there are probably still issues to be resolved; it also remains to be seen whether the ones with credit have actually been assimilated as well...

Insights from others welcome, of course :-)

Cheers - Al.

P.S. It comes as no surprise that the new validation mechanism doesn't seem to be doing anything with non-MCM1_9xxxxxx workunits. I do hope that is a feature of their testing, not an indication of a [future] problem...

Additional Notes:
I have seen none of my WU's in this control batch error out yet.
There are 265 WU's IN PROGRESS on my systems at this time.
If successfully completed, they are all expected to transition to PENDING VALIDATION. So, time will tell where things go from here,.

I think the scheduler is a bit busy now. Most of the time there is no problem, but from time to time the response to a request is:

Scheduler request failed: Error 408

Unfortunately, as reported by several users, the current mass of MCM1 work seems to be producing some bad validation-related entries in the database, including cases where one result is marked Valid (and probably has the claimed credit granted (see below) whilst the other is still Pending Validation. I'm not sure whether the transitioner might spot that at some point and mark an error, but it's not a valid workunit/record state!

And as for the claimed credit -- one of my systems seems to have ended up with a 202.5 credit claim for the normal-length tasks instead of a far more reasonable 80 to 85!

I hope these things turn out to be code bugs rather than a side-effect of some [accidental?] interference with BOINC's normal processing sequences for returned work[*1], as the code in places like the transitioner and the wrapper programs for validation and assimilation is interlinked in all sorts of ways...

Cheers - Al.

*1 Standard BOINC validator will tag the [two] validated results as Valid and mark the WU as ready for assimilation, whilst setting the "next transitioner check time" of the WU to a value designed to tell the transitioner to ignore the WU! It is then up to the assimilator logic (which can be a null action as far as individual records are concerned) to mark the WU as potentially ready for deletion and set a more normal check time.