Sunday Report

10,768 units have been validated this week

Assuming that a full generation 182 will be the last, there are 1,626,322 units still outstanding. I will re-start my forecasting once output stabilises.

The definition of accelerated and extreme are unchanged have moved up a generation.

There are 34 Extremes and 58 Accelerated units listed as one has crossed over. The numbers in their generations are 2,357 & 3,824.

The end daye for the project is too unpredictable but seems to be hovering around 2027 at the moment.

Mike

As it happens, luckily, after scrolling back quite some lines in my terminal's history (consisting of 5,000 lines), I found a moment on that screen where task _0 was still Pending Validation (!) and that task _1 had turned into 'No Reply':

 <1>   ARP1_0010993_138_0  Linux openSU  Pending Validation  2023-05-06T09:00:27  2023-05-07T15:09:59  5.54/5.62
 <1>   ARP1_0010993_138_1  Fedora Linux  No Reply            2023-05-06T09:03:59  2023-05-12T09:03:59  0.00/0.00
 <1> * ARP1_0010993_138_2  Fedora Linux  In Progress         2023-05-12T09:04:12  2023-05-15T09:04:12  0.00/0.00

(Of course, when I made my selection at the time, there was no number <44>, it was a different selection and ARP1_0010993_138 came out on top as number <1>; different selection, different rankings.)

So, of course at first there were only tasks _0 and _1. Then _0 was turned in, waiting for task _1 to complete. Task _1 got a No Reply after 6 days, so task _2 was issued to me. That's the situation that you see above (copied from my screen).

Now let's have a look at number <44> again from a selection that I made later on:

<44>   ARP1_0010993_138_0  Linux openSU  Error      2023-05-06T09:00:27  2023-05-07T15:09:59
<44>   ARP1_0010993_138_1  Fedora Linux  Valid      2023-05-06T09:03:59  2023-05-12T22:28:59
<44> * ARP1_0010993_138_2  Fedora Linux  Valid      2023-05-12T09:04:12  2023-05-12T19:59:14
<44>   ARP1_0010993_138_3  Linux Ubuntu  Valid      2023-05-12T21:41:47  2023-05-13T06:31:57

(It's an exact copy of the situation that you saw earlier in this thread.)
You can see here that the OpenSUSE task _0 has changed into Error. What happened there?
Well, one of the plausible reasons that task _3 was created, is that once my task (_2) was returned at 2023-05-12T19:59:14, tasks _0 and _2 were compared and - we don't know what happened exactly, but another task was needed for verification of the result - so task _3 was created and sent to another device at 2023-05-12T21:41:47. Quite a few minutes later, task _1 was returned at 2023-05-12T22:28:59. With task _3 still In Progress, my task _2 was declared Valid at Fri 12 May 22:55:23 UTC 2023 and so was task _1. We can't tell when the OpenSUSE task _0 was declared as Error, maybe at the same time as when tasks _1 and _2 were declared Valid. Anyway, when task _3 ended, they got the predicate Valid, too.

Adri

Sorry. I don't understand what is hinky. I'm missing something.

We have a validation issue at the moment. I think we all have a bunch of older WUs with 2 results that just sit in pending validation. Also newer resends may have validated just fine even though older WUs are still pending.

Does the hinky fall in this category, or is there something else to add to the list of problems?

- I'm so looking forward to the boost that will come when all these WUs are validated.

Al

Another reason for the last copy going out might have been that the 2 PVs didn't match, so an extra was sent to check.

Mike

Mike,

Thanks for that -- _0 was definitely late (and the system should've queued a retry at 04:35 or so on the 13th!), but I misread the outgoing time on _5 so that association was incorrect.

However, if the validator found a mismatch it should've set them to Pending Verification, so I'm unsure what actually happened here as a standard BOINC server platform should not behave like this if all validators, transitioners and feeders are running constantly![1]

The sooner they can get enough hardware to run on, the better :-)

Cheers - Al

[1] "Managing" the ARP1 feeder would explain why retries sometimes seem to take quite a while to actually go out [Waiting to send?]... There also has to be something strange regarding the validator(s) (or they're turning it/them off for prolonged periods...)